Selling hardware on a pay-per-use or subscription model is a 'lie' created by marketing bods

Selling tech based on consumption is predicated on a "lie" and a "fallacy" created by the powerful marketing machines at some of the largest hardware vendors. So says Mike Norris, the straight-talking chief exec at Computacenter, Europe's largest reseller, who made the comments via a webcast interview broadcast at the Canalys …

[A Non e-mouse]

Of course all the vendors want "as a service" It keeps the money coming in when the end user would rather just stick with the five year old (or older) version of the software they were previously running. As to hardware, the pace of change/improvement has slowed over the past decade so there is less drive to upgrade. I've seen people running 10 year old network switches as they're "good enough". (Sure, there might be some security issues, but with a bit of skill you can block that off and the switch keeps running)

[Flak]

Spot on!

It is not just the vendors, but also their investors - businesses with long term recurring revenue streams are far more valuable than businesses that sell devices and have to 'run to stand still'.

The article is a bit too negative, as there are some real benefits for the customer:

- avoidance of capex (and potential debt/difficulty to get loans)

- avoidance of maintenance cost

- avoidance of obsolescence

- elasticity/flexibility up and down and in terms of contract terms

- immediacy of services (for cloud services at least) with no lead times

- outsource of infrastructure expertise

- outsource of hosting and connectivity

For some (types of) hardware I think you are right. Laptops and printers fall into that category.

[Roo]

Re: Spot on!

I think some negativity is inevitable with respect to renting hardware...

- avoidance of capex (and potential debt/difficulty to get loans)

- avoidance of maintenance cost

- avoidance of obsolescence

Not convinced the customer avoids any of that - as a money making business will pass those costs + margin onto the customer. Unless of course they sell at a loss (hint: vendors might do loss leaders - but they actually do need to make money at some point in order to survive).

- elasticity/flexibility up and down and in terms of contract terms

In principle yes, but it does require the vendor having a pile of idle kit hanging around burning money - they have a strong incentive not to do that. What actually seems to happen is that they have a ~3 month lead time on the kit getting put into their datacenters... Oh and good luck enforcing an SLA on that.

- immediacy of services (for cloud services at least) with no lead times

IME that can be true if you are lucky or you are a very small customer. ;)

- outsource of infrastructure expertise

Vendors have a tendency to skimp on expertise, choosing to keep costs to a minimum (as you'd expect) - their idea of adequate skill levels and acceptable SLA is somewhat different from the customer.

- outsource of hosting and connectivity

I think that point stands. :)

For the record I definitely see valid cases for renting kit - but I haven't seen it work out cheaper yet. :(

[Dave314159ggggdffsdds]

Re: Spot on!

It isn't necessarily cheaper, but when it is, it's usually the cost of sufficient expertise to make informed decisions in house which makes it so.

Cheaper isn't quite the same as more profitable, anyway. If I pay a little more to hire in a service, but it frees up my time to do extra jobs that bring in more than the additional cost, it's well worth it.

And do bear in mind that specialisation and economies of scale make things cheaper to buy in than doing them in house, generally. No one suggests offices should contain paper factories, for example.

[Roo]

Re: Spot on!

"And do bear in mind that specialisation and economies of scale make things cheaper to buy in than doing them in house, generally. No one suggests offices should contain paper factories, for example."

I agree with your point, but your strawman is a bit pants. Case in point it's generally cheaper to cook + eat at home than it is to go to a 3 star Michelin Restaurant. :)

[Mr.Nobody]

Re: Spot on!

Not sure the paper analogy fits here. Paper doesn't need troubleshooting when it is slow or broken. It certainly doesn't need upgrading or replacing.

[martynhare]

Re: Spot on!

Let us debunk these claims:

"avoidance of capex (and potential debt/difficulty to get loans)"

Capex for most IT hardware comes without potential debt. An employee either needs a workstation or they don't. It's that simple. Same for printers, where cheap and reliable B/W LaserJet printers win the day. If you can't afford the hardware, you definitely can't afford the employee in most cases.

"avoidance of maintenance cost"

Maintenance costs for most IT hardware is minimal. We are talking about parts replacement. It's technical support which still incurs the bulk of the costs and renting your hardware doesn't fix that because by and large, vendor support is inferior to your in-house (or even outsourced) IT support when it comes to general troubleshooting.

"avoidance of obsolescence"

OEMs like HP offer 5 years of next day on-site repair coverage as an option for both servers and workstations. Microsoft keeps server variants of Windows updated for 10 years while workstation variants now get lowest common denominator rolling updates "for as long as the OEM supports the computer". This means businesses in general need only set an alarm for 5 years time and then begin rolling out new hardware as and when appropriate.

Due to the coming end of Moore's Law, good quality hardware will now easily outlive the 10 year mark and continue to work with the latest Windows releases. That is why these vendors can get away with selling these products "as a service" in the first place. Esepcially with AMD processors, where vendors can keep the same motherboards in service for many years, while offering "the latest" business-class CPUs.

"elasticity/flexibility up and down and in terms of contract terms"

Existing cloud solutions which don't involve hardware-as-a-service can cover temporary excesses. There really is no need to rent physical hardware to achieve flexibility, when the hardware you're purchasing can easily last a decade.

"immediacy of services (for cloud services at least) with no lead times"

People can immediately evaluate most products/services. Many evaluations can be extended with a quick phone call to the sales department. It's common for evaluations to last 90 days or more for many services. Renting hardware doesn't offer immediacy unless it's IaaS which is outside the scope of this.

