back to article Lies, damn lies, and KPIs: Let's not fix the formula until we have someone else to blame

Monday has arrived once again and with it the sweet, sweet music of a reader's darkest IT misdeeds in The Register's weekly Who, Me? feature. Today's story comes from "Alban", who in the mid-Noughties worked at a mobile telephone company "in a country more known for its musical than technological prowess". "At the time," …

  1. Oliver Mayes

    If this calculation was so important to the company, why was it written by one guy with no oversight?

    1. big_D Silver badge

      Because that is how such things always turn out.

      Best practices and good intentions have nothing to do with reality.

      1. DBH

        The script that decided how many people would be involved with writing and checking the KPI script unfortunately had a cancellation error that resulted in the output being a random number, which by chance was usually about 1

    2. LeoP

      Been there, done that

      Yours humbly is from a country that is better known for its musical than for its technical prowess and whose biggest telco was at that time a mostly state-owned partner of Vodafone (now a daughter of America Movil after a more than boitched privatization job).

      Their claims of 95+ percent availability seemed rather absurd then, about 90 was what everyone would have guessed (remember that this means double the downtime).

      Now: The clue for the why is in the first paragraph: It was a mostly state-owned telco. Management was picked by their party affiliation and willingness to do what some politician wanted, not by what was best from a technical perspective.

      1. BillG
        Pirate

        Re: Been there, done that

        "Some managers – including the chief of operations – had part of their bonus tied to this indicator."

        Dum dum DUUUUUUUUUM!

        That's the music that went off in my head when I read that sentence. When managers protect their bonuses, it's gloves off. But when CxO level protect their bonuses the laws of the jungle apply and if you're not careful you'll wind up in the dunpster out back.

    3. phuzz Silver badge
      Facepalm

      I guarantee that in every single company, there will be an irreplaceable business process which depends on an undocumented spreadsheet.

      It doesn't matter if the entire company is ISO9000 accredited and running the latest and greatest ERP. Somewhere there will be a spreadsheet, set up years ago, which calculates all of the stock replenishment rates, or something equally important. Every, single, company.

      (At my last job it was a spreadsheet that pulled data from an Access DB, which was just a means of querying a SQL database, but the entire month-end procedure relied on numbers that it spat out.)

      1. A.P. Veening Silver badge

        At my last job it was a spreadsheet that pulled data from an Access DB, which was just a means of querying a SQL database, but the entire month-end procedure relied on numbers that it spat out.

        You are lucky the spreadsheet pulled the data. In a previous job the data had to be put into a csv-file and about half the time it wouldn't import correctly because of PC-settings like delimiter and decimal separator. And yes, there were completely valid reasons for staff to have different PC-settings as they were working in different languages. The problem with the delimiter was finally solved by inserting a single line to the front of the csv with "sep=<delimiter>", but the decimal separator remained a pita.

      2. Prst. V.Jeltz Silver badge

        often accessing the code of that spreadsheet is passworded and the owner long gone

      3. Anonymous Coward
        Anonymous Coward

        In one of my previous contracts, I was asked to rewrite a pricing calculation spreadsheet because the original data source was being decommissioned (after a six-month search had located said database on a dedicated server under someone's desk in a different country).

        In the process I discovered several significant errors where the data being pulled did not match the documentation, so the calculation was effectively pointless. Obviously, that hadn't stopped them using it as the basis for all customers' subscription renewal quotes for about five years!

        Anonymous for obvious reasons.

        1. Mark 85

          n the process I discovered several significant errors where the data being pulled did not match the documentation, so the calculation was effectively pointless. Obviously, that hadn't stopped them using it as the basis for all customers' subscription renewal quotes for about five years!

          As long a there's PROFIT, manglement won't care if it's pointless.

          1. Anonymous Coward
            Anonymous Coward

            Management don't care about profits at all; that's for the owners. As long as management salaries are enormous and rising, profits are no more important than processes or customers or the rank and file.

      4. david 12 Silver badge

        I'm just old enough to remember that spreadsheets predate spread sheet programs. They used to be big sheets of paper on which the numbers were pencilled in, then later inked in.

        So when we first moved to Supercalc macros, one of the things we discovered was that the departmental spreadsheets didn't actually add up and agree with the global monthly reports, or the annual budget spreadsheet that had gone to Board of Directors.

        Fortunately, the Board of Directors doesn't actually need to know the real numbers anyway....

    4. Anonymous Coward
      Anonymous Coward

      Likely because it wasn't ever known or cared about until someone saw it broken out as a number, and made it important.

      Along the lines of:

      <slaps server> "This one has more gold on the riser cards than the previous model, which makes calculations less error prone!"

    5. Sureo

      "why was it written by one guy with no oversight?"

      Very simple - the results looked reasonable so everyone assumed it was working as intended.

      Proper testing would have revealed the problems.

      1. DavCrav

        Re: "why was it written by one guy with no oversight?"

        "Very simple - the results looked reasonable so everyone assumed it was working as intended.

        Proper testing would have revealed the problems."

        I just found a problem in a short script that I wrote a couple of weeks ago. I had a list of variables and as I slowly solved the equations I substituted in the determined variables. What I hadn't noticed was that I searched the list for the variable name and then changed it. If one unknown was proved to be exactly equal to another, then my script changed the first instance of it, not the correct instance of it.

        This problem did not appear until quite a long time after I started using it.

      2. 's water music

        Re: "why was it written by one guy with no oversight?"

        Proper testing would have revealed the problems.

        This is why nobody in management likes Proper Testing

  2. Efer Brick

    Bonus?

    Would've thought it'd be the tin tack for the local bigwigs,

    Said bigwig did the pragmatic right thing

  3. Anonymous Coward
    Anonymous Coward

    Building entry log

    I worked at a company that implemented an ID-card based building entry system. There was an additional module that could be used for timekeeping, but the company decided not to buy that.

    They then sent out an email to staff saying that the entry system was tied to the alarm system and when the last person left the building, the alarm would automatically be turned on. That meant that everybody had to book in and out, otherwise, if you forgot and worked late, the alarm would suddenly go off as you moved in your office... They also informed everyone that the system did not have a timekeeping function and they would not be monitored.

    It worked very well.

    Then management discovered the log file and that, from the log file, you could extrapolate the times people entered and left the building... They quickly turned the log file into an Excel spreadsheet with entry and exit times and a sum of hours on the premises per employee.

    This report was then handed over to the CEO every week. He even went one step further and had a special version written to track individual employees over the whole year - he used it to sue one of the other directors out of the company.

    Talk about toxic!

    1. paulf
      Big Brother

      Re: Building entry log

      "Paulf & co" moved to a new building about 6 years ago. The new building had entry gates so every entry and exit could be logged (if they wanted to). As a fully paid up cynic, that wouldn't bet against how low Manglement can go (especially what I've seen at paulf & co over the last few years), I've always assumed they do use it to log presence and hours in the building even though this is not officially confirmed/denied. That they don't use it for anything useful (e.g. has everyone successfully evacuated after a fire alarm activation) just increases my cynicism level to "Bitter".

      Any one who thinks Manglement / Human Remains wouldn't take the opportunity to log stuff that could be used against employees is being incredibly naive. I cover my arse by keeping my own record of my hours, along with a reason if I have to leave early/arrive late in case I get called on it.

      1. CrazyOldCatMan Silver badge

        Re: Building entry log

        has everyone successfully evacuated after a fire alarm activation

        If your access control system doesn't unlock the doors (and hence disable the entry system) in the event of a fire alarm then it's a very, very bad access control system is is quite possibly illegal..

        Counting people out in the event of a fire alarm is why you have fire wardens.

        1. paulf
          Holmes

          Re: Building entry log

          It would be illegal, in England & Wales at least, as it would breach paragraph 14 (2) (f) of The Regulatory Reform (Fire Safety) Order 2005:

          (f)emergency doors must not be so locked or fastened that they cannot be easily and immediately opened by any person who may require to use them in an emergency;

          but that wasn't the point I was making.

          My point was that Manglement are quite happy to use the access control system (like ours and the one mentioned by the AC OP) to log how many hours people spend in the office for their own nefarious purposes. What they won't do is use the access control system for good purposes e.g. checking how many people have successfully evacuated the building in an emergency. This can be achieved by placing badge readers at each Assembly point for people to confirm they have left the building and have made it to the specified muster location.

