How to make clouds and influence accountants

Slipping in between the beancounters and the CEO's PR bird


The cloud might mean that the corporate accountant becomes your new best buddy. Appalling thought, I know, but beancounters aren't all that bad: they can do sums even if they can't do algebra, which puts them a step ahead of the marketing department. Of course, this time of year, when bonuses are being decided, is a great time of year to go and explain to said corporate beancounter why you're about to become his best buddy.

The reason that you, that professional master of the company's computing, are going to have the pinstripes slavering over you is this cloud thing. For it moves the vast majority of the iron, the boxes, from being an "investment" to being a "current expense". There's even some likely-looking tax-dodging that can be done at the same time.

Apologies, but a little bit of corporate tax background is necessary here ... no, don't run away, it'll be simple. Roughly, and we'll stay simple, the things that the company uses are divided into two types. There's those things that are bought and sold: the actual goods, the stock.

In this class, for these purposes, we can include the labour – your and my wages for example – and we also put in there pretty much anything that's going to get used up in the year it's purchased. Cartridges for the printers for example, paper, a keyboard for the salesman who breaks one every three weeks.


When it comes time to do the sums about how the company has done (and no, I will not entertain corrections from accountants who feel this is too simple) we treat all of these things pretty much the same way. They are current expenses; subtract them from gross sales and there's our gross profit. We don't make much difference between that ink, the time spent by the bloke using the ink and the paper the ink is used upon: it's current stuff, all the same stuff.

Then there's all the interesting stuff that you lot do. Play with machines that are bought for use for two, three years, like the PCs. Write the code that might still be there in three decades, underlying some application that has been migrated through five or eight different sets of big boxes (just ask a COBOL bloke if there's still 30-year-old code running somewhere).

Now, we can't treat all of this as just this year's expenses. If we did then we'd show a huge great whacking loss every time we upgraded the “enterprise server” and then much larger profits than we are actually making for the four years we keep using that box.

Or rather, we could do this, but then the accounts wouldn't actually show us what we really want to know: how's the business doing? We might end up showing that investing for the future has just bankrupted us all, or alternatively that we're making a fortune but ignoring that we've got to replace the kit next year.

The way we deal with this is “amortisation” and I promise that's the only technical accounting word I'll use. So we go and buy a MacSoft server, which we think will work for four years, to run the CEO's new Twitter account (and keep him happily occupied with the new PR babe who actually writes it). We want to spread the cost of that box over the four years of its life, quite naturally.

So in this bit, where we write down cash movements, we show we bought it this year, and that the cash has gone. But over here, on this other bit of the accounts, we show the value of the box, minus in the first year 25 per cent of its value, 50 per cent in year two and so on.

We can get more sexy (no, accountants do think things like this are sexy) and use straight line, like that, descending balance (ie, 25 per cent of the value each year, starting with the value we inherited from the year before), make an estimate of final value and only amortise the cost minus final value. Pulses race at this sort of stuff you know!

The thing is, all of this is a right pain in the bum if truth be told. For each of the different things that the company buys to use over the years comes with a different life span. That blast furnace over there might be good for 50 years with a refit or two. That InfOracle system you've just migrated the corporate database might be good for maybe five years. PCs? Well, it depends on whether the apes in Sales get hold of one or not, really.

Which means that all of these different things need a different amortisation schedule and this is what those agonised meetings where they hardly touch the biscuits over in accounting are all about. Not what are the sums, but how do we allocate these various more difficult expenditures?

Just to add spice to it all, there are two entirely different sets of such rates. There's what the taxman will let you do (and the associated tax breaks he'll let you use) and what you want to report to the shareholders. For the former you'd like as low a profit as possible, for the latter the largest. Management accounts, as always, are the Third Way truth between the two.

What the cloud does here is simply remove all these arguments about such rates. For you're not in fact buying the big boxes any more, are you? You're renting time on them, sure, but that's easy to account for. We just don't have to worry about when the box is going to be worth nothing because we, the company, don't own the box. We also don't have to worry about the taxman because again, we don't own the box. It's a current expense, just like the CEO’s new juicy PR bird is, simples.

It's not just this that will please the beancounters though: by shoving everything off to someone else we've also drastically reduced the uncertainty faced as a result of any purchasing decision. When we decide to buy a big box, one of the things we worry about is whether the company will be there in five years to service it. Or even whether the OS will be: HP's just shown that you can't even rely upon a still extant company to keep one of them going.

Is Moore's Law going to speed up in the next few years? Certainly a worry if you, but not your competition, are going to buy your own processors this year. What are memory prices going to do? Do you buy to system limit now as they'll go up, or buy development amounts only now and restock when they've gone down and you're ready to deploy?

Who in buggery knows, really. And that's the other thing that the cloud does for you. It removes much of the uncertainty of how the underlying hardware of the systems is going to develop over the lifetime of your systems. You just don't care anymore, they're all someone else's problem, hidden in full view beside the pavilion at Lords. And yes, the beancounters do like the removal of uncertainty.

And finally, to a possible bit of tax-dodgery. Working out how to actually use the cloud effectively is going to mean quite a lot of research. Some development even. Stuff that can and should be done in-house as being better at this than the competition will indeed be a competitive edge. And you know what? There's a special tax allowance just for this R&D stuff. If you spend £100 on it it might be possible to knock £175 off the company tax bill: there's even a set of circumstances where HMRC will actually send back cash to the company.

Now if you can wangle that, not just reduce complexity and uncertainty, but get cash back off the taxman then you really will be the accountants' best buddy. They might even approve choccie biccies rather than just the plain ones at the next departmental meeting. ®

Sponsored: Minds Mastering Machines - Call for papers now open

Biting the hand that feeds IT © 1998–2018