"outsource of infrastructure expertise"

Outsourcing your IT is a bad idea, outsourcing your infrastructure outside of your outsourced IT is even worse. That's self-flaggelation of the tallest order.

"outsource of hosting and connectivity"

That's outside of the scope of renting hardware from a vendor like Lenovo.

[Roo]

Re: Spot on!

"OEMs like HP offer 5 years of next day on-site repair coverage as an option for both servers and workstations"

Not convinced about the value proposition because the vendors don't seem to carry enough spares to actually fix anything - which is a problem when boxes are dropping like flies (5% failure rate within 12 months) and their suppliers retooled before the boxes were even shipped. In once instance the vendor (one of the pillars of the PC vendor establishment) shipped the machine with a slew of BIOS versions from the previous generation Xeon boxes - all of which had *different* configuration options. The boxes were 2 years old (and substantially reduced in number through failure) before the vendor managed to come up with a BIOS that supported the processors *and* a way to remotely update the BIOS.

It would be really nice if some vendors took advantage of the slow down in Moore's Law to have a box in production for 3+ years rather than a few months - so they have a realistic chance of characterizing the weaknesses of the product and creating enough spares to support them properly (and profitably) for another 3-5 or even 10 years.

[Roland6]

Re: Spot on!

Lets further debunk these claims...

"avoidance of capex (and potential debt/difficulty to get loans)"

If you don't take out and repay any loans, you won't build up a credit rating and so will find it hard to get loans...

The only reason (other than tax) to avoid capex, is to avoid having to pay out a lump sum and maintain it as an asset.

It should also be noted what happened in the late 1980's as companies rushed to convert assets such as property into opex, whilst they did liberate monies that could be invested or (more usually) returned to investors, they also significantly increased their opex, with a massive downward pressure on prices these companies found that they could not compete on price with (mostly) foreign companies... the companies that retained their property assets (ie. continued to own their factories) could leverage this asset to keep their opex low and thus continue to be price competitive...

Third-sector organisations that rely on grants tend to find it is more renumerative to purchase stuff outright as grant giving organisations tend to prefer to fund capex.

"avoidance of maintenance cost"

The only thing avoided is explicit maintenance costs - your subscription to say Microsoft 365 includes maintenance.

"avoidance of obsolescence"

I laughed at this one, a client has their 15+ year old fork lift's via a service contract, the service company has no interest in replacing the fork lifts with more modern ones...

Which really gets to the bottom of the xyx-as-a-service lie. Once customers are hitched to the service there is little real need and incentive to innovate at the pace we have seen in the IT industry these past 40 years. So in fact there is the potential to miss/avoid the 3~5 year refresh and 10~15 years down the road, discover the world has moved on but your xyz-as-a-service provider hasn't. Additionally, we have the converse, a business has to suffer the service providers upgrades delivered at the service providers convenience not the user organisation.

"elasticity/flexibility up and down and in terms of contract terms"

The benefits only really accrue if you have very dynamic or loads with uncertain scaling requirement - something that was common 10+ years back when businesses first opened web stores.

"immediacy of services (for cloud services at least) with no lead times"

The real value of cloud/xaas is lead time/time-to-market reduction (of the IT infrastructure) and the size of investment needed to trial something.

"outsource of infrastructure expertise"

Many business, particularly small businesses that don't run an IT organisation beyond a single person have already done this, so it isn't something special to xyz-as-a-service.

[veti]

Re: Spot on!

Most of those "advantages" are not really advantages to the business. They're advantages to some part or unit of the business that wants to do its own thing without meddling from central management, but not to the business as a whole.

That's why the service model is so popular with vendors: it lets them sell to lower-level managers, because opex budgets are typically much less tightly controlled than capex ones.

[Mr.Nobody]

Re: Spot on!

You want people to come up with a business case for why they are building an application?

[Grunchy]

Re: Spot on!

I personally don’t pay for any “software as service”, nor “hardware as service” either. I also actively campaign against all such schemes, too.

The only cloud services I subscribe to are free ones such as gmail.

Opex is nothing but a nightmare scenario which any video gamer has first-hand experience with: when the day comes to shut the service down it’s always for the vendor’s advantage, never for the subscribers.

Oh and that crack about avoiding maintenance and such? Not only does the vendor expect that the subscription fees pay for all of the capital cost and ongoing operating costs, he also wants a sizeable profit to boot, in addition to complete control including the power to shut down the service at any second (or else exercise his monopoly powers and jack rates up 10 fold).

I’m a gamer. All cloud schemes are bad. I know firsthand.

[Blackjack]

Software tends to be used longer than hardware. Windows XP machine finally died? Get me this program working on Windows 7. Doesn't work in Windows 7? There is XP mode. Need a XP program to run in Windows 10?

https://www.tenforums.com/tutorials/135551-hyper-v-add-windows-xp-mode-virtual-machine-windows-10-a.html

There is also Dosbox, Freedos on a VM, Windows 3.1 on Dosbox. A Windows 3.1 emulator for Windows 10, WINE all you want... this old software is here to stay.

[Charlie Clark]

Cashflow & tax rules

Vendors like subscription because it provides regular cash flow. It potentially also reduces competition and, thus, price wars once customers are locked in. Companies will generally go with what's most tax efficient and there is huge pressure to go from capex to opex for all manner of purchase. But they also like being able to outsource and offload the maintenance of devices: subscription can also include onsite replacements which don't need negotiating separately. You can see why third parties don't like this, because they get cut out. For customers the net effect may initially be small. But, as we can expect additional concentration in the industry, we'll no doubt see potential savings on the manufacturers side only and further concentration of capital.