          Fire Wardens counting people out isn't that helpful if they don't know how many have been counted IN, in the first place. This is where the access control system comes in as it does know not just HOW many but also WHO!

          1. Martin an gof Silver badge

            Re: Building entry log

            placing badge readers at each Assembly point

            How's that going to work if the computer which produces the report is in the building? If the power is off because there really is a fire? If it's a fire in the server room? If you can't get on the internal network from a laptop when stood at the mister point?

            Paper records kept at the desk and grabbed by a nominated member of staff on the way out, on the other hand...

            M.

            1. doublelayer Silver badge

              Re: Building entry log

              That's why a system actually used for safety purposes would need to have a remote backup. If there were multiple sites, a mirrored version between the two would help. If not, the records could be stored in any number of remote places. It's quite doable. However, it's quite unlikely ever to be considered a priority despite the required tech already having been installed and the real benefits it could provide.

            2. ricardian

              Re: Building entry log

              I worked for a large Government department in a secure building. We logged in and logged out so security always knew who was in the building. In the event of a fire the security staff would print out a list of who was in the building and use it as a check that everyone was out. Alas, the security staff were low in the pecking order and were only supplied with an elderly PC and an even older dot matrix printer. Printing a list of several hundred staff would take longer than it would for the building to burn to the ground - as long as the PC didn't crash, the power didn't fail and the printer didn't run out of paper or ink ribbon

              1. Terry 6 Silver badge

                Re: Building entry log

                As soon as I read "In the event of a fire the security staff would print out a list " I pretty much knew what was coming next......

        2. Grooke

          Re: Building entry log

          How do the wardens know how many people were in the building to begin with?

          The access control logs should tell you. They don't have to lock the doors in an emergency to help judge if an evacuation was successful.

    2. MJB7

      Re: Building Entry Log

      "They also informed everyone that the system did not have a timekeeping function and they would not be monitored."

      If they were in the EU and they tried that trick now, they would be in all sorts of GDPR shaped trouble. One of the requirements of GDPR is that you have to tell people what data you are collecting (they did that, sort of), *and what you are going to use it for* - they didn't do that.

    3. Anonymous Coward
      Anonymous Coward

      Re: Building entry log

      , from the log file, you could extrapolate the times people entered and left the building...

      I wrote some code to log who logged in ,where and when, etc for a pretty toxic nhs outpost

      1. Anonymous Coward
        Anonymous Coward

        Re: Building entry log

        Is the 'pretty toxic' not otiose to mention of an NHS outpost?

    4. Anonymous Coward
      Anonymous Coward

      What about tailgating?

      Unless they have mantraps at every entrance/exit, there will be people who won't card in to enter. Even if you have security watching and enforcing everyone carding in on entry, I've never seen a company that enforces a "no tailgating" rule on people leaving. The issue of non-authorized people getting in only exists on entry, building security doesn't care about multiple people exiting together.

      So I guess the lesson is you should make sure you tailgate people whenever possible to make the data set useless - at least where you are concerned. Except when you are coming in early or working late, then you should make sure to card out properly, so anyone who tries to manage based on "butts in seats" will know about the long hours you pulled.

      A company doing this will probably also put trackers into your work computer that will know when you are active, and these days maybe even facial recognition in the hall in front of the bathrooms so they can see how much time you spend in there. Probably no escaping companies run by Big Brother wannabes, other than to quit and work for someone else who values your work output not your time spent.

      1. swm

        Re: What about tailgating?

        I once forgot my badge so, entering with a bunch of people, I just made a motion with my hand like I was showing my badge - my hand was empty. I had no problem entering

    5. Mark 85

      Re: Building entry log

      Then management discovered the log file and that, from the log file, you could extrapolate the times people entered and left the building...

      Many of the places I've been, the troops don't know that it's related to timekeeping. Thus, Joe swipes open the door, and 5 people following him go into the building. The only reason we never heard someone scream about paychecks was that the office was salaried and the shop was hourly but had their own timeclock for clocking in and out.

      The places were the employees were told it was also for timekeeping, everyone swiped their badge at the door.

    6. Bibster
      FAIL

      Re: Building entry log

      All this talking about entry logs...Makes me think of this bit on internet-lore that I once read, about a door where you'd have to badge to open it, but it sort of randomly responded.

      One night a bloke heard the door opening, and opening again etc. whereas he was pretty sure he's the only one in, and he decides to figure out what was happening, and the story ends with a windows machine with an enormous log file or so, causing the program that handles the badge, to be slow as hell, causing the door to 'randomly' open.

      No idea where to find that story. Would be a fun re-read!

      1. Psmo
        Go

        Re: Building entry log

        I think you mean this one.

        It's a classic.

  4. Anonymous Coward
    Devil

    "blamed the less rosy new numbers on them and their new software"

    Well, if the KPI were computed correctly before, they would have just got the blame earlier... if they were smart, they could also blame the supplier management had chosen...

  5. A.P. Veening Silver badge

    "As you can expect," Alban said sadly, "this didn't improve my trust in managers."

    I'd say Alban was hopelessly naive if he even imagined trust in managers can get above absolute zero.

    1. Tweetiepooh

      Optional

      The managers have to be above absolute zero else there is no where to rank lawyers and salesmen.

      1. Anonymous Coward
        Anonymous Coward

        Re: Optional

        Hey! As a sales manager for a law society I resent that comment!

        Signed:

        O.Kelvin

      2. A.P. Veening Silver badge

        Re: Optional

        But we can trust lawyers and salesmen ... to fuck things up. Managers can't even manage that.

        1. ICPurvis47
          Boffin

          Re: Optional

          I was a member of a team that was developing a new Motor Control Centre, which used Vacuum Contactors to switch the current on and off. The trucks containing the contactors had to be physically moved onto and off the BusBars for isolation purposes, and I was detailed to do the development on the mechanical linkages. When we had a working prototype, it was presented to Production Engineering department for them to make any modifications they required to make it easier or more productive to be built on the production line. Unfortunately, one of their "Improvements" made my linkage mechanism non functional, so we had to have it back in our development workshop so that I could make it work again without infringing any of their "Improvements". This took about two weeks (because a lot of the parts were made outside the company), and we were then blamed for the product being late to the market, even though the delay was initially caused by Production Engineering department.

      3. Mark 85

        Re: Optional

        The managers have to be above absolute zero else there is no where to rank lawyers and salesmen.

        I thought that's what negative numbers were for?

    2. smudge

      I think Alban is being a bit churlish, since the manager could easily have thrown him to the wolves...

      1. codejunky Silver badge

        @smudge

        That was my take on this. The manager probably saved him from losing his job or taking unnecessary flack for something that while wrong didnt cause irreparable damage.

    3. Pascal Monett Silver badge

      Well what was he expecting ? That his manager would commit seppuku and admit that the Board had been chasing ghosts and granting bonuses for a year ?

      It was the only rational choice - for a manager.

      And the alternative could have been Alban getting fired. He should be grateful.

      1. Doctor Syntax Silver badge

        "And the alternative could have been Alban getting fired."

        Not really. He now knew and could presumably prove they'd been submitting false returns all this time. Not a good idea to fire someone who has your balls in a vice.

        1. Mark 85

          Not really. He now knew and could presumably prove they'd been submitting false returns all this time. Not a good idea to fire someone who has your balls in a vice.

          Isn't the the first rule for any BOFH? Any BOFH knows very quickly where the bodies are buried and who's doing what to who. It's called "insurance".

          1. Doctor Syntax Silver badge

            Knows where the bodies are buried and where to buy the quicklime.

        2. veti Silver badge

          No, because he was the one who'd been giving them the false information, which they'd been submitting (as far as we know) in good faith. There's no way for Alban to spin that as him being better than his management.

          Owning up would certainly have been embarrassing, and might have resulted in management casualties - but Alban would definitely have been first out of the door.

          1. doublelayer Silver badge

            I'm not sure that would necessarily have been the case. The data was wrong, that's true. But the data was also untested. It could be argued or inferred that the managers were responsible for at least attempting to verify that the data was correct, and that they had failed to do that. If someone in HQ was of a suspicious type, they might assume that management had specifically engineered the script to function improperly and was trying to blame the person who caught them, or that management and the developer conspired to do it incorrectly and could both be held responsible.

