One of the joys of social media is that strangers will stop by your profile every now and again to tell you that you’re completely wrong about something. Often something that you’re an expert in and they are just over-confident amateurs.
The area that happens most for me is when I talk about whether there is a true RoI (Return on Investment) for a particular course of action.
Bear in mind that I’m a qualified accountant, with many years assessing investment proposals across a wide range of sectors. So it’s always a delight when someone working in tech who has just asked ChatGPT what a business case is decides to stop by to tell me I don’t know what I’m talking about.
That’s partly because tech folk suffer from a common ailment (not known only to tech folk, to be fair) of deliberately “failing to understand something when their salary depends on them not understanding it”.
Tech folks’ share award vesting depends on nobody ever thinking that might not be the case. So you can forgive them for finding it easier not to ask themselves the hard questions. Their future wealth depends on nobody, themselves included, asking those.
It’s also partly because tech folk think that knowing how to populate formula for calculating something is the same as understanding what the formula means.
To be fair, that is true when writing code, which is why this is a frequent blind spot for tech folk.
You only work with surface-level thinking when you write code – you need to explain to your computer, one logical step and a time, exactly what you want it to do next. The computer can’t think for itself – a programmer has to tell it what the sequence of events, and decision points, are.
So, in the tech world, understanding the formula (or programming language, if you like) is the same as understanding how it all works, because tech has no depth to it.
Sure, some people carry out the task a little more elegantly than others, but fundamentally the only way your printer is going to print a sheet of A4 paper is if pretty much the same instructions, word for word, flow from your computer to your printer. Your computer has to communicate in the rigid, pre-determined language structures your printer understands, or no printing is going to happen any time soon.
In the interests of full disclosure, though, this is not just a disease tech folk suffer from. I’d be lying if I said I haven’t come across quite a few accountants over the years who made most or all the same mistakes that tech folk commonly do. And other professions are not immune to it either.
So even if you’re not a tech person, there’s something in this article for you.
Stay tuned, and you’ll discover what a real business case RoI looks like – believe me, this is a lot more complicated than knowing the right formula in Excel.
There are seven areas where people drastically misunderstand what a real RoI looks like. So let’s dive in:
1-It’s not the numbers
The most common mistake too many people make is to take numbers at face value.
I’ve worked with numbers for a long time and I’ve got to say a staggering amount of the numbers people throw into a business case are wrong.
Not mathematically wrong, usually. Most people can get Excel to add up a column of numbers at least semi-competently.
But wrong in the sense that the numbers are a biased view of the problem (consciously or unconsciously), or that they address only a partial view of the data, or that they sound pertinent to the case being put forward but in reality are only tangentially relevant.
My most recent interaction with an over-confident tech person was them claiming that a particular app (which they may or may not have had a hand in writing or specifying) was generating “high levels of satisfaction”.
When I checked the App Store, this app had a 3.8 rating – which isn’t terrible, but isn’t stellar either.
But this tech person missed the point: people didn’t want to use the app at all. They wanted a different solution. So they were dissatisfied. But if the app was the only way to get the service they needed, then, their experience was that the app itself wasn’t terrible.
I don’t know about you, but that’s not a set-up I’d use the expression “high levels of satisfaction” to describe. But this tech person had missed the bigger point that people fundamentally weren’t happy about their experience.
Pretty much every number presented as part of a business case for anything is either less than the full story, or capable of one or more possible interpretations which are different from the rationale being put forward.
If you think a business case is just about finding some numbers and slotting them into a model, you’ve missed the point. You need a much deeper understanding of reality to truly understand whether a proposed course of action is likely to have a positive RoI.
2-Task not system
As I often say, one of the biggest lies in business is the old mantra that to solve any problem, you break it into its component parts and solve each component separately.
This mindset tricks a lot of tech folk into thinking they are making a positive impact on the bottom line when they are doing the exact opposite.
Imagine your business wants to dramatically grow its revenue line. There are many, many components to make a sale in most businesses, but every sales process starts with finding a lead.
Now, if it’s my job to generate leads, I guarantee you I can generate cheaper leads very easily.
Over the last couple of decades I’d have switched from off-line to on-line. I’d have switched from SEO to social media. I’d have switched from text to video. I’d have switched from long-form video to short-form video. I’d have switched from Facebook to TikTok. I’d have integrated AI into whatever I was doing. And so on.
Each one of those – at first, anyway – was cheaper than whatever most businesses were doing before. (In time, they also all trend back to being about as expensive as whatever solution they replaced, but that’s not my point here.)
But, as each wave of new tech is introduced, and as costs reduce dramatically (at first, anyway) that sounds like a positive RoI to most tech folk.
Except your business objective is to grow the revenue line. Lead gen that costs half as much but converts half as well is no better than whatever you were doing before. And even reaching that conclusion requires you to ignore switching costs and training costs.
Focusing on a narrow task can make it look like a business is generating positive bottom line results at the same time as the business is actually going backwards at the “whole system” level.
It’s rare I come across a tech person who is able to do that – especially when the definition of “whole system” encompasses things that can’t be done with tech alone.
3-Visible vs invisible
Tech only works in a visible dimension – what you see is what you get.