The tax effect is to shift the write-offs to the manufacturers. This first started with the card industry in the early 2000s with leasing the preferred way of disposing of fleets of cars after that particular financial crash. And car manufacturers discovered the joy of regular cashflow combined with the ability to write off the excess vehicles it was producing.

[Jellied Eel]

Re: Cashflow & tax rules

But, as we can expect additional concentration in the industry, we'll no doubt see potential savings on the manufacturers side only and further concentration of capital.

Yup. Generally these changes don't seem to benefit the customers. I 'buy' tin, I pay X for annual support/maintenance and the costs are easy(ish) to understand. I might even decide that after 5 years, I don't want to pay support any more and just replace kit as it breaks.

What I don't like is to 'buy' tin, and at some random point in the future it goes EOS/EOL and stops working, forcing an upgrade. Or having to allow tin to phone home to some licensing server or it stops working. Thus adding security risks in exposing that tin, or just having to depend on connectivity to a 3rd party server which I don't control.

But a couple of examples of suckiness spring to mind. Way back, I was evaluating Alcatel kit for a new network. It was nice stuff, and had a very nice OSS. But the OSS also came with licence 'credits' for doing MACDs. The cost of the kit was comparable to other vendors, but the uncertainty around having to continually buy credits made the opex hard to model.. Plus additional headaches wrt purchasing those.

Or more recently, I've been musing about setting up a machinists school. Partly so I can play with CNC machines, but also because I think it's a GoodThing(tm) to train future makers of things. So I looked at AutoCad, and discovered that was 'sold' with annual licences, and cloud 'credits' for doing rendering or CFD in the cloud. It wasn't clear if some functions could be done locally, because compute is relatively cheap or if it was forced into their cloud at $1 a credit. Plus additional complications. So classroom licences could be cheaper educational versions, but production might need commercial licences, and licence challenges taking something produced in educational into production. But trying to figure out opex on that gave me a headache.

But such is 'progress'.

[Doctor Syntax]

Re: Cashflow & tax rules

"So I looked at AutoCad"

What was the eventual verdict?

[Jellied Eel]

Re: Cashflow & tax rules

What was the eventual verdict?

I'm still working on a business case. Main snag is I'm not a machinist, so having to make a lot of desktop assumptions. Main challenge is I want the training part to be low cost, but the startup and opex costs will likely be high to get premises, machines, classrooms etc up and running. So I've been looking around at potential for grants etc to defray some of those, plus whether it'd be viable to do some commercial work to offset the educational. I'm also a tad nervous about insurance and liability given kit like lathes & mills can easily eat the unwary or inattentive student. Or me..

[Doctor Syntax]

Re: Cashflow & tax rules

Best of luck with it. And if in doubt, let them eat the students.

[keithpeter]

Re: Cashflow & tax rules

@Jellied Eel

Find a seasoned workshop technician/instructor from your local further education college / general college. One with a decade or so of experience who is up for the challenge of setting up a workshop that runs to their rules.

https://steamhouse.org.uk/

These people (who I have no dealings with at all other than looking at the projects being completed each day as I walk past) add undergraduates to very big and sharp woodworking machines together with basic metal work with apparently no issues. Currently mothballed because of COVID.

Best of luck. Sounds a blast.

[Jellied Eel]

Re: Cashflow & tax rules

..add undergraduates to very big and sharp woodworking machines

Hmm.. not sure I'd phrase it that way in a prospectus. We'll introduce our students to a woodchipper.. :p

But thanks for the tip. That's the general idea once I've got a business case roughed out because I have very little knowledge of the practical side, other than IT stuff. I'm also keen to have it 'learning by doing' as much as possible, and to graduate, I'm thinking of challenges like making these-

https://www.youtube.com/watch?v=xw3GzqgWg-8

To test practical skils. Plus if there are trade certs they could pick up along the way, all for the good. I've firmly decided that the first thing taught would be getting students a first aider qualification :)

[Boris the Cockroach]

Re: Cashflow & tax rules

Quote:

"So I looked at AutoCad, and discovered that was 'sold' with annual licences, and cloud 'credits' for doing rendering or CFD in the cloud. "

We looked at that sort of thing recently, a very good idea someone thought until it was pointed out that we'd have to make sure the internet connection never dropped out, that where our data went was'nt exactly known (goodbye any government/military stuff we do) and that we'd be at the mercy of someone who was happy to low ball the price to lock us into their system,then jack the prices when their system was all we could use and our data was in their format.

So we stuck with what we've got now, a rival to autocad loaded onto the 4 laptops used by the setters with 4 licences(well 5 .. bulk discount on 5+)

[Jellied Eel]

Re: Cashflow & tax rules

We looked at that sort of thing recently, a very good idea someone thought until it was pointed out that we'd have to make sure the internet connection never dropped out, that where our data went was'nt exactly known (goodbye any government/military stuff we do)

Yup. I'd also be concerned about NDAs for general work, especially things like prototyping. Kind of hard to protect clients commercially sensitive data, if there's no control over how that's processed.. especially if any is archived. I can see the cloud being a tempting target for any industrial espionage, especially if you can snatch drawings and models handily labelled by client.

And although AutoDesk offers industry specific bundles, it doesn't look especially well integrated. That seems to be due to their borging of competitors. So as a start point I've been looking at the product design collection. Includes all sorts of useful looking stuff, but then..

Autodesk HSMXpress is a free SOLIDWORKS plug-in for basic milling, including 2D roughing and pocketing, drilling, facing, contouring, and more.