            Blame is a complex substance; when it's dropped from the ceiling it never just falls straight to the floor. Every time, it splatters everywhere.

  6. Alister

    We had a similar thing happen with one of our clients - big rail travel provider - who we used to send performance figures to every month for their public facing website, page views and click-throughs and so on, which they based their ongoing strategy on. This is in the days before Google Analytics was a thing, and we had been using Webtrends to gain the metrics, but the client decided they didn't want to pay Webtrends prices, so they asked us to switch to ClickTale.

    The following month, the results were completely different and bore no resemblance to the historical results from Webtrends - as I suppose Clicktales' algorithms were different to those used by Webtrends. This caused much consternation, and lots of angry emails as to why the website was no longer getting the same traffic as it used to.

    If they had looked at actual online ticket sales, they would have seen that not much had changed, but it took a lot of explaining.

    1. Doctor Syntax Silver badge

      "it took a lot of explaining."

      The explanation should have been simple, although probably beyond manglement's understanding: you get what you pay for.

      1. Nick Kew

        The price of snake oil varies rather a lot - as in this instance.

  7. jonathan keith
    Joke

    It's his life.

    And that, my friends, is the story of why Alban left IT, gained a Dentistry degree, and recorded a worldwide hit single.

    1. A.P. Veening Silver badge
  8. Bronek Kozicki

    bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

    I wouldn't worry too much about it - if they were not interested in the testing of the formula, that means they were not interested in the results. So, business as usual - get paid for attending meetings, not for the results these meetings supposedly are meant to deliver.

    1. Warm Braw

      Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

      To the extent that you can actually measure the effect of management on business (and it isn't easy because of varying accounting methods and the many external variables), studies seem to conclude that it is actively malign. Most acquisitions and mergers destroy value, most actively-managed investment funds perform worse than trackers if you don't get to pick the measurement period.

      So, people make absurdly huge amounts of money out of, effectively, pretending to manage businesses: huge bonuses from fake KPIs, free shares, compensation and new jobs for agreeing to acquisitions, bonuses for acquiring and massive consultancy fees for "expertise".

      It may seem strange that this situation persists, but the shareholders of actively-managed businesses, tend to be the active managers of pension funds, so as long as they all get their cut, the employees and the pensioners aren't really considered in the whole process and all that matters is that the charade continues to play out.

      1. SVV

        Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

        I have come to the conclusion that most large companies perform more or less as well as they can, given the managers, staff and budgets in place, and all KPIs and metrics are totally irrelevant compared with all the time wasted on company politics and manouverings for personal advancement within the organisation.

        This is a wonderful example of something that probably did the company a lot of good by preventing any changes that may have been made as a result of the targets not being met, and therefore risking changes that were ineffective, counterproductive, incompetently managed or simply woefully misconcieved.

        1. Anonymous Coward
          Anonymous Coward

          That's a big assumption

          You're basically assuming that nothing managers do can have any positive impact, only negative. If a telco knows availability is below 95% and wants to reach that threshold, they will try to make some changes. If they get proper feedback they'll eventually find something that works and know how to actually improve it. It sounds like based on these random numbers everything they tried "worked" in the sense that next month's numbers were probably better so they "fixed" problems that didn't exist.

          I mean, some guy could say "let's spend $2.5M a year having someone pressure wash bird shit off our cellular tower antennas" and when availability improves that guy gets a pat on the back and a bonus, and they will pressure wash birdshit from then on. That $2.5 million a year they spend on something useless is money they can't spend on something useful, like adding new towers in areas where they are lacking, or upgrading connectivity to towers that are overloaded.

          I can't think of ANY case where having bad information would be better, because it will prevent the company from doing something dumb. Having bad information makes it almost certain that something dumb, when proposed, will appear to "work" and continue forever. You seem to be assuming management as actively hostile and trying to make things worse. They may want to cut spending, and sometimes cut in places they shouldn't, but if they have no way to see the bad effects of their cuts they could use this information to cut budgets far beyond what even the most evil manager would - they'd say "hey we are still obviously cutting fat because our metrics show availability is still over 95% despite cutting tower maintenance expenses by 90%!"

          1. 's water music

            Re: That's a big assumption

            You're basically assuming that nothing managers do can have any positive impact, only negative

            That is obviously a somewhat facile view but if you accept that the things that managers are trying to influence are so complex and the tools that are available have very nebulous links to the outputs (which themselves may be difficult or impossible to measure effectively) then there is a tremendous moral hazard that tempts everyone to seek out evidence to support the expected outcome from the action taken. So even if we assume that more managers are acting in good faith than bad, confirmation bias will apply selection pressure in favour of the thing we did over the thing we didn't or aren't doing regardless of its actual impact.

    2. imanidiot Silver badge

      Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

      I would say that's business as usual anyway. Most of the time the "numbers" the bigwigs work with are so abstracted and convoluted someone might as well have pulled them from their ass

      1. Nick Kew

        Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

        Particularly true of the work for MoD clients I encountered back in the 1980s.

        Need a number, but have no data? Pluck one from the top of your head. Use it, and explain that it has no basis in reality. Try a couple of other numbers to see how much difference it makes: this is called a parametric study and should be good for a few more taxpayer-funded man-hours.

        Next person to need that number then references you. In a couple of years, all the caveats are lost, and your number has become the definitive reference. Noone questions it any more. Or rather, anyone who questions it soon finds myself out on a limb and looking for the next job.

        Next time you watch a war film and the hero speaks dismissively of those backroom boys who have no clue, you know he's right[1]!

        [1] Am I still OK to refer to a hero of a war film as male?

        1. A.P. Veening Silver badge

          Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

          [1] Am I still OK to refer to a hero of a war film as male?

          As long as it is the hero, I don't see a problem. It sure would be different if it were the heroin of a war film.

          1. Warm Braw

            Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

            the heroin of a war film

            I suppose they can be addictive and are pushed by a shady cartel...

            1. A.P. Veening Silver badge

              Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

              I suggest you look up the differences between heroine (addictive drug) and heroin (female hero).

              1. Anonymous Coward
                Anonymous Coward

                Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

                Before you start hurling brickbats, check your facts. Hint: You have it the wrong way round.

              2. John Doe 12

                Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

                Ha ha - this is the best fail I have seen in a while here. Rule #1 when being a pedantic prick on any forum - triple check you have your facts right :-D

                1. A.P. Veening Silver badge

                  Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

                  So I mix up two words in a foreign language, influenced by the fact that in my mother tongue the correct spelling for that drug happens to be "heroine" with a terminal "e". Have fun, but better make sure you native speakers of English don't mix up "their", "there" and "they're" (alphabetically ordered).

                  1. nichomach
                    FAIL

                    Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

                    See, the problem isn't that you mixed the words up, the problem is that you attempted to lecture someone else who picked up on that and had a little (non-malicious) fun with it. Don't blame other people when you pwn yourself?

                    1. A.P. Veening Silver badge

                      Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

                      No problem at all, I upvoted John Doe 12 for it and corrected my down vote for Warm Braw to an upvote, just returning the favour with a pet peeve.

                    2. Anonymous Coward
                      Anonymous Coward

                      Re: and had a little (non-malicious) fun with it.

                      As he seemed to be replying to a post including the words "pedantic prick", I think maybe "non-malicious fun" might be a not terribly good description.

                  2. Kubla Cant
                    Headmaster

                    Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

                    in my mother tongue the correct spelling for that drug happens to be "heroine" with a terminal "e"

                    It may seem presumptuous to query your knowledge of your unspecified mother tongue, but I find this surprising. "Heroin" is Bayer's original brand name for the drug, spelt in a similar way to their contemporary drug Aspirin.

                    Also, your excuse only holds water if your mother tongue spells the lady hero "heroin".

                    1. FrogsAndChips Silver badge

                      Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

                      In French the word is "héroïne" for both the female hero and the drug.

                      1. Anonymous Coward
                        Anonymous Coward

                        Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

                        Both addictive in small quantites.

                    2. A.P. Veening Silver badge

                      Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

                      It may seem presumptuous to query your knowledge of your unspecified mother tongue, but I find this surprising. "Heroin" is Bayer's original brand name for the drug, spelt in a similar way to their contemporary drug Aspirin.

