I know Behavioural Science is a bit of a buzzword at the moment (can you have a buzzword made out of two words…? I’m not clear on that, but that isn’t my point, so we’ll press on…).
The concept behind Behavioural Science, to horribly over-simplify a complex topic, is that understanding how humans behave in the real world can help businesses understand their needs better and thereby get them to do something the business would like them to do, whether as customers or employees.
In marketing and advertising, people like Rory Sutherland and Richard Shotton showcase how the way we talk about products and services can make it more or less likely that someone is going to buy the products a business is selling. The differences are often subtle, but with out-sized returns if you make the right choice.
My pal Christian Hunt has done a tremendous job implementing the concepts behind behavioural science into the seemingly unrelated field of compliance – but it turns out that skills of a marketer or advertiser are very similar to the skills an organisation needs to get people to follow the rules, especially in areas with significant regulatory oversight, like in financial services, for example.
Yet behavioural science is not a concept I hear people talking about in accounting all that often.
To be fair, that’s partly because accountants have a (generally well-deserved) reputation of not engaging with any airy-fairy hippy nonsense that you can’t capture on a spreadsheet.
But I think that’s a mistake because understanding human behaviour is a shortcut to much better, bottom-line-boosting business decisions.
And if your CFO isn’t wanting to make a lot more of those, frankly I don’t know what else they think they ought to be concerned about.
Traditional financial analysis
Somehow, there’s an expectation – both of accountants by non-accountants, and of accountants themselves – that to make any important decision you need a 90 day-long project to extract mountains of data which someone is going to produce enough charts and tables from to explain every possible outcome and variation.
Let’s be up-front about this.
That’s always an expensive process. However well-intentioned it might be, tying up senior people in weeks or months of meetings, analysing spreadsheets, charts and graphs until the numbers give up their secrets is horrifyingly expensive.
Sometimes people tell me that doesn’t matter because these people are paid a salary “so it doesn’t cost any more to get this work done”.
When people tell me that, I know they’re either not an accountant or, if they are an accountant, they’re a particularly third-rate one.
Internal projects are a huge time suck…and therefore cost fortunes, although generally nobody joins the dots to work that out.
And they often lead to very equivocal conclusions – “well maybe this might work, but equally it might not”.
Frankly, you don’t need half a dozen very well-paid people spending 90 days together to work that out. I could probably have come to that conclusion myself with 10 minutes thinking time and a calculator.
If you didn’t do the analysis project at all, and fired half your finance department because they were no longer needed, you’d make a guaranteed bottom line return on Day 1. And your business would probably be no worse off because the project’s conclusion was only ever going to be “meh” anyway.
So what do you do?
Now, at some point in the process, you do need to do some thorough analysis. I’m not suggesting for a moment that you commit $ millions to an idea someone had in a fever dream without checking it out thoroughly first.
However, my suggestion is that you don’t knee-jerk your way into an analytics project at the start.
I kick things off with two other tools long before any serious analysis takes place.
The first of these, I call “a rough cut”.
That really is me spending 10 minutes with a calculator and a bit of paper (I wasn’t joking about that).
Doing rough cuts is not the point of this article – we might do that another time. But, briefly, I’m trying to work out “what’s the prize?” in this stage. Put another way, is this project likely to be worth the time and effort required to explore it further?
You’d be amazed at the number of projects which have floated across my desk over the years which can’t even meet that hurdle rate.
In my home life, I once had a salesperson try to convince me to switch my utility provider because they could save me something like £2.73 a year compared to what I was paying now. There is almost no amount of work I’m prepared to put into something that costs less than a cup of coffee in Costa, and spending an hour on a phone with the utility company’s salesperson didn’t sound like the best use of time to me.
At work, I’ve often had people try to flog some system or other which requires me to spend £1million up-front in return for a £50k-£100k annual cost saving, after factoring in the cost of the system.
But pause for a moment and cost in the disruption, training, and general inefficiencies in implementing any new system…together with the fact that you can be pretty sure that anyone flogging a system to do anything will, at most, deliver half the benefits the salesperson claims…and that project is a breakeven project at best. Certainly not one I’m going to commit hundreds of hours of highly-paid people’s time to analyse in excruciating detail.