Neat, I'll take it. But is Solidworks, not AutoCad, and works with Fusion360, which is the cloudybollocks version. And last time I mused about doing this, Solidworks was an AutoCad competitor.. And my headache is coming back.

But this all started when I got roped into speaking to a room full of kids during a school careers event, and explaining product development. So I asked the students to put their 'smart' phones on their desks, and make them move without touching their phones. Which got quite a few kids thinking about the interface between the real & the virtual, and the joys of engineering.

[whitepines]

Re: Cashflow & tax rules

For the small amount of mechanical CAD our company does we actually settled on FreeCAD for the reasons outlined above. We tend to need to keep designs around for quite a while, and the subscription model meant that in addition to all the other risks there was a distinct risk of a 10 year old design simply not loading on the latest (and only) cloud version, forcing redesign / revalidation / recertification of the design.

[Jellied Eel]

Re: Cashflow & tax rules

Yes, and similar concerns over some of the machinery. I think CNC stuff is the closest thing to a replicator that we've come. Take a virtual object from a screen, send it to the machine, and watch a finished/near finished product appear. From some casual conversations with young people, they're generally similarly enthusiastic about the art of the possible. Or for me, I find YT channels like Clickspring's recreating the Antikythera Mechanism fascinating.

But I also want to make sure students can still run tools like lathes, mills etc, so aren't overly reliant on computers so they understand the basics before diving too deeply into CNC/automation. From scanning auction sites, there's usually a steady stream of machines, and from making some enquiries, that can often be shops closing because they can't find staff to run them.

[whitepines]

Re: Cashflow & tax rules

I think CNC stuff is the closest thing to a replicator that we've come.

Spot on. How is it not near magic to throw a concept into FreeCAD, run Slic3r, then have the physical object from a Marlin-driven 3D printer an hour or two later. And all without anyone else's cloud or other leased machinery* in the mix, making it pretty much an exact comparison to the Star Trek replicators in terms of personal liberty. Yes, it's plastic and usually some assembly required, but still this is unprecedented and useful capability compared to even 20 years ago.

* We actually run the 3D printers on OpenPower hardware (no Intel ME / AMD PSP), Linux, and Arduino (no locked loaders etc.) at the machine end, This was carefully chosen so that the apparatus will work until some key part breaks and for some reason can't be manufactured and replaced, versus until some license runs out or planned obsolescence hits.

[Dave314159ggggdffsdds]

Re: Cashflow & tax rules

You need a laser scanner as well. Slap a solid object in a scanning booth, press the big button, return later to find a duplicate.

[Jellied Eel]

Re: Cashflow & tax rules

This was carefully chosen so that the apparatus will work until some key part breaks and for some reason can't be manufactured and replaced, versus until some license runs out or planned obsolescence hits.

This is also part of the magic of skilled craftspeople. Things break, and get thrown away for want of a part. But give or take any legalities, creating parts is possible. Especially I guess if the original vendor is defunct, or has just decided old hardware is abandonware.

Neat example here-

https://www.youtube.com/watch?v=qzUMYc-bnmY

Where the dynamic duo of blacksmiths are bringing a vintage power hammer back to life. They have a neat motto of "Need a tool? Make a tool!".. But not sure I'd want a forge as part of the training. See previous comment about insurance/liability risks. Plus I have a healthy fear of lumps of red hot metal being waved around.. Or despite intending to be student #1, I skip that module. Being able to forge, heat treat and weld stuff would also be useful skills to teach.

[hoola]

Re: Cashflow & tax rules

This is also the only way they have of competing with "Cloud". Everything that is supplied as a service is based on monthly subscriptions. Buying real hardware is not unless it is leased of financed in some way. This makes those purchases stand out now so people ask questions and critically it is in boardrooms where various management droids will have bought into the "Cloud is best, cheapest always available" sales pitches.

Seeing large capital expenditure come though does not fit with these peoples understanding of how things work. Almost everything now, whether business or personal, is based on monthly subscriptions that in themselves appear cheaper. There is no escaping from this even if the capital purchase it cheaper. The level of scrutiny on capex is far greater as these jobsworths to ensure that the overall increased cost of subscriptions is kept under the radar.

[Roland6]

Re: Cashflow & tax rules

The tax effect is to shift the write-offs to the manufacturers. This first started with the card industry in the early 2000s with leasing the preferred way of disposing of fleets of cars after that particular financial crash.

Car manufacturers also discovered the enhanced revenues to be gained by selling the same car multiple times - a feature of the lease system. Not sure how applicable that is currently to (mainstream) computing.

And car manufacturers discovered the joy of regular cashflow combined with the ability to write off the excess vehicles it was producing.

They were already doing this in the 1990's, having many cars 'sold' to rental/leasing companies on very short leases (sub 6-months) then getting dealers to sell them as low mileage (nearly) new cars. I think it just took a little longer for the manufacturers to finesse the lease proposition to the general public.

Tesla, notably, seem to be the only manufacturer who have extended the lease model to the software installed in a car, separating it from the car itself, so the new owner has to purchase it all over again...

[ReadyKilowatt]

IBM until the 1990s

"Nobody ever got fired for buying IBM" should have been "Nobody ever got fired for leasing IBM." This is not new.

[bombastic bob]

Re: IBM until the 1990s

the only time leasing (over buying) makes sense is when the hardware is prohibitively expensive, like those old IBM mainframe and mini-computers (and the lease no doubt came with a service contract).

Modern computer hardware is generally inexpensive, doesn't really need constant maintenance, and lasts for YEARS if you're not addicted to to the "always upgrade" hype.