                      Heroïne, analogous to morphine.

                      Also, your excuse only holds water if your mother tongue spells the lady hero "heroin".

                      Nope, my mother tongue (Dutch) spells that "heldin", female form of "held".

                    3. Michael B.

                      Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

                      A quick scan of Wikipedia shows that at least French and Dutch spells it Heroine despite the Bayer name.

                      1. A.P. Veening Silver badge

                        Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

                        Correct, words ending with "in" (and "innen" for plural) indicate female persons, words other than names ending with "ine" (plural "ines") indicate a substance. Examples includes (next to the already mentioned heroïne) words like benzine, vitamine, cocaine, codeïne and morfine. As a rather famous Dutch junk (Herman Brood) once remarked: 'You should try everything ending with "ine" at least once, except for benzine'.

                  3. xeroks

                    Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

                    I thought - at first - you had got them the wrong way round deliberately. All you needed to do was add "/s" to pull off a double bamboozle.

                    1. A.P. Veening Silver badge

                      Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

                      I needed proof again I am still human ;)

                      1. Fading
                        Terminator

                        Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

                        Can be upsetting to fail a Turing test when one is talking to one's self.....

                        1. A.P. Veening Silver badge

                          Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

                          "To err is human, ...".

                          1. veti Silver badge

                            Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

                            "... but to really fuck things up, you need a computer."

                            Wisdom from the 1970s, there.

                            1. A.P. Veening Silver badge

                              Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

                              I was referring to that other variation, "... but to forgive is divine."

                              Not many gods around here ;)

                              1. Anonymous Coward
                                Angel

                                Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

                                Hello, my friend. I am everywhere. What can I do you for?

                2. Psmo
                  Headmaster

                  Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

                  Pendantic pricks seem to be what language is for!

        2. irrelevant

          Re: bigwigs had spent the last year working on ... "numbers totally unrelated to reality"

          "Next person to need that number then references you. In a couple of years, all the caveats are lost, and your number has become the definitive reference. Noone questions it any more. "

          My dad was a mathematician. He used to tell me the story of when he worked for Avro, designing (parts of) aircraft. Must have been 1960s or so. Now the wings of a plane are filled with ribs, which give them the strength. But when he asked how they were specified, dimensions, separation, etc., he was referred to the previous model of plane, just take it's values and adjust a bit for the different weight. He did though have the formulae for calculating all this based on wing loading, etc., but nobody actually used them, they just took the previous plane's values, and fiddled a bit, and had been doing this for years!

          So, my dad being a perfectionist, decided to do it the hard way and worked everything out from scratch. As this needed to be done individually for each rib it took quite a while, even with the help of the nascent computer department. (I think this was where he learned PL/1, his only ever experience coding computers. It was still a "submit a job, wait a few days for the results" mind.) In the end, though, he came up with a wing design that took up far less material, and was thus lighter and more efficient, and saved the company much money! He was encouraged to write it all up as a thesis for a degree, but got it turned down as it was "too industry specific." It should still be sat on a bookshelf at mum's., I really should go borrow it sometime..

  9. Dave Watts

    Hardcoded KPIs

    Yup, I've experienced these as well... My favourite from a previous employer was a 'very important' dashboard of service management stats for an exec team. If a measure fell below it's pre-ordained acceptable level teams would be rewarded with great scrutiny, micro-management and potential loss of bonus; Bad Things (TM). One of the measures was "customer satisfaction",and it had been difficult to get agreement of the definition of this subjective measure, people's pay and happiness were at stake, so that metric just had a default value and no calculations. And thus it stayed for many months. Everyone was happy...

  10. Anonymous South African Coward Bronze badge

    He decided to hold back Alban's fix for the KPI until "the mobile access network guys had a big upgrade of their systems and blamed the less rosy new numbers on them and their new software".

    The BOFH in me agrees 100% with this. Muhuhahaha.

    1. Wellyboot Silver badge

      One has to ask.

      A more fitting BOFH fix would be to ensure the system continued, possibly with the average being slowly nudged towards 96/97% for the improved benefits.

      Did Alban get any sort of promotion for his efforts in identifying this key strategic information for management? After all, the boss was obviously not caring a jot about publishing the revised previous numbers merely looking for a patsy for when the numbers 'go bad' in future.

      1. davetalis

        Re: One has to ask.

        A similar issue is covered rather well in From the Earth to the Moons "Spider", probably the best way of dealing with it https://youtu.be/XuL-_yOOJck

  11. jmch Silver badge

    KPIs

    One of the things about KPIs, performance measurement, reward / bonus systems etc, is that whenever these are introduced, employee behaviour stops optimising for best practices, product delivery, better service to clients, and starts to optimise for the KPIs.

    If the KPIs are well chosen and accurately reflect reality, that's great for the business. All the more commonly though, KPIs tend to measure things that are easy to measure / calculate, and important things that are difficult / impossible to calculate (eg customer satisfaction) are ignored.

    1. Whiskers

      Re: KPIs

      When these were introduced in my last workplace someone asked what the initials stood for. I came up with "keeping people ignorant".

    2. Anonymous Coward
      Anonymous Coward

      "starts to optimise for the KPIs."

      You mean just executives started to optimize for meeting shares target values instead of actually managing a company? What a surprise...

    3. TRT Silver badge

      Re: KPIs

      Customer satisfaction is hard to rate. I never believe ours. You see people tend to use satisfaction questionnaires as a means to express displeasure, but not necessarily when things went as could be expected. So you have to apply a correction factor for that tendency, calculated from response rate.

      1. Andytug

        Re: KPIs

        Yep... if you use NPS (which has a scale from 0 to 10) there are very rarely any scores between 3 and 8, you're either brilliant or cr@p.

        1. doublelayer Silver badge

          Re: KPIs

          And that's often because I know any score below a nine (or sometimes a ten) is seen as an indicator that someone has failed. Even though eight out of ten is pretty good, I know there will be a discussion about why there weren't two more points there. Maybe this will cause a lot of problems for the person concerned. Maybe they'll send me a dozen more surveys to try to extract the reason for my withholding those two points. And I succumb to laziness and just assign nines and tens if the thing was fine or above.

          1. A.P. Veening Silver badge

            Re: KPIs

            Repeat surveys will only see declining results until they get the message. And it is pretty rare for me to give above 7.5 anyway.

            1. Terry 6 Silver badge

              Re: KPIs

              There's a genuinely quite good app on my phone. I forget what it is. And it keeps popping up with "give us a rating" messages. With stars, but no option for a comment.

              So over 12 months it's gone down from 5 to 1 because I reduced by 1* for the annoying pop up each time. It's not the asking that annoys. It's the repetition. Do they think it'll get magically better?

      2. CrazyOldCatMan Silver badge

        Re: KPIs

        You see people tend to use satisfaction questionnaires as a means to express displeasure

        AKA "one IT support-related screwup erases 99 IT support-related successes".

        (Also AKA: no-one notices if you do a good job. Everyone notices the one time you don't)

        1. Anonymous Coward
          Anonymous Coward

          That's why asking for "satisfaction" with IT is the wrong question

          Ask them only what they are dissatisfied with, and let them list as many separate items as they wish. Then you go through their lists and pick out the things that are repeatedly mentioned, like "access to files on server is too slow" or "login in the morning takes forever" etc.

          You measure improvement by the total number of items they list, and you know what to target based on the top 5 or so items across everyone's lists. If you can make changes so that "login in the morning takes forever" goes away from the list, then you know you've improved something. Of course "I hate the new GUI" after an upgrade to Windows 10 might take over as the new #1 in the meantime, and there's not much you can do about that, but you target the things you can and ignore the stuff they will just have to live with.

          Would work the same for say telcos or cable/satellite companies. People care about the things that frustrate them and aren't likely to give any positive feedback (other than maybe being happy if you put up in a new tower near their house so they now don't have to be on the top floor to get reception) so you just want to ask what they are dissatisfied with. But yeah I know, upper management doesn't want to hear "dissatisfaction is down" because that's measuring a negative, they want to hear "satisfaction is up". That's just not realistic, people don't work that way.