There just isn’t enough upside in it to make it viable.
Behavioural accounting
To be fair, this is a term I’ve probably made up. (If it is, I hereby claim the copyright. If it isn’t, the expression is the property of its current owners.)
But what this is about is looking at what people actually do, and extrapolating a decision from there.
Now, it has to be what people actually do. That’s really important.
It’s not what they say – whether that’s people answering a survey or a business claiming to have thousands of happy customers.
It has to be what they do.
And you’ve got to be especially careful when the proposal in front of you suggests a course of action which is not consistent with what people actually do. The cost of behavioural change on an organisation-wide level is prohibitively expensive and while I wouldn’t say there is never a business case for it, the number of times you are likely to get a positive RoI on an investment of that kind is no more than once or twice in the course of your entire career.
Looking at what people actually do is a great guide to the decisions you should take – and, often more importantly, the decisions you should avoid.
It’s been said the eyes are the windows to the soul.
Well, the decisions people make are the windows to their soul too.
Perhaps a couple of examples would help…
1-Taxis
If I was in the market for a relatively inexpensive, good value, reliable, easy and cheap to repair sort of car, I could conduct an extensive piece of research into every model offered by every manufacturer on the market, compile huge cross-model comparison spreadsheets, and go on dozens of test drives.
Or…I could just look at what taxi drivers choose.
If there is one group of people who are optimising for exactly the attributes I’m looking for in a car, it’s taxi drivers. (Not London-type black cabs, but the normal cars which regional taxi operators tend to use.)
Everywhere I go, taxi drivers overwhelmingly choose diesel Skoda Octavias for the job. And if they don’t drive one of those, they almost certainly drive a Toyota Corolla Estate Hybrid. While there is the occasional other model, 80% or 90% of the taxis I see are one or other of those models.
So I don’t need to conduct extensive research over the course of several months to choose a good value, reliable car. I just buy one of the two models that the overwhelming majority of taxi drivers actually drive on a day-to-day basis.
Equally, if I ran a taxi fleet and someone pitched me on buying 100 of some other brand for my fleet, I’d be very suspicious because if it was such a good idea, I’d see a vastly higher number of Brand X on the roads working as taxis than I do.
Decision made. I’m buying 100 Skoda Octavias, probably.
2-Inexplicable inconsistencies
The time I’m most sceptical of a proposed course of action is where the people who are proposing it are not acting consistently with the opportunity they are pitching.
A great example of that at the moment is people selling AI solutions.
People claim that AI will save businesses $ billions. But if that’s true why are all the AI companies so focused on helping you make animated videos of your dead cat?
Frankly, the real world business case for videos of people’s dead cats is pretty much zero. Sure, people might enjoy playing around with that, and posting their videos on social media. But there is not a billion-dollar market in consumers ponying up thousands of dollars a year to make videos of Muffin, their much-beloved, sadly-departed cat from when they were a teenager, brought back to life.
Pretty much no-one is going to be handing over more than pennies a month for that.
On the other hand, AI companies claim to have technology solutions which will drastically reduce companies’ operating costs through automating business processes.
That market genuinely is worth billions of dollars a year.
Yet, in recent weeks, MIT have published a study suggesting that businesses see no real world benefits from AI in 95% of the projects they analysed. And there are also some stats suggesting that most organisations have not reduced headcount even when they have implemented AI solution. That’s because a large number of humans are required to check that AI is doing the job properly…because, by and large, it doesn’t.
Now, I am quite convinced that there are some excellent real world applications for AI – running datacentres, perhaps, or automating low-end computer programming. But those are vanishingly small areas of operation for a typical business.
So, my question is, if huge tech companies who claim to have a magic solution to problems worth $ billions to businesses around the world, actually spend most of their own time and money perfecting videos of dead cats, why would any rational seller of tech solutions do that?
And the answer is that AI doesn’t work all that well outside a computer lab. Maybe it will one day, but right now tech companies are turning their back on a business market worth billions of dollars to service a consumer market worth pennies on the dollar.