Might as well put a coin box on the side, like a payphone. [but the coin box and hopper would probably cost more than the computer, and you'd need someone to collect from it periodically]

Is THAT something they might consider? The extra cost associated with subscriptions and leases, vs the (alleged) benefit to customers?

I use a pre-paid cell. It costs me $100/year to do this, less than $10/mlonth. I always forget to re-charge it until after the minutes expire. I never use them all anyway. And it's the cheapest I know of, and I rarely use the phone anyway. It makes sense for ME to do a "per-usage" because of the obvious COST SAVINGS.

Would COST SAVINGS _ACTUALLY_ happen_ for customers for per-use charges on HARDWARE? I think it would be like MS wanting subscription based Windows OS, so that it would be a constant funneling of money from you to them...

[alain williams]

A big problem with opex ...

is that you need to be able to continue to service the on-going cost. This is especially painful when your income suddenly drops - maybe due to a certain virus. If you have bought the hardware outright then no one can take it away because you stop paying them. OK you do need to pay when something breaks, but that does not need to be who you bought it from (unless you bought a John Deere tractor).

Yes: capex means that you need to be able to stump up the cash when you buy it, this can be especially hard for a new/growing org.

[Graham 32]

Re: A big problem with opex ...

Depends how the licensing works. If it's paying to have access then yes you need to keep paying regardless. If it's paying for the usage then no.

Example: I pay my ISP a flat rate for access to the internet. In recent months I've been working from home a lot more, using a lot more data, but I pay the same amount. My mobile is on a pay-as-you-go model. I'm going out a lot less nowadays so using much less data and saving money. But if my landline fails and I have to use mobile data... ouch!

The real difference is that if you buy on capex you can make a decision to defer the next upgrade. eg replace laptops every 5 years instead of every 4 years. Instant saving for the buyer, instant penalty for the seller. There's no deferring on opex.

[Anonymous Coward]

Re: A big problem with opex ...

With hardware you usually pay just for having it inside your company - even if it's still in a box collecting dust. That's not cloud services when you're paying (dearly) for the time something is consuming resources.

Your ISP flat rate is probably low enough it doesn't make a big dent in your pay, so you don't usually risk to lose it if your income suddenly decreases. But think about people who rent an house, or even a room, and suddenly lose their income and also have to move away.

Would you like to be a professional who, for example, has a sudden health issue, and see his or her devices - with all their data - seized because of unpaid rates? Maybe while you can't care about the last backup?

[Dave314159ggggdffsdds]

Re: A big problem with opex ...

You say you can defer capex, but you haven't really thought it through fully. If the capex is justified at a given time, then you'd be taking a hit to profits instead of cashflow if you were to defer it.

[bombastic bob]

Re: A big problem with opex ...

this can be especially hard for a new/growing org

E-bay is full of used, reconditioned, and old stock computers, and I've never had trouble buying old hardware on e-bay [except maybe having to replace a power supply or hard drive earlier than usual]. But that's way less cost than a new machine. If you get inexpensive monitors locally, and the rest of the system in a marketplace venue like E-bay, you can avoid excessive 'capex' and get started on a much lower budget. Then, upgrade/repair/replace as needed later on... with minimal 'opex'.

(welcome to 'small business and startup' world, where I prefer to be anyway)

[DavCrav]

Capex to opex in a world of low interest rates

SaaS is already bad enough for the majority of instances. One can see that it's dreadful for the customer by the quoted stock price of Adobe. It only goes up that much because customers are paying a lot more. HaaS is insane unless you very rarely need the equipment.

I recently had to buy a machine for compute. The budget was about £6k, so not massive, but you can get something fairly decent for it. We were offered HaaS by one of our departments. A 5-year contract, two logins, some amount of data throughput, and it would cost more than buying a computer direct from Lenovo, which I can run as long as I want with as many users as I want. The previous machine is still going after 8 years, so their HaaS looked even more like a terrible deal.

Add into the mix low interest rates, which makes it much cheaper to move open to capex, and I don't see why anyone who isn't forced to would do it.

[Anonymous Coward]

Re: Capex to opex in a world of low interest rates

"I don't see why anyone who isn't forced to would do it."

I used to work at an IT consultancy where the IT director was Wined & Dined at least once a week by some traveling salesman.

Every so often our hardware and software portfolio was changed and in less than a year the preferred backup software we sold to customers were rotated between Tivoli, BExec, Veeam and Avamar (can't remember the order), and the same with hardware: IBM, Netapp, HPE and EMC.

No obvious reason was given, he just praised (to us) how the new ones brought more income and were somohow better than the old stuff. A world class bullshitter, adept at talking himself out of every sticky situation whenever he was questioned by the not-very-technical bosses.

[DavCrav]

Re: Capex to opex in a world of low interest rates

OK, I was wrong, so let me rephrase.

"I don't see why anyone who isn't forced or bribed to would do it."

[Cuddles]

Follow the money

Vendors want to sell subcriptions because they think it will make them more money. It follows that everyone else should avoid them as much as possible because the entire point is that it will cost you more money.

[nintendoeats]

Sod IT

As a consumer, if a fixed-cost item says "as a subscription", I walk the other way.

I mean, ok. If Photoshop was sold as a subscription dirt-cheap, that might be ok because it provides a means of staying up-to-date for the same price as buying every few years. Hell, if you use a lot of creative cloud it's a really good deal. But if it were cheap, Adobe wouldn't be making bank. So nuts to them.