        2. 's water music

          Re: KPIs

          AKA: no-one notices if you do a good job. Everyone notices the one time you don't

          A young man moves to a village in Wales and gets talking to an old man from the village. He asks the old man what his name is; the old man gets very irate at this point and says: "See that line of houses over there? I built them all, but do they call me Jones the house builder? Do they hell! See those railway lines over there? I laid them all, but do they call me Jones the engineer? Do they hell! See those bridges over that river? I built them all, but do they call me Jones the bridge builder? Do they hell! But, a long long time ago, I fucked *one* sheep..."

        3. ICPurvis47
          Angel

          Re: KPIs

          We used to have a saying "Doing a good job here is like wetting your pants in a dark suit - it gives you a nice warm feeling, but nobody notices". I don't know where it originated, but it seemed germane at the time.

      3. Anonymous Coward
        Anonymous Coward

        Re: KPIs

        Years ago I worked on a hell desk and Manglement decided we should follow up the faults we cleared with a satisfaction survey... except they were too cheapskate to employ an outside company to do the surveys and we would have to do them ourselves... "would you like to take part in a survey?"... "F%$k off, your $H&* service means I'm half a day's work behind!!"

        We happened to have a friendly regular 'customer' who said just fill it in yourselves and, provided it's honest I'll back you up. At the end of the year we got an award (only a certificate!) for the number of surveys submitted (near 100% with the same name)... proving that no one, apart from us, actually analysed any of them for problem areas or better ways of working, and were just looking at the numbers.

        Having moved on from helldesks, my faith in customer surveys was not improved more recently when I discovered that, with KPI, picking a 'neutral' score of 5 for "I don't have an opinion either way" or "I've never used that service" is far from neutral and I would be flagged as a detractor

        1. Terry 6 Silver badge

          Re: KPIs

          We've probably all filled out those forms.

          And sometimes we just want to say "fine, whatever" I'm sure. because, frankly, the person did what they were meant to do. Which was what we had expectation that they would do. And what we wanted them to do

          So do we put a 5? After all they did no better or worse than expected. And there may well have been no opportunity to do better, anyway. Because they did their job.Or is just doing you job worth an "excellent" in 2019?

    4. Terry 6 Silver badge

      Re: KPIs

      And there you have the whole education system of the UK summarised. Only we call them SATS not KPIs

      1. Andytug

        Re: KPIs

        Same goes for GCSE results, where certain academies have been caught "off-rolling" some of their more challenging pupils (including those with special educational needs) in order to keep their pass rates high and look good.

        1. CrazyOldCatMan Silver badge

          Re: KPIs

          keep their pass rates high and look good

          In the mid-80's the Comprehensive I went to had one of the highest exam pass rates in London. With a mysterious dip in the year I took my A levels..

          Which, I'm sure, is pure coincidence.

    5. nichomach

      Re: KPIs

      Isn't the axiom that when a metric becomes a target it ceases to have any value?

      1. Paul Kinsler
        Alien

        Re: Isn't the axiom that when a metric becomes a target it ceases to have any value?

        I find that targeting the correct metric is absolutely vital when building wormholes and other Einstein-Rosen themed constructions. :-)

      2. Mark #255

        Re: KPIs

        Indeed, the (possibly apocryphal/exaggerated) canonical example being that of Soviet nail factories; their output being measured by a single, simple metric of either "number of nails" or "mass of nails" produced, and producing either millions of tiny tacks, or tons of railroad spikes, respectively.

      3. jmch Silver badge

        Re: KPIs

        Indeed - Goodhart's Law

    6. Paul Cooper

      Re: KPIs

      It's a well-known trend in the design of racing yachts, dating back as far as handicapping rules have existed. You set up handicap rules for a class of yachts, with the aim of ensuring that it is the skill of the crew that wins races, not a cunning design. Immediately, designers start to figure out how they can get a better handicap without damaging the performance too much. The long overhang at bow and stern of some 1920s yachts (which are now regarded as classics) resulted from that - waterline length was penalized in the rules of the time, but length when heeled wasn't. Similarly, at one time there was a rule that sail area behind the rudder wasn't counted in the handicap rules - so you got lots of yawls (which have a second mast behind the rudder).

      1. JimC

        Re: Yacht Racing Rules

        In fact the situation is worse than that. In many cases, especially at the higher levels the rating rules are required to both ensure the best crew wins and to encourage design innovation. The logical flaw is ensured because there will be a lobby for both aims, and like any good management team the rule administrators wish to satisfy both lobbies, leaving the backroom techies to sort it out.

        The other issue is that its much easier to design a boat that's slow, but not as slow as the rule thinks it is, than it is to design a boat that's fast, and even faster than the rule thinks it is.

        All the resulting situations and recriminations are familiar territory to anyone in business.

        There have also been some monumental throwing of toys out of prams when some very well heeled executive (because such commission racing yachts) commissions a very expensive "innovation" intended to be extremely slow in the eyes of the rules, and the rule writers take a different interpretation. This results in said executive having spent a lot of money making his boat go slower, only to find that the rule makers have adjusted his handicap fairly, so all the money has bought him no competitive advantage at all!

    7. Anonymous Coward
      Anonymous Coward

      Re: KPIs

      It's the classic trap: once a metric becomes a target it is no longer an effective metric.

      1. Anonymous Coward
        Anonymous Coward

        Re: KPIs

        Sounds like a variant on quantum physics - a particle is a wave is a particle, depending on how you try to measure it.

    8. Baroda

      Re: KPIs

      (IMHO) here follows a wonderful quote: I hope I have it down accurately. It repays careful reading.

      'When the right thing can only be measured poorly, it tends to cause the wrong thing to be measured only because it can be measured well. And it is so often much worse to have good measurement of the wrong thing - especially when, as so often is the case, the wrong thing will IN FACT be used as an indicator of the right thing - than to have poor measurements of the right thing.'

      Tukey, J.W. (1979) methodology and the statistician's responsibility for both accuracy and relevance, Journal of American Statistical Association, 1974, No. 368, pp 786-793

    9. Anonymous Coward
      Anonymous Coward

      Re: KPIs

      The UK NHS has been guilty of this too, with doctors doing all sorts of under-hand tricks to make their statistics look better than they were in reality.

  12. chivo243 Silver badge
    Trollface

    Kick that can full of fudge down the road

    Fudge IT! It's the next guy's problem!

  13. Anonymous Coward
    Anonymous Coward

    Reminds me of two things..

    In a previous life we had to report on KPIs but, for one reason and another, they weren't particularly accurate or informative but took a while to compile.

    At one team meeting we were asked to sugest ways of improving efficiency and I suggested scripting the output of the KPIs so generally they were okay but randomly one or two would drop into amber or even red. The values would be stored and the following months would incrementally move back into the 'green'. I proposed that this would save an awful lot of time and be just as accurate as the current reports but my sugestion was rejected.

    We also had a target of answering phones within 3 or 5 rings (I forget) but of course people realised that if you didn't get to the phone by the requisite ring (or weren't sure that you had heard the first ring) the best thing to do was let the call ring out. There was no measure for 'lost' calls.

    As somebody above said, 'measuring what's easy to measure rather than what you need to measure'.

    1. phuzz Silver badge

      Re: Reminds me of two things..

      "We also had a target of answering phones within 3 or 5 rings"

      I'd have tried changing my ringtone to something that was at least minute long. Easy enough to get there in time then :)

      1. Doctor Syntax Silver badge

        Re: Reminds me of two things..

        Carefully balance the handset so it falls off after the first ring. Who said anything about talking to the customer?

      2. CrazyOldCatMan Silver badge

        Re: Reminds me of two things..

        I'd have tried changing my ringtone to something that was at least minute long

        *Waves*. Got plenty of Prog songs for that. 29-minute ringtone - yup.

        1. quxinot

          Re: Reminds me of two things..

          I'd have suggested a certain John Cage piece. It's only a bit under 5 minutes, but if they're not tracking if it doesn't get answered anyway....

    2. thomasallan80

      Re: Reminds me of two things..

      You think that's fun, my brother used to work in a supermarket and as a section manager he would get KPIs run based on what things were sold at (usually). When he was in charge of bakery he really blew his lid because if you made rolls with a total cost of say $0.20 or $0.30 and a retail sell price of $2 if you then discounted the goods to $1 to get old stock out of there the KPI would mark you at -$1 but if you threw the same goods in the bin you were marked at -$0.30.