That’s only a rational decision for tech companies if AI solutions don’t actually work as well for businesses as the people pushing AI solutions claim.
Otherwise I have an inexplicable inconsistency between what tech companies say and what they do.
Faced with an inconsistency like that (in my new, made-up discipline of behavioural accounting) I look at what tech companies actually do, and pay very little attention to what they say.
At least for the moment, AI is an easy “no”. If they people pushing the solutions demonstrate by their behaviour that they are more concerned about perfecting cat videos than automating credit control processes (or whatever other business activity), you can be pretty sure that their business solutions don’t work, or they’d be pursuing a market worth $ billions over a market worth pennies.
3-“Playing what’s not there”
Celebrated jazz pioneer Miles Davis once said the secret to a great performance was not in playing “what’s there” (ie the notes on a page) but playing “what’s not there” (ie how you play the notes).
Another of my behavioural accounting techniques is to look for what’s not there…but should be.
This draws a little from my early career as an auditor – if a company claims to have banked £10 million in sales this year we didn’t see £10 million or so being deposited in the company bank account, we were taught to immediately become highly suspicious of everything that organisation told us, and to make sure we confirmed every piece of company data with independent sources in case the organisation, or its officers, were lying to us.
To apply this in practice, what you do is ask yourself “what would need to be true for X to be true?”
In auditing, you quickly discovered that people tend not to pay you until you submit an invoice, for example. And also, they don’t pay you if they don’t think you’ve done any work for them requiring payment.
So if one company make a payment to another company – evidenced by a payment flowing into your client’s bank account – you can be pretty sure a genuine piece of work was carried out, and duly invoiced. (You do need to check that the second company doesn’t somehow funnel the cash back to the first company again, directly or indirectly, but I’m trying to keep this example as simple as possible.)
When I worked in the printing industry, we didn’t have a product without buying paper or carton board to print on. So if sales were high, but purchases or paper were low, on the face of it (adjusting for any stockholding) either the sales number is wrong, or the company hasn’t recognised enough cost for the paper it must have bought to make the products it sells.
Because buying paper was a necessary precursor for making a sale, so you would expect those two numbers to move more or less in lock-step.
In the part of the printing industry I worked in, you could even prove the sales made to each client if you wanted because we used a lot of special colours of ink (think M&S green, Sainsbury’s orange, or Cadbury’s purple).
If we claimed to be making lots of sales to M&S but were not buying much M&S green, on the face of it, our claimed sales to M&S are unlikely to be accurate. Buying the right shade of ink was a necessary precursor to making products that M&S were going to buy, because they were printed in M&S’s house colours.
Behaviour first, numbers second
Where a lot of decisions go wrong is that there’s a (generally well-intentioned) drive inside organisations to launch a huge analysis project of some sort when faced with a big decision.
Let me be clear, there will times you do need to do a full analysis. But that’s going to happen less than 100% of the time you’re faced with a decision. Significantly less.
What you should do first is triage the problem or opportunity.
First, a quick rough cut to make sure the project is worth doing at all – that there’s enough of a potential upside to make all the analysis, delivery, and ongoing operation worthwhile. If it’s not looking attractive at that level, no amount of data analysis is going to make that into a good project – generally things only look worse when you dig into the detail, so if it doesn’t look good at a “headline level” it’s never going to look good “down in the weeds”.
Then check the behavioural accounting. Look at what people do and ask yourself whether the decision you’re being asked to make is consistent with what you see people doing.
If the answer is “no”, then the quicker you shelve that idea, the faster you’ll stop wasting money on it. Your chances of making that work at all are 100-1 against – and likely to take vastly longer than you think and cost a lot more in the process.
The best decision you can make for your bottom line is to bail out early and do something more productive with your time.
As Charlie Munger (Warren Buffett’s long-term business partner) said: “It is remarkable how much long-term advantage people like us have gotten by trying to be consistently not stupid, instead of trying to be very intelligent.”
Make sure you’re the right side of the line on that, and your business bottom line will thank you.