If a few lines of code say “send a message to the printer to let it know I’m about to send a document to print and it needs to wake itself up”, then that’s exactly what happens. Every. Single. Time.
That code will never do anything else because it’s incapable of doing anything else. That’s an entirely visible process.
Where humans are involved – which sooner or later they will be, until tech folk work out a way to sell only to other spambots – not only is an exclusive focus on the visible elements less than the full picture, it’s often positively harmful.
For example, I know very few humans who are enthralled by the constant diet of AI spam which pollutes most social media platforms these days. Yet there seems to be no end to it.
The visible dimension might still be getting attended to – maybe your AI tools are auto-posting to Instagram for you. And maybe you’re even still getting clicks and impressions (although it’s no coincidence that those measures are down across most platforms as AI spam overwhelms our social feeds).
Here, the visible dimension of “still posting 3x a day” doesn’t help in the slightest when it comes to assessing how your social media followers feel about your business. The clicks and likes are circumstantial at best…and nowadays are probably mostly automated anyway, so they mean little or nothing.
When we focus only on the visible elements of whatever we’re doing, and forget about the seven-eighths of the invisible elements below the waterline, odds are we’re going drastically off-course and won’t notice until it’s too late because all the visible elements seem OK.
The invisible is a valuable leading indicator of the visible which you ignore at your peril – yet most tech folk never let it enter their thinking.
4-Logical not illogical
Tech is entirely logical. That’s how it works. Logic plus a lot of 1’s and 0’s.
The problem is that only perhaps as much as 5% of everyday life is exclusively logical. So the thinking processes deployed in tech are only valuable 5% of the time.
The tech view of the world is that to sell a pair of jeans you run lots of online ads, get people to click on them, join a sales funnel of some sort, go on at some length about features and benefits, and eventually X% of the people you shovelled into the sales funnel will buy some jeans.
Back in 1985, top London ad agency BBH got Nick Kamen to take off his Levi’s and throw them into a washing machine in a laundrette while Marvin Gaye’s “I Heard It Through The Grapevine” played over the top.
This is a classic, much-admired, award-winning ad.
But not a single element of that was logical. Not the actor. Not the laundrette. Not the song.
None of them had anything to do with jeans in general or Levi’s in particular.
The brand wasn’t mentioned in a voiceover. Only in the last couple of seconds of the ad do you discover what the ad is for. There are no long lists of features and benefits. No sales funnels. No tech.
Yet sales of Levi’s went up 800% in the aftermath of that TV ad running.
I can pretty much guarantee that nothing you do by way of AI-generated Facebook ads is going to grow your sales 800%.
That’s because we live in a predominantly illogical world. Logic only gets you so far. But, in the world of tech, logic is all there is.
5-Tech view not user view
Tech folk get very excited about the tech, but often forget about the user.
The problem with that is, ultimately, the success of your business is determined by your customers buying from you, not by how clever your tech is.
For example, I’ve yet to come across a chatbot which wasn’t complete garbage. So the minute I see one of those popping up on a website, I know the business concerned doesn’t really care about me as a customer.
If it did, either I wouldn’t have had the problem in the first place, or there would be an easier way of sorting it out than going through a series of inane questions which I had to rephrase 5 or 6 times because I didn’t use the precise language the chatbot was programmed to understand when describing my problem.
Almost every piece of tech I interact with has this problem in spades. It’s one of the main reasons I’m deeply sceptical that tech can provide a positive RoI solution to many more of the world’s problems.
However I have previously written about my experiences of trying to book my car in for a service before which I won’t repeat here.
All you need to know is that poorly thought-through tech is just about the most unhelpful way of getting simple things done that you can imagine.
6-Real world vs laboratory environment
A lot of tech works fine in theory.
In a lab, where they can control all the variables, feed a model a pre-determined case study to demonstrate how smart the tech is, and the responses are selected from a pre-determined, pre-optimised list, tech folk can make almost anything look slick.
Trouble is, the real world looks nothing like that.
In my days running a 1,000 seater call centre, while of course there were some well-worn topics people used to call up about, scarcely a day went by when a caller didn’t raise something completely random that we hadn’t heard before.
Real life humans are vastly more unpredictable than tech folk think they are. Even when we thought we had made everything simple and obvious, to a fair proportion of the people who called us up 24/7/365, they weren’t.
And that’s where the tech falls down.
Doing something clever in a lab proves nothing about how it’s going to work in the real world, with a whole boatload of extra randomness thrown in.
The thing is, in our call centre we already know how to respond to things that happen regularly. When someone called up to change their address on our records, pretty much everyone in the call centre could do that competently with little or no tech.
These calls were simple and easy and dispatched in a minute or two by a human for next to nothing, without requiring us to invest in a £10m data centre for our online AI chatbot and integrated website first.
And tech was pretty much useless for anything outside the mundane because the developers had no idea what the real world looked like for us and our customers.
While humans are really good at responding to new and unusual challenges on the fly, tech just drops into a doom loop of serving up essentially the same FAQ write-up which didn’t fit the customer’s needs 10 minute ago when it served them the same article either, but still refuses to connect the customer to a human.