      I'm assuming the business was trying to get everyone to keep everything as close to the ticket sell price as possible because other departments would be totally stuffed if they did that, but loosing 'half' the retail price and making the company some money then gets you penalised for making money but throwing the stuff away and not making any money at all and it looks a lot better on the report?

      Since I wasn't involved in those stupid KPI's I don't know what the f they were thinking but I know they scheduled everything (wages for the week/month, the cost of goods being used, cost of facilities shouldn't be too hard to calculate) and then measuring the dollar value of actual goods sold should be easy to compare to estimated cost as another way to double check ANY KPI. Then having a section in the KPI's for total dollar amount discounted, thrown out, etc and compared to previous months shouldn't be too hard to add either.

      1. Doctor Syntax Silver badge

        Re: Reminds me of two things..

        It's what I call the seductive power of numbers. If you produce a number it must be right and it must be meaningful. That doesn't encourage taking care in measuring things. Measuring properly can be hard, at best it needs thinking about and at worst involves expensive equipment (CERN!) and processes; if you've already got a number why put in the extra work to get a "bettter" number?

        Back in the early days of the OU I had a couple of students who were science teachers. Why faff about with the single balance and its weights when they had digital scales at school that read out numbers? So where do your numbers come from? How do you know they're right as opposed to just believing what the display says? The first time we got a digital balance in my lab I paid a visit to the local Weights and Measures department - handily our neighbours - to borrow a couple of their standard weights so I could check.

      2. drunk.smile

        Re: Reminds me of two things..

        This KPI is probably about right because they'd need to calculate uplift.

        I don't imagine that the reduction in price causes large amounts of people to buy bread that otherwise wouldn't. This means the people buying $1 bread are mostly those who would otherwise have purchased the same bread for $2.

        So a $1 loss vs a $0.30 loss for throwing it away.

      3. cdrcat

        Re: Reminds me of two things..

        But the KPI is correct - there's a sunken cost fallacy in there somewhere.

        Sell 10 stale buns at $1 each = $7 profit.

        Throw away 10 stale buns, make 10 new buns, sell at $2 each = $14 profit.

        Throwing away buns is likely to increase profits (assuming most new buns get sold, ignoring elasticity or price discrimination, and ignoring some other issues).

    3. Anonymous Coward
      Anonymous Coward

      Re: Reminds me of two things..

      "We also had a target of answering phones within 3 or 5 rings (I forget) but of course people realised that if you didn't get to the phone by the requisite ring (or weren't sure that you had heard the first ring) the best thing to do was let the call ring out. There was no measure for 'lost' calls."

      You don't run the statistics office for Network Rail by any chance, do you?

      1. Killfalcon Silver badge

        Re: Reminds me of two things..

        I think it was Arriva who got caught with some of their Cardiff train routes ending a stop or two past the main station, and that last leg - that almost no-one took - had an extra 10-15 minutes in it's schedule beyond what was needed. Net result was the train could drop 200 people off late, then finish the route "on time", and in the stats that "whole train journey" was what counted, not the number of passenger journeys.

    4. TomPhan

      We have a target of patients not waiting longer than 10 minutes - so the trick is to call patients before that then have them sat in a room till someone can get to them, because that time isn't monitored.

    5. Anonymous Coward
      Anonymous Coward

      Re: Reminds me of two things..

      BT had a requirement that a certain %age (lets say 95%) of calls to directory enquiries had to be answered within 1 minute. The easy way to meet that was to adjust the call queues so that any calls coming in when the queues were almost full were dropped with a busy signal. Since only the calls which actually made it to the queue were eventually answered, meeting the 95% target was easy.

      It was assumed that about 10% of calls to a particular DQ centre got dropped, but one day someone actually hooked up some measuring equipment, and discovered that the centre was so undersized that 90% of calls were being dropped with busy, yet the 1 minute time was consistently met...

      1. Killfalcon Silver badge

        Re: Reminds me of two things..

        That's often a problem: you fudge the KPIs to avoid getting in trouble, but then the KPIs look like you don't need any assistance.

        How do you tell if a KPI is meant to indicate "we need more staff" and when it's meant to mean "this team is useless and should be outsourced"? Well, you can't. The people making that decision know what will happen if they tell you, so they're not gonna. And even if they do, their replacement next year might have a completely different idea.

        1. Terry 6 Silver badge

          Re: Reminds me of two things..

          Or does the KPI mean "This team are underperforming" or " this team are hampered by poor management decisions".

          Any guesses which way that toast will fall?

  14. Anonymous Coward
    Anonymous Coward

    Randomised network risk

    We studied the the effect of randomised replacement versus targeted replacement for a complex network. Perhaps surprisingly, it has similar effect on total risk as targeted replacement (especially for large populations and a slow replacement rate, say 50 years average life and 10,000+ items). With a filter to eliminate recently replaced items, random is nearly as effective at dealing with risk as a targeted approach.

    Clearly random is much lower an overhead too, compared to targeted risk. Only compliance commitments and safety legislation rule it out. For non critical hardware I’d suggest random is a valid model.

    Overly complex models are overrated and put faith in black boxes rarely earned or tested. The opinion of a good engineer is far more reliable, however government funding quangos don’t trust such answers preferring to rely on the black box for fear that the engineers will pull a fast one to print cash. Somewhere in here is an operating model that has the advantage of human factors without being compromised by human factors either.

  15. ColinPa

    KPIs are insensitive

    In the days of 6 sigma a large software project I was on had a formula to tell us our projected defect rate.

    - The design phase will introduce 5 defects per line of code.

    - Unit test will remove 70 % of defects and introduce 4% more defects

    - FV will remove...

    and the answer is 6 defects over the life of the project. This means we do not need any support teams.

    As we went through the development process the numbers got tweaked. . Unit test will remove 72% of the defects etc... and the answer is still exactly 6 defects over the life of the project.

    I set up a spread sheet to adjust the "constants" by small amounts and I could get an answer from 6 to 6000 defects, but management were not interested ( we have to have no more than 6 defects or we cannot release it)

    Within one month of releasing it, we had hit over 20 defects - as it would not install. We quickly set up a support team

    1. Anonymous Coward
      Anonymous Coward

      Re: KPIs are insensitive

      Yes, I recently discovered a bug which was the result of an omission from specification. One system sends a list of three values, unsorted . The receiver (in another company) just takes the first value, no sort.

      So two pots of six sigma design principles applied in two companies didn’t catch stupidity.

      test, test, test some more, and then throw production at it for six months. Assume you need devs for at least a year post launch to fix post launch errors. More for complex systems.

      1. A.P. Veening Silver badge

        Re: KPIs are insensitive

        Nice one, any indication which was correct? Or were both wrong and should it have been five values?

        Absent other specifications, as a programmer I normally assume single values.

        1. This post has been deleted by its author

        2. Anonymous Coward
          Anonymous Coward

          Re: KPIs are insensitive

          The correct value is the minimum of the three. Under certain circumstances, for example cold weather, high wind or heavy rain; performance could temporarily be improved by taking advantage of the extra cooling that the air/water/cold provides.

          Items two and three in the list indicate the limit of the next item in the chain, therefore, are an indicator of whether it was worth requesting an enhancement to item 1 or not. I.e. if it's cold out, and there's a big gap between 1 and 2, then a weather related enhancement may be possible.

          There are a series of mistakes here on both sides.

          1) In sending the list of three, we assumed the receiver was intelligent enough to deduce that the lowest value is the limiting value. (No, because it's a dumb computer reading the file with no instructions)

          2) The receiver assuming the data are sent in particular order (no, the order in the file is simply the order the plant data were entered into the system - future modifications to hardware would result in changes in the order)

          3) The specifications for both send and receive application not defining any particular rules.

          The lesson is to specify absolutely bloody everything in size 60 font, Bold, Italic, Underline because you can never underestimate the incompetence of an offshore programmer that does not understand your problem.

          1. Mark #255
            Trollface

            Re: KPIs are insensitive

            In the problem space where the implementer doesn't know everything you do, is it their fault when they fail to read your mind correctly infer the 'correct' implementation which matches the complete model, or is it yours for relying on (unspecified) assumptions?

