Budgetary myopia and the hand-me-down laptop

Splitting atoms or roasting beans

There’s a great line in Winning with Software where the author’s trying to emphasize software’s critical importance to almost every business out there today. This importance holds up irrelevant of the business’s core function. It doesn’t even matter whether anyone in your organisation has ever written a single line of code; you could be roasting coffee or you could be designing light-water reactors—either way, you will live or die by your systems.

More succinctly than I could, he puts it like this:

“Every business is a software business.”

I find this to be overwhelmingly true. An enormous fraction of businesses are now largely driven by their software and systems. It’s a point that I’ve found lost on decision makers, who often seem to think of IT equipment and software spend as an unavoidable evil; a pesky and persistent demand for resources that could be better used buying raw widget material or funding a few more trade show visits.

Costs and benefits

The problem with the above position is that good tools are not some unrecoverable sunk cost. They’re an investment—one that can offer an incredibly high return. From my personal experience, there are two stand-out examples of this position that have stuck in my mind ever since. Both are actually more about computing hardware than software, but I think the culture of these things is the same.

Productivity can be cheap

The first of these unfolded about a decade ago, when it was decided that the next year’s intake of new graduate engineers would no longer need two monitors, and that one would suffice. On the face it, this might make sense. You could hire thirty new grads in a year, so that’s thirty monitors you don’t have to buy. At a few hundred quid each, that’s an easy £10k saved.

However, the average loaded cost of each of these hires was probably in the region of £50k, plus whatever their software licenses cost—which could easily have you pushing a £100k per person total. I would think of £100k as a pretty significant investment, so I’d want to make sure that my new hires were as productive as possible. Given that research suggests a second monitor can increase productivity by as much as 42%¹, I still find it astonishing that this policy was ever even considered, never mind actually implemented. You don’t even need to do any sums to reveal that a one-off £300 spend that lets you wring 42% more out of a recurring £100k cost would be a good use of the funds; it’s a complete no-brainer.

“That would be 1% of our budget”

The second happened a few years later, and was thankfully much more understandable—but still offered some interesting insight into the decision making behind things. I was working for a much more cash-strapped organisation, at least by the industry’s standards, and a significant chunk of the engineering function were using crusty old laptops that were nearing the end of their service life†.

In my case, I was working on something that involved running a lot of simulations and my machine was struggling to keep up. The equipment was costing me a pretty significant amount of time, and I was concerned that this would have knock-on effects for the people who were waiting for the results of this work. When I raised the fact that the kit was becoming a pretty significant limitation, this was run up the chain and a pragmatic response came back. It was along the lines of:

“We can’t afford it right now, but we understand the issue and we’ll get to you with new equipment as soon as we can. What you have to remember is, if we replaced your kit, we’d have to replace everyone else’s as well, and that would be 1% of our budget.”

I have a high opinion of how this org was run while I was there, and I think this was handled in a completely reasonable way—but that 1% bit has always stuck with me.

In the time that I was with this organisation, I was usually at my desk in front of a computer for anywhere between 50 and 80 hours a week. The same applies to the dozens of other engineers who I worked alongside. For each of us, almost every single thing that we ever did involved some interaction with a computer:

• You have a new concept idea? Fire up Visio or PowerPoint, draw up a schematic, then e-mail it to someone.
• You need to do some calculations? Open up Excel or MATLAB, run some numbers, then write up your results in Word and send them out.
• Ready to work up a detailed design? Wait for CATIA to load, then get modelling.
• Had some queries about stress concentrations? Abaqus.
• Part design finalised? hyperMILL.
• Can’t find the information you need? Google.

This list goes on. The point is our computers are now our first step to almost everything that we do. They’re the tools we use the most often, and they’re the tools that we use to do the most important things. If you’re spending 50-80 hours a week in front of a computer, then having those tools be as good as they can be is almost always going to be a good investment.

Ultimately, it seemed to me then, as it seems to me now, that spending 1% of your budget on those tools every single year would be a wholly rational use of that cash.

If you’re going to cut down trees, you should spend a little bit of money to keep your axe sharp.

†I should say that I absolutely don’t have a problem with this. I’ve bought second hand equipment for myself and for Anneal employees in the past, and I’m happy to run almost any machine into the ground before replacing it. It’s only a problem when it becomes a problem; when your workload exceeds what the machine can handle, and when it starts to hold things up.

Budgets and boots

Having mulled this phenomenon over for a while, I’ve wound up thinking of it as budgetary myopia. I suspect the thinking is along the lines of: “You can still run your CAD system with one display, and you can still run lap simulations on that laptop—so you don’t need anything else. What’s wrong with the stuff you’ve got?”

With a short term focus, this all stacks up. You could save a little bit of money by not buying the tools. However, zoom out a bit and the false economy becomes apparent. The tools are the things that let you do the work, and that let your business function. By choosing to save that £10k, you could be costing yourself orders of magnitude more—by missing deadlines, losing races, or failing to get your product to market before a competitor.

It’s a bit like a self-imposed version of the Sam Vimes Boots theory of socioeconomic unfairness². You can save a bit of money by not buying the tools, but you’re going to pay for it in the long run. You’ll pay for it in time, in quality, in lost opportunities, and in the frustration of your staff.

Drawing boards, draughting film, and rose-tinted glasses

I’ve also spent a lot of time trying to work out why some organisations have this culture. With respect to engineering businesses, I think part of it is that the people who make these decisions are often a bit removed from the coal face, so they don’t feel the pain first-hand.

I also think things have changed a lot in the last few decades. When I started work as an engineer about ten years ago, many of my managers had started out working with drawing boards and using draughting film. In one case, I remember a manager telling me that they still drew on linen when he’d started at British Aerospace. There wasn’t a computer on every desk, nor an enormous software bill, never mind a second or third display for every person… yet these organisations still produced incredible designs.

My suspicion is that these guys look back on simpler times with rose-tinted glasses, with an internal monologue something like:

We didn’t need any of this shit in my day, and we designed Concorde. Let’s use the money for something else.

It’s a pretty understandable position, and I certainly envy those who got to do hand calcs in a smoke-filled room without ever having to unpick somebody else’s Excel workbooks. But, for many good reasons, we’ve structured our engineering functions such that CAD, CFD, FEA, MBS et. al. are now critical parts of our workflow. Products today are more complex than ever before, with higher demands for both performance and reliability—so we need all the help we can get.

However, these systems have a serious and growing demand for compute, and they churn out data by the bucket load. One way or another, you’ll need better tools to manage that—and the rose-tinted retrospection won’t help

You need a sharp axe.

References