7-Cost to customer, not cost to firm, is what matters
Whenever I hear a sales pitch about how much money a piece of tech will save, I silently roll my eyes. (And sometimes, not so silently.)
Of course, I don’t want to run my business any more expensively than I have to. But that’s not the dimension that really matters here…except to the tech folk who have hung their whole sales pitch on that.
What tech folk (and quite a few accountants) fail to understand is that the dimension we are really trying to manage here is the RoI to the customer of dealing with our business.
Not the cost to the firm.
Those are two entirely different – and often completely unrelated – concepts.
Put yourself in a customer’s position for a moment.
Let’s say I used to be able to ring your call centre and get my address changed in the course of a 2-minute phone call. That was a typical call duration for us in my old call centre.
But now, to save the firm money, the business fires everyone in the call centre and invests half of the savings in a new tech stack to automate all their customer interactions. That’s an impressive RoI, right…?
Wrong.
Now I’ve got to go to your website and navigate some crappy chatbot for a few minutes to find out what to do.
Then I need to enter my customer account number – which I can’t remember off-hand, so I need to search through my emails to find that information. Thankfully “in the interests of security” you only ever show a line of asterisks and the last four digits of my account number in the emails you send me several times a week.
So now I need to go back in my emails for three or four years until I find the one that has my full account number on it from when I originally set up my account.
Five or ten minutes later, I’m back at the login screen with the required information.
Except now I need my password. A password I input when I set up the account four years ago and have never used since.
So I click the “reset password” button and go back to my emails to pick up the magic link – only valid for 15 minutes – which lets me reset my password.
Now I need to guess which rules your tech person has randomly decided I need to follow in setting passwords – capitals, or not? A number, perhaps? Maybe some special characters – but obviously not all the special characters as tech folk like to give us a challenge. And, often, not a password I’ve used before on your website.
So some hacking around to find a combination acceptable to the glitzy tech website takes another couple of minutes.
Next, I’m allowed to go back to the login screen again with the details I’ve now written down on a Post-It note next to my PC. Even though your website told me I’m not allowed to write those details down “because of GDPR” or something.
Ten minutes later, after hunting through a range of menus and options, I’ve finally been able to update my new address in your system
Except now I need to go back into the email system I just signed out of to click the acknowledgement that it really was me who changed my address on your system.
Which takes me to the screen which forces me to set up two-factor authentication in order to be allowed to purchase the notebooks I buy from your business once or twice a year.
The net effect of all this from a customer perspective is that this business has converted what was a 2-minute job, for both me and the business, into a 10, 15, 20 minute job entirely for me where I am required, for free, to give up my valuable time to do something your business should be doing for me if it wants to retain my custom.
Here, whatever the tech people said the RoI to the business might have been on a spreadsheet, I’ve gone from being a loyal customer of many years’ standing to someone who is unlikely ever to buy from you again, thanks to an interaction with your tech.
Ultimately, your customers make buying decisions by factoring in the RoI on their time too. Any business which becomes too demanding on their customers’ time, with too little upside for their customers in the process, just encourages their existing customer base to buy elsewhere instead.
The RoI you really need to work on with tech projects is the RoI to your customer, not the RoI for your firm’s operating costs.
Making your business simple and easy to deal with is a sales growth superpower – because many businesses in many sectors are neither of those things.
It becomes easy to make a sale when your service is both first-rate and reliable, and doesn’t require me to take 20 minutes of my time, without remuneration, to navigate my way around the third-rate tech that some tech evangelist has convinced you will deliver a great RoI for your business.
It’s not just tech folk
As I said earlier, these problems are not exclusively caused by tech folk – I find engineers and accountants also particularly susceptible to most of them.
But a mindset of “all tech is good” or “tech is always a guaranteed money-saver” is almost certainly taking your business in the wrong direction, in the company of grifters who make big promises but don’t deliver.
Not that they don’t deliver to the narrow tech spec which you signed off. They nearly always do that.
But because they don’t deliver where it really matters – with your customers.
Because the RoI on tech projects tend not to consider the points listed above on anything other than a superficial level, it’s becoming rare to see genuinely value-adding tech in a business setting.
I’m not suggesting all tech is bad – computerised accounting was a definite improvement (albeit not without its downsides), as is email as a form of communication. The banking app I use is genuinely excellent. And Google Pay/Apple Pay have made paying for things in shops much easier.
But increasingly Big Tech is running up against the problem that only 5% of the world’s problems can be solved exclusively with logic under laboratory conditions. Those things already have a range of perfectly serviceable tech solutions in place – like Xero and Sage for small business accounting, for example.
Everything else in the world is varying degrees of not logical and/or not operating under laboratory conditions.
While ignoring non-logical and non-laboratory factors can help you show a positive RoI on a spreadsheet relatively easily, the problem is that business case doesn’t reflect reality – either for your business or your customers.
Price in the non-logical and non-laboratory factors, and increasingly I suspect tech solutions deliver a negative real-world RoI.
If you want to avoid doing the same, make sure you check off the seven points above next time you consider making an investment into new tech.
Sometimes the best solution is not to invest in new tech at all. But to do something else instead.