The same applies to almost every subscription-for-fixed-cost-good I have ever seen. It's only cheaper if you were already on the upgrade treadmill.

No, I'm not bitter. Just learning Krita 0_o

[Doctor Syntax]

Re: Sod IT

I suppose if you only need Photoshop occasionally the pay as you go option works out cheaper. But so does Krita.

[Dave314159ggggdffsdds]

Re: Sod IT

If you only need it occasionally the pixlr.com is probably going to have the features you need.

[Bruce Ordway]

Depends?

I personally hate "as a service" but... I have recommended it.... to some sites.

These are smaller outfits that (often don't even know they ) struggle to maintain a capable workforce to run their current IT systems.

e.g. site running that 15 year old ERP system and they only have one (usually old) employee who has any idea what to do if anything goes wrong.

So my advice to those outfits is that the next time they think about an upgrade.... also look at "as a service" options.

This is purely from a disaster recovery angle other than anything else.

[Doctor Syntax]

Re: Depends?

So they get rid of that employee when they go to ...S. Now they have nobody who knows what to do when something goes wrong.

[Charlie Clark]

Re: Depends?

But I think that's the point: what's the company supposed when the employee leaves? Renting the hardware/software can be an option where the company does not have the necessary skills. But the risks of benefitting from an outsourcers economies of scale need to be offset against the risk of lock-in; or, and this is more common decline in quality of service as the outsourcer continues to try and save money by cutting costs and replacing trained local specialists with cheaper ones in Elbonia.

[Bruce Ordway]

Re: Depends?

>>But I think that's the point: what's the company supposed when the employee leaves

Yes, this is what I have been seeing.. more than companies getting rid of people.

Where ( I think ) personnel just tend to move on - especially SMB's.

Right off the top of my head I know of several sites where either IT personnel has voluntarily moved on and/or senior management/company ownership has changed.

i.e. normal evolution that is leading to vulnerability... and it's low on their awareness scale.

[Anonymous Coward]

Re: Depends?

I have a client that has this exact problem. They employ thousands of people. Only 3 of them know how to keep their ropey old ERP chugging along. It's so customised no external vendors could take over. The software vendor has long since deprecated the product. If the 3 people leave, the company is screwed...so keep throwing money at them at bonus time...

[Doctor Syntax]

Re: Depends?

Oh, I get the point, OK. But if ERP, on-prem or in cloud, is at the core of what the business does there needs to be some expertise in-house or maybe on-call. Maybe the OP performs the latter function as I did for a small engineers' supplier. The same applies to any sort of business which depends on IT.

[Donn Bly]

CapEx vs OpEx

I personally prefer to reduce operational expenses through strategic capital expenditures, however, there are often very good reasons for the reverse - at least in the eyes of the beancounters (and shareholders) - and it comes down to taxes.

If I buy a machine as a capital expense I have to pay property taxes on it every year until it is amortized off of the books, and often it is on the books longer than it has a useful life.

If I expense the equipment as a service, not only do I avoid property taxes on it but I can use pre-tax money to pay the bill. Couple that with not having to pay taxes on it after its useful life and the total cost of ownership may actually be lower.

The loser is the local government in the form of property tax income (since the company "servicing" me pays the property tax in their jurisdiction and not mine) but the government is the one who makes the rules, the rest of us just have to play by them.

[T. F. M. Reader]

Re: CapEx vs OpEx

@Donn Bly: I don't think the tax question is that easy, though I do think (at least) some parts of your argument have merit, though a bit of a refinement may be in order.

The fees you pay for the "service" certainly include property and other taxes that the service provider pays. That's the not so easy part.

You still may be absolutely right on that for one or more of the following reasons:

1) The service provider may deploy their property in a place with an advantageous tax regime, whereas you cannot.

2) The service provider, being a big whale, may have negotiated particularly advantageous tax terms with whatever state or local government rules the place where they pay taxes. You cannot do that.

3) The service provider will typically "multitenant" the services, thus reducing the property taxes on HW per customer. You cannot do that with your HW, unless you multitenant different activities in your organization on the same HW and save all of them the service costs. That would be difficult in most cases, I imagine, and less flexible/efficient.

Ultimately, the service provider may pass a part of the savings to you, so your would pay less in service fees than you would pay in CAPEX+taxes. So, you are probably right, but it is only due to the economies of scale available to the service provider, not so much due to some intrinsic advantage of OPEX.

[Insert sadsack pun here]

Re: CapEx vs OpEx

"If I buy a machine as a capital expense I have to pay property taxes on it every year..."

What are you talking about, a property tax on IT products?

[Robert Carnegie]

Good example.

Some householders were sold a solar panel deal where they don't actually own the solar panels, they are sort of leasing them... and when they don't get as much useful and/or re-sellable electricity as they expected to, it isn't the provider's problem. In other words, someone has managed to make free energy unaffordable.

[Lorribot]

The whole Capex v Opex thing in driven by Finance departements that live within a year and just want simplicity. Opex offers that. Capex needs the business to do stuff and is unpredictable (but not really), comes and goes and is planed by IT and ignored by business (no maybe next year's budget, what do you mean you said that for the last 3 years and there is no support any more why didn't you say that last year....).

Companies are not geared up to fund IT to the degree it needs to be funded and not prepared to make the business comprimises ito fit the level of its funding. Opex ends up being the best comprimise.

If finance could think out the box and apply different metrics other tha quarterly/yearly results the world would be better for it.

[ortunk]

Kill All Beancounters then?

[Wellyboot]

A lot of effort needed there, Just a few in every organisation would encourage the others to think differently.