            (Icon because I realise the tone of this is rather provocative)

            1. Anonymous Coward
              Anonymous Coward

              Re: KPIs are insensitive

              Not provocative at all. The purpose of the process and tools were to aggregate up the performance data for multiple equipment items; in such a way the limiting items are identified. This is stamped all over the business requirements, and as such, the "minimum of" rule is written everywhere with the exception of in the xml schema that carries the output over the fence.

              The third party developer reading this file in did not understand this requirement (despite the entire business requirements saying as much) and the formatting of the xml was such that they could, in ignoring the requirements, make erroneous assumptions about what to read in.

              The lesson remains, to prevent mistakes like this being made, to specify absolutely bloody everything, no matter how seemingly obvious it is to everyone else concerned.

          2. A.P. Veening Silver badge

            Re: KPIs are insensitive

            In one word: Lovely

            And you are right, everything should must be specified, otherwise you end up with military acronyms.

  16. A K Stiles
    Facepalm

    Measurable KPIs

    We had a good laugh when a KPI for a system was implemented that suggested a target of "X% of everything would be recorded in the system" - It took a surprisigly long time for those who had picked it to understand that the derision was because we had no way of knowing what stuff wasn't recorded in the system...

    1. A Non e-mouse Silver badge

      Re: Measurable KPIs

      Our contracts people love putting KPIs in contracts. The problem is that mot of them are unmeasurable or the sample size is too small to be meaningful.

  17. aj68
    Facepalm

    No PORT in a storm

    Years ago, I worked for a well known decision support software company where the MD took pride in using their own software tool to query data, produce charts and track the business. One of his key data sources was the spreadsheet updated by the Finance team on the shared network drive of all customer shipments so the MD could keep his finger on the pulse.

    Being a curious database man, I looked at the underlying data and my eye was drawn to the Port column which had entries such as "Win", "Windows 3.1", "Unix", and also lots of what seemed to be random letters, which I thought a bit strange. Later, stood in the MD's office while he brought up some of his charts and drew some conclusions, I saw him enter the letter "P" as a parameter for the column in his query, so asked him what the Port column was supposed to be.

    "Ah" he said, smiling in his condescending style "If you'd been here since I started the company, you would know it's not 'Port', it's 'P or T' - Production or Trial and I only want to track the Production customers, not those who are only trialling the software".

    Obviously that information was also news to most of the Finance team who had been entering the data based on their assumption of 'Port' for many years! :-)

    As they say, Garbage In Garbage Out.

  18. Anonymous Coward
    Anonymous Coward

    I was responsible for running bonus calculations for our sales team. The rules for these were adjusted from time to time. One year, one of the KPIs had its sales threshold doubled, but the time restrictions removed. I duly implemented this and carried on... for about 15 months, when one of the managers queried one of the results and it emerged that the threshold had never been doubled, so for the last year and a bit we'd potentially paid a bunch of bonuses that hadn't been earned.

    Fortunately most of the affected items did go on to clear the higher threshold, but there were several bonuses that paid out when they shouldn't have. The management decision was not to revert any that had already been calculated, but just go on with the correct threshold in place for the future.

    1. 's water music

      The management decision was not to revert any that had already been calculated, but just go on with the correct threshold in place for the future.

      So not Oracle then?

  19. lglethal Silver badge
    Holmes

    hmmm

    The bigwigs had spent the last year working on what Alban delicately described as "numbers totally unrelated to reality".

    Since when have managements numbers ever been related to reality????

    1. Doctor Syntax Silver badge

      Re: hmmm

      Since when have managements numbers ever been related to reality????

      FTFY

  20. Anonymous Coward
    Anonymous Coward

    Ignore waste if your job depends on not reporting it

    Many years ago I was contracted to a large service company whose contracts tended to be upwards of $5m. HQ had been asking for Cost of Quality (CQ) figures for several years but management had managed to stall on the basis it was too hard to measure and, besides, almost all contracts met or exceeded their budgeted profit. One of my roles was to seek areas for improvement so I set about getting a measure of waste. I selected a sample range of projects (about 25% of the previous year's work - at random) and trawled their management accounts, together with the project accountants. I also looked at the nonconformance reports for those projects and, where I found a report, tied it to the relevant costs. (Still with me?) My total for the past year came to around $25m was waste (e.g. purchased parts that had to be reworked due to faults, extra work needed that couldn't be invoiced, etc.). Since I'd only looked at a quarter of the work, my report to the General Manager (for the country unit) presented a figure of at least $25m but could rationally be projected to be $100m (or more as there was the possibility one of the project I hadn't examined was a doozy)!

    The annual turnover for that part of the business was around $500m pa so I added a reference to research that showed businesses that don't actively measure loss, and actively work to reduce it, waste between 20% and 30% of turnover. At 20%, the business was at the better end. However, there was no way the GM was going to tell TPTB that he was wasting $25m pa, let alone and estimated $100m. My report was quietly filed in the round cabinet. I was paid for the work and, as a contractor, had no personal interest in any associated bonuses. Mind you, I decided to move on shortly after.

    The GM? I checked back with some colleagues a year or two later and he had "moved on" - I believe to a division that was shortly wound up.

    Anonymous to protect the guilty (and me)!!!

  21. Admiral Grace Hopper

    Spectrum is green

    When you have a lovely big shiny display on which to put your customer dashboard, it may useful to have a script called, of, let's imagine, green.bat, that magically resets all the red alert-y things when you have a customer visit you operations centre. Apparently. Obv.

    1. lglethal Silver badge
      Trollface

      Re: Spectrum is green

      No, No, No. You have to have at least a couple of yellows, and maybe an orange thrown in as well. Otherwise the customer knows you're bullsh%tting!

  22. Anonymous Coward
    Anonymous Coward

    Many years ago, when working on a busy support team for a major utility company, we were all asked to record and report the number of phone calls we made and received each week. This information was already available as everything came through the support desk. But no, a senior manager thought it would be a good use of everyone's time to record and report another number every week. No detail, just the number of phone calls.

    We soon rustled up a vaguely technical looking 'call recording system' in MS Access where the team leader could select a team member from a drop-down list and generate a random number which she duly wrote down and submitted on a weekly basis for no-one to look at, ever.

    1. A.P. Veening Silver badge

      Re: Call numbers recording

      I would have requested a project number to write the time necessary to record the call numbers. And I probably would have written about an hour a day for that, one hour I could not spend answering calls.

      1. ICPurvis47
        FAIL

        Re: Call numbers recording

        When I was working as a delivery driver for a local pharmacy, Head Office decided that all drivers had to fill out several more forms each day to record mileage, time taken, number of drops, etc. every day. This resulted in every driver adding ½ hour per day for doing the paperwork. Manglement were not impressed, and decided it would be quicker to put the timesheets on the computer network. However, the lack of available computeer terminals, coupled with the complexity of the reporting system, meant that the ½ hour went up to almost an hour. I left before that conundrum was solved (if it ever was).

  23. Anonymous Coward
    Anonymous Coward

    Melons everywhere

    I work with a lot of major corporations and typically when I dig into their KPI's they are WaterMelons - Green on the outside and Red on the inside. Always impressed how the numbers are engineered to show a compliant status to the senior management whilst the users/customers get the full impact of the reality where things have failed or are sub-optimal.

    1. Notas Badoff

      Re: Melons everywhere

      Seems to me that ElReg needs to add an informational category alongside the treasury of measurements. This highly technical term, watermelon, deserves to be widely deseminated. "Is this a watermelon?" will cause many a manager to feel faint, yes?

    2. Anonymous Coward
      Anonymous Coward

      Re: Melons everywhere

      That's what happens when management wants to manage by a few simple numbers rather than digging in and actually understanding the inputs to those numbers and how they are changing over time. In other words, evidence of incompetent management who know they don't understand what those under them are doing so they ask for shorthand numbers.

      Then if their unit is failing, they can pin the blame on the people who came up with the overly rosy numbers they were given - basically by claiming their underlings were lying to them. Which, I'm sure is sometimes the case. Did the underling create those rosy numbers to hide his OWN incompetence, or because his boss didn't give him any direction how to construct the KPIs being asked for making it more a matter of ignorance?

      1. Doctor Syntax Silver badge

        Re: Melons everywhere

        Or because the boss didn't want to hear bad news. This is how the reality distortion field works.