[DavCrav]

"Kill All Beancounters then?"

Pour a lorryload of beans on them. I hope they would appreciate the irony.

[TheWeetabix]

Knowing accountants....

I think you will find beans depreciate once taken out of the tin.

[Doctor Syntax]

"Kill All Beancounters then?"

Just can them.

[Flatlander]

No subscriptions here

Which is why being a small business we still use Adobe CS6 Suite.

[Martin 47]

I wonder whatever happened to Radio Rentals?

[Peter Gathercole]

Radio Rentals!

Radio Rental's business model originally worked because radios, TVs and video players were expensive, and people could pay small amounts of money regularly over a long time to have what they could not afford to buy. In additional, valve TVs and early video players went wrong often, and the rental model also provided maintenance built in. People were happy renting things they could not afford to buy.

Their business model broke down when things that were expensive became cheap, these cheap devices didn't need constant maintenance, and people could borrow money and eventually own what they wanted rather than paying a similar amount and never owning it.

RR tried to diversify into white goods, but the same thing happened, and eventually there was no money in it and they shut down.

We've seen this in the computer world. Computers were expensive, and they were leased, rented or timeshared. They became cheap and small and companies could buy them outright.

Computing appears to be going the other way now. But as an analogy, so are cars (PCP), phones (plans paying for calls with the phone part of the deal), and even media (think, we don't buy CDs and DVDs any more, now we have Spotify, Netflix Amazon Prime et. al.)

Some of us don't want to own things. So why is it a problem for companies to want the same!

[Anonymous Coward]

Re: Radio Rentals!

> Computing appears to be going the other way now. But as an analogy, so are cars (PCP), phones (plans paying for calls with the phone part of the deal), and even media (think, we don't buy CDs and DVDs any more, now we have Spotify, Netflix Amazon Prime et. al.)

I don't know why you got the downvotes. It's clearly A Thing that a substantial number of people would rather pay £10 per month for music, than build up a collection of recordings that they can enjoy in their retirement without continuing to pay for the rest of their lives.

There is a convenience aspect at play: you don't have to drop a CD or DVD into a player, and in particular you don't have to wait for all those tedious DVD copyright warnings that you can't skip over.

There's also some logic in the car leasing model, *if* you're the sort of person who upgrades your car every 3 years. The same applies to phones, *if* you're the sort of person who upgrades every 2 years.

The imperative to do this for servers is rather dubious. A new server will have *slightly* higher performance and/or lower power consumption than a three-year-old one, but that by itself is unlikely to justify the cost of a new server.

[Doctor Syntax]

Emigrated to Oz.

[chuBb.]

Glad about this

"cloud computing has grown "hand in hand" with cloud computing"

I would be surprised if it didn't...

[Dave314159ggggdffsdds]

This is completely counter to the business trends in all sectors, globally, of the last few decades. Buy in the stuff that's ancillary to your core business so you can concentrate your capex in the bit you specialise in and know enough about to get right. Only do what only you can do, and leave the rest to external specialists.

That's quite apart from computer hardware depreciating so fast it ought to be considered a consumable rather than an asset.

[DavCrav]

"That's quite apart from computer hardware depreciating so fast it ought to be considered a consumable rather than an asset."

UK research councils classify computer hardware costing less than £10k to be a consumable.

[Lorribot]

Depreciating assets are just a idea manufactured by accountants in order get tax breaks and do other stufff that makes no difference to the day to day running of the business. IT stuff are not assets it something you bought to do something till it no longer adequately serves that purpose and gets replaced, by which time there is a very tiny number of people who may have a use for it, but are unlikely to give you anything for it or may even charge to come and get it.

Computer hardware is a consumable. Other things dictate if they continue in use such as OS/Hypervisor support etc. teh business and accountants don't see any of this or understand it until stuff breaks big time and there is no support.

[Doctor Syntax]

"Buy in the stuff that's ancillary to your core business"

This assumes you're capable of working out what's core and what's ancillary. Do you need IT to run production? To take orders? To move product off the shelves? Is IT core or just ancillary?

[Anonymous Coward]

A major downside from my viewpoint is a loss of autonomy. Renting always puts your business at the mercy of the property owner's whims.

[John Smith 19]

I own your data. I own *you*.

You pay what I tell you to pay or I cut off access.

You don't have the bandwidth to download all that data and you can't collect it because I will not tell you where it is even stored

I think Kevin Spacey in Horrible Bosses put it quite succinctly.

"You are my b**ch"

Or to mis-quote Benjamin Franklin "They who give up essential (data) freedom to purchase temporarily lower costs deserve neither freedom nor lower costs"

Any business leader who doesn't get this is no leader, just a follower of fashions with no understanding of the issues.

[martinusher]

Old As The Hills

The first person to figure that renting data processing equipment was more lucrative than selling it was Herman Hollerith who coincidentally was the first person to market punched card tabulating equipment. Rental or leasing models for IT equipment were pretty much the rule untilt he 80s when the price of computing equipment plummeted to the point where it effectively became consumer equipment.

The trend is reversing because of the same logic that Hollerith used. If he sold his tabulators to the Census Bureau then its a one-off transaction. If he rented them then he could sell exactly the same equipment to them twice or more. The cashflow model is enticing. These days there's an added twist in that users are becoming increasingly granular -- with the trend towards classifying employees as individual contractors (even if just for internal accounting purposes) its possible to enscapulate all their expenses along wiht their renumeration and figure out if that particular work unit is yielding enough profit. (It also keeps other work units in employment, continually measuring and verifying the productivity of others, an entire vortex of work that doesn't necessarily produce anything but certainly creates a lot of employment.)

[bigtreeman]

Re: Old As The Hills

Hollerith machines were used by the Third Reich for the Holocaust.

So that means IBM owned and profited handsomely from the genocide of the Jews.

[bpfh]

Full circle?

“Selling tech based on consumption is predicated on a "lie" and a "fallacy" created by the powerful marketing machines at some of the largest hardware vendors.”

That’s how Big Iron used to be sold. Plenty of stories about how you had to buy credits from HP to unlock CPU’s for your end of months accounting, or Big Blue field engineers reading books in a server room for 4 hours after flipping a couple of DIP switches to make it look like mainframe maintenance was hard - and expensive...

[MachDiamond]

What you don't own, you don't control

"and to move fixed costs to variable costs, which is critically important in today's fast-moving business environment."

For today's "fast moving business environment" (buzz words all), if your tools shut off because they can't verify the rent has been paid or the company that rented the gear is updating and will be coming by to collect what you have and drop off the new stuff, you are dead in the water. When I shift computers, it's often a whole day or even a weekend to transfer everything over and get it all connected back to license files. Often the upgrade will be coming with a new OS version and some of my applications and drivers will need updates (or I'll have to get a new peripheral). For these reasons, I buy good quality computers so they will last a good long time. With a rental, I may not be allowed to break a seal to clean, troubleshoot or upgrade without a big penalty.

If you need a supercomputer to run some complex CFD analysis, that's a good place to have computing as a service and there are a bunch of firms that do that. For day to day operations, if you can't afford a desktop/laptop computer, the company is underfunded or not in a position to add another employee and keep with with paying them and taxes. This is like companies that need salaried staff to put in 60 hours/week to make the business model work. Bonkers.

[bigtreeman]

fashion

Rent or buy ?

In house or outsource ?

So The Reg is just an advertising machine for cloud providers.

Questions, questions, questions flooding into the minds of today's young executives.

Will I make the decision that experienced, old executives know is bullshit.

I choose fashion and I'm hypnotised by that delicious sales pitch.

[Valheru]

Re: fashion

There is an insightful haiku hiding in there somewhere...

[Anonymous Coward]

Depends on the size

Consumption model is around since ever, mobile phones with the subscription, cable boxes, internet routers, etc.

Now the real issue here is about cash flow either from individual or a company.

With the trend of home office average people is now obliged to get more equipment to be able to work from home, equipment that otherwise they would not. While pandemic is a catalyst, most companies will keep the model no matter what.

Now companies that is an all new level of problem with cash flow. The problem is that a lot of companies need a head start in CAPEX to provide services, and will get return in years. Now again is about cash. Then there are some other issue, particularly in Europe, VAT. If you buy 1M worth of equipment in CAPEX, on top you will have to get in average additional 200k out of your bank. Now imagine you can for the same estate pay 50k a month even if you add up the VAT.

But this is just the tip of the iceberg, most of the models aren’t just about the hardware, is about managing the logistics. I have to deal with a large estate and massive refresh programs. Lets say I’m doing a refresh’s of 400 laptops, in average they cost 1200 usd x 400 + VAT. Do the maths, but I will have to inbound all the assets, put them in some financial system, install, store, distribute, and by the time I have completed the process I already lost several months in support, likely and all the real estate cost of storing and processing. Most companies offering asset management they will also install the laptop for you and send it to the user and you will start to pay only at that time.

Then there is the issue of claiming back the assets, a lot of the services not they will collect and dispose the equipment, often even with programs of reselling in secondary market and they credit you part of the profit for future purchases. Most companies would have to pay someone to dispose the asset. Anyone running an IT department will tell you that disposals are a big problem, and companies end up with piles of junk all over IT departments.

Another thing people disregard is the complexity of dealing with assets from the financial prespective. They need to be recorded, depreciated and disposed. This is a huge problem for CFOs as well, I regularly have to evidence for financial audits where are the assets, even for other things people forget, like insurances.

In summary these open models that are having success aren’t just about the financial side, it is one of the most important ones, but is about the general asset management life cycle and at scale, this makes a huge difference in all aspects I mentioned and many others I did not. I moved laptops and other assets to this model and the company saved a lot of money in thing you wouldn’t think off at the first look. And remember companies end up refreshing hardware every 3 to 4 years in average, in my company I do not see that this model have changed our patern of consumption, just made my life much easier and my business it cost more predictable and easy to allocate to departments.

[ForthIsNotDead]

Fallacy is correct

The whole model doesn't really work with projects based companies. I work in SCADA/telemetry, most of it CNI. If we take on a project to move your SCADA and telemetry based data gathering systems 'into the cloud' (i.e. to somebody else's hardware - don't get me started) it will be an OPEX cost for us while we're developing it, but then when we hand it over, it's a CAPEX cost for you! You've gained nothing.

[John Smith 19]

...AsAService. Looks cheaper.

Isn't.

"Outsourcing" for the 2020's.

Big fun.

[matjaggard]

I don't disagree but from you?!

I don't disagree with anything in the article but of course a hardware reseller thinks that buying hardware is important, isn't this article just an advertorial?

[Anonymous Coward]

I think you're all missing the point

For the vendors it's about two things.

1.Customer retention, "as a service" doesn't pop it's head up and ask to get three prices or go to tender, it just ticks along and is easier/cheaper to stay with the same one.

2 Salespeople are a huge expense, and on this model they are selling themselves out of a job, e.g. they don't get paid on ongoing revenues from O365 forever.