        1. A.P. Veening Silver badge

          Re: Melons everywhere

          Or because the boss didn't want to hear bad news. This is how the reality distortion field works.

          The correct solution to this problem is to leak the bad news past your direct boss to his (and if necessary even further up). Unfortunately, it doesn't work if your direct boss is already C-level (in which case you are D-level and should urgently start looking for the nearest exit).

        2. Anonymous Coward
          Anonymous Coward

          Re: Melons everywhere

          If they boss didn't want to hear bad news, why order his underlings to measure KPIs at all? He can claim he "has his ear to the ground" and knows what is really going on, and people know most KPIs are of limited value.

          Managing based purely on P&L isn't necessarily a terrible thing, and a lot of managers succeed looking only at the "KPIs" from the accountants. They might miss potential early notification of problems that KPIs could have given them, but problems will show via increased cost or reduced sales eventually. Where it hurts is if they are trying to make cuts, or trying to decide where to invest, and don't have the knowledge to know where the money is most/least useful.

  24. PickledAardvark

    Not a KPI but daft metrics

    The marketing manager decided that all staff should have the intranet portal as their default home page for the standard web browser (Internet Exploder). A number of other managers reckoned it was a bad idea, so I was given the task of determining whether/how people switched their default home page after it had been imposed as a non-permanent change. Using logon scripts, I captured each user's URL for "day zero" (actually a few days after the imposed change) and for day 30ish afterwards. All I had to do next was send the data to the manager who had requested it.

    First attempt: Send the manager the raw data -- an anonymous CSV file with two columns (day zero, day thirty) and an Excel copy of the same. [Required Excel actions: Alphabetic sort and count.] A junior manager intercepted it and requested something simpler.

    Second attempt: Send the raw data and a summary Excel sheet, showing the number of repeated URLs (70% Google search home, unsurprisingly) and number of unique URLs. [Required Excel action: Convert to percentages.] Again, I was asked to simplify.

    Third attempt: The summary Excel sheet and a three paragraph summary that 10% of staff had stayed with the imposed change, 70% had switched to Google search, 10% had switched to a different intranet page and 10% had picked something unique. Pretty much the same analysis could have been performed by a crude sort and eyeball of the raw data...

  25. Sequin

    I worked for a large government department - let's call it the Bome Boffice, and we wrote a system to be used at ports and airports for one of the Immigration departments.

    After it was installed at one port, they asked us where they entered their correction factor. "What's that?" we asked. The reply came that it was a fudge they applied to all of their figures so they conformed to the averages across the country. "It's because we're a small port, and only a few lorries ship out from here" was their justification. We later found out that each office had a different fudge factor, for different reasons, just to keep the official figures in line!

    I think it was Rudyard Kipling who said (to paraphrase) that the Indian Government was the only one in the world that analysed their figures and stats in more ways and to greater depth than the UK government, but that when it gets right down to it, the data is entered in pokey offices in the back of beyond, by a lonely clerk who writes in what the hell he likes, so long as it doesn't make his bosses look bad.

    1. Anonymous Coward
      Anonymous Coward

      At least its an improvement over the previous fudge factor which assumed all immigrants arrived at Heathrow or Gatwick .... that's why for a couple of years after the expansion of the EU to eastern europe that the Labour government continually maintained that immigration from Poland etc was very low and people were exaggerating the numbers that they were seeing around .... until they eventually twiiged that eastern europeans flew in on cheap aitflights to regional airports, or came overland on coaches so the immigration figures had been wildly innaccurate. I think they then tried to improve counting but the ONS have finally acknowledged this year that the methodology that's been used to generate immigration figures for ever is so badly flawed that the figures can no longer be relied on for policy decisions.

  26. Sequin

    On a monthly basis I have to produce a huge report for one of our clients, which takes about 30 elapsed hours to compile, process and report the previous month's figures. This process was created by somebody who has since left the company, so it's not my fault!

    They asked us for a new report last year, and my boss told me that I could use the first report as a basis for the new report. This I did, but when they checked the results, I was told that the figures were nonsense. I then went back to my boss and told him that as the new figures were crap, the originals must be too. I was told in no uncertain terms to keep quite about it - They pay us a fee for it, and If they didn't complain, we wouldn't fix it!

    I have also said that I don't think anyone actually reads the original report - I once proposed replacing the contents with a message saying "If you can see this, please give me a call" to see what the response was. Again this was rejected.

  27. Waseem Alkurdi

    Also known as Passing the Buck™

    Known since before written history, but immortalized by the BOFH.

  28. Anonymous Coward
    Anonymous Coward

    We used to have a loophole in our call logging software where you could "steal a resolve" off of other analysts.

    This way it would make it look that you were resolving more calls that you were. All you had to do was run a search on a pending result of a closing call and force it to resolve early, it would replace the name of the initial resolver with the new resolver giving false stats.

    Luckily it was reported after it was being massively abused by some lazy staff, once it was patched over so no one could abuse the system it was quickly noticed who was shirking work and stat padding to get their yearly bonus.

  29. Anonymous Coward
    Anonymous Coward

    KPI's generated by a spreadsheet ..... reminds me of when I was a school govenor ~15 years ago where we got a clearly templated letter from our LEA (Bristol) which "congratulated" the school on its previous years SATS results and suggested that to aid the LEA improve its overall rating could we target increasing our results in the next year whch was followed by a list of their suggested targets which included increasing the Science SATS pass rate from 100% to 105%!

    1. A.P. Veening Silver badge
      Facepalm

      increasing the Science SATS pass rate from 100% to 105%!

      ====>

  30. IceC0ld

    you called him Alban, with THAT manager just call him Dilbert an be done with it, a classic PHB move if ever I heard one :o)

    1. A.P. Veening Silver badge

      Re; Calling him Dilbert

      There is a small problem with that as there is a copyright on that name, but they could have asked Scott Adams.

      EDIT:

      And we would have missed this post by jonathan keith.

  31. Anonymous Coward
    Anonymous Coward

    lies, damned lies and ... management reports

    Posting anonymously as my former colleagues may recognise this and/or still be working there.

    Back around the turn of the century I worked on a defence contract. The last week of each month involved a team of about a dozen people creating the MoD report [it was 2 weeks and even more for the MoD Quarterly report].

    Given this effort one of the directors brought in a management consultancy to try to improve matters [I know, I know.....]

    They set up a room with paper all on one wall,coloured pens,post-it notes -- the works. They started from the left side listing all the data that was collected and from the right side all the fields in the report. The aim was to see how the data was traced through and processed, with a view to optimising and possibly automating the process.

    Soon afterwards that room was always locked. The consultants ended their job. Talking to one of them over a pint it seems the big issue was the uncomfortable revelations that (a) a whole shedload of data was collected and never used at all and (b) some of the report outputs appeared to have no inputs (they just magically appeared - or, if you prefer, were just made up).

    Nothing much changed and, as far as I know, this practice continued (and may well continue still to this day).

  32. Jou (Mxyzptlk) Silver badge

    There would have been better manglament method!

    Correct the bug, but insert a correction to be always at least a 95%. To be removed during the next upgrade.

    That way they would have noticed where the bad areas are, and do some real work.

    But wait... That would be management and not manglament...

    1. Anonymous Coward
      Anonymous Coward

      Re: There would have been better manglament method!

      Better yet, have any "bad" areas corrected to 95.05 - 95.5%, just high enough to be green for upper manglement, but close enough to the line for lower manglement to want to fix it before it drifts down and makes them look bad.

  33. irrelevant

    Manglement Reports

    Ah yes...

    I remember, company I was working for was installing new computerised booking system in a council-owned leisure centre. The council wanted a particular report, some sort of analytical thing, which previously the staff used to produce via spreadsheet. Feed in the daily figures, out pops the report with some statistics on the end. We were given a few samples, but even in the original spreadsheet file, there was no calculation coded, just typed numbers in the relevant fields. So we asked, and were given the spec for the report, and coded it up accordingly. But when we ran it using the figures on the sample spreadsheets, we ended up with different values at the end compared to theirs. So we asked the local staff, and were told "Well, we just fiddle about with them until they look right!" Or look good, perhaps ... :)

    (We ended up leaving the report calculating the statistics as the council wanted, as they were paying the bills. I have no idea if the staff submitted that, or carried on using their spreadsheet..)

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like