It’s easy to forget that metrics alone tell you very little about what’s happening in your business. So little, in fact, that I’m surprised so many people are quite so obsessed with them.
Please note: I’m not saying you don’t need any metrics at all, just that most businesses could be run with a fraction of the metrics they collect at the moment, and almost certainly don’t need the cottage industry in management reporting that sustains a fair part of the overhead costs in the P&L.
At best, most metrics give you a surface indication of what’s going on. They’re a comfort blanket of sorts…providing cover for people to justify their existence, mostly, without necessarily moving the business forward.
You see, once someone convinces you that the management fad of the day is vitally important for the future success of your business, they’ve effectively also sold you on the need for a department full of people managing whatever that thing is, a daily/weekly/monthly reporting cadence for that thing, a slot at every board meeting to talk about that thing (leading, they hope, to a place on the board for themselves in due course), and a pretty secure job for themselves.
After all, you couldn’t stop doing something that was so vital to the company’s future well-being could you?
Now, lest you think I’m a grumpy old Hector complaining about “young folks nowadays”, that doesn’t mean that I think all modern management fads are entirely without merit.
Mostly, there’s at least some point to them. And I, for one, would like to think I left the world a better place than I found it, so I have no problem at all with most of the things businesses are expected to do these days.
It’s turning these noble-enough objectives into metrics I’ve got more of a problem with. Because then the activity – however worthwhile it may be – turns into a process of managing the metrics, not a process of delivering the outcome which the process is supposed to deliver.
That might sound a little convoluted, so let me explain.
Many business processes are terrible
Firstly, and perhaps most obviously, many business processes are terrible.
They’re often badly designed, customer experience and user experience are generally woeful, they’re usually inflexible, and they arrogantly assume your customers have all the time in the world to interact with some unholy combination of your website’s FAQ page and some soulless, AI-powered chatbot.
And that’s just the good bits.
Business processes are also often under-resourced, poorly integrated with a company’s wider systems, and unsympathetic to anything other than pre-determined, pre-defined scenarios. The minute anything happens that doesn’t fit neatly into a box on a process chart somewhere, the system grinds to a halt because nobody knows what to do.
If you think I’m being a little unfair, I used to deliver business process re-engineering programmes.
All of these things happen on a regular basis, and every time you find one of those conditions, it means your business costs are much higher than they need to be – there’s no cheaper way to run a business than to “get it right first time” and resource your activities properly. (And before you think it, no AI doesn’t help that at all. It just makes it cheaper to deliver the existing terrible service.)
But here’s the bigger problem
Terrible though many business processes are, you create an even bigger problem when you overlay whatever’s going on with a set of metrics, and a regular reporting cadence.
Then, the activity becomes all about the metrics, almost to the exclusion of the original objective.
This is even worse in the age of AI because it has effectively become free for people to game their metrics. A random person in your company might be no closer to achieving the company’s objectives, but they keep their job because their monthly RAG-rated report is all in green territory.
Gaming the metrics without impacting the outcome is, by definition, a complete waste of time and money. And, to be fair to the tech bros and gals, this went on long before AI was invented. AI has just turbocharged the pointlessness of it all.
Let’s take one everyday business example to illustrate.
You are no doubt familiar with the old sales and marketing mantra that you need seven touchpoints with a potential customer before you make a sale.
Let’s skip over the fact that this is nonsense – there is no magic in the number seven, any more than there is in 3, 5 or 12.
It’s very likely to be “more than once”, unless your business is an extremely well-known brand, or you’ve had a personal recommendation from a close family member into the decision-maker.
But enough articles have been written about seven being the magic number…enough motivational speakers have quoted that statistic…and enough sales and marketing textbooks include a reference to this particular “golden rule” that pretty much everyone believes it now.
And, as I said at the start of this article, there is some underlying merit to this principle. The answer is almost certainly “more than once”. But plenty of sales are made on the 8th, 10th, or 42nd touchpoints too.
It’s the distilling of a reasonable enough principle of “more than once” to “exactly seven” that I have a problem with.
However, since everybody now believes the magic number is seven, I can now set up all my sales and marketing processes to hit those magic seven touchpoints. If you’re my boss, you’re unlikely to challenge the conventional wisdom when I tell you I need the resources for seven touchpoints, because everyone knows that’s the magic number, right…?
Spare my inbox
So now we have all the ingredients in place:
- a terrible process to reach the magic number of seven touchpoints, by hell or high water
- a metric we can track (is it more or less than seven, at the time of writing this month’s report?)
- and a system we can game to make it look like we’re indispensable to the business.
You’ll note something subtle has happened here.
The only real objective of a sales process is to make a sale. Everything else, no matter how elegantly it’s performed, is just window-dressing.
However, my objective now is to engineer seven touchpoints. It’s not to make a sale.
This is especially true if I’m in the Seven Touchpoints Department and some other department has to make the sales.
Then I’m sitting pretty even if we never sell a thing because I’ve held up my end of the bargain: “Hey, check my metrics – I hit the seven touchpoints. It’s not my fault the sales department can’t do their job properly.”
However it’s a great protective mechanism even if you, or your department, is responsible for both making a sale and achieving the seven touchpoints. That’s because I’ve persuaded you to believe in the metrics, and switched your attention away from the outcome you wanted…at least most of the time.
Nowhere is this phenomenon more obvious than my email inbox at the moment.
The amount of AI-generated garbage sales messages which end up in there is beyond a joke – they’re poorly targeted, inelegantly delivered, and terribly written.
Yet I can guarantee you that someone somewhere is counting every one of those as one of their seven touchpoints.
The problem with that theory is that, at least in the old days, the concept behind the seven touchpoints was that you had to do seven different things, not the same thing seven times over.
At some level, I get the need for a degree of persistence in making a sale. That’s true when humans try to make a sale and it’s true when our robot overlords try to make a sale too.
But persistence alone rarely makes a sale. More often, what persistence results in is a stiffening of my resolve never to buy anything from your business for as long as I live.
Last time I checked, cheesing off all your potential customers wasn’t a great way of trying to build a business, but hey-ho, no tech bro or tech gal seems to have worked that out yet, so the daily torrent of garbage continues for now.
It’s clearly news to those responsible, but I’m not likely to buy anything from people who irritate me. There is no shortage of arrangers of business finance and utility brokers in the world – if I ever need one of those, it won’t be hard to track down someone who hasn’t irritated me intensely up to that point by sending me a never-ending flood of clumsy sales emails first.
The metrics that matter
Metrics like the seven touchpoints, or the number of daily emails sent, are largely pointless.
I get it that if you send zero sales email, you’re unlikely to make a sale from an email. But if you’ve sent a thousand emails and I still haven’t responded, the likelihood of me responding on the 1,001st is as near to zero as makes no difference.
So the benefit of tracking a metric like “emails sends” is also pretty much zero.
The metrics you should track are the activities or events which indicate you’re moving closer to your end goal – which, remember, is making a sale, not sending out seven identical emails.
And to do that effectively, you’re by and large looking for information external to your business – or, at the very least, external to the process you’re running here – which indicates a changed state at the customer end of the equation.
1,000 emails I don’t respond to means nothing. Except the fact that I almost certainly will never buy from your business, of course.
A single email – whether that’s the first, the seventh or some other number – that I click the “book a demo” button on? Now there’s a changed state.
There’s engagement. There’s an indication that someone is moving down the decision-making process. There’s an early indicator that a potential sale might be in play.
The changed state, ideally from outside your business, or at least outwith the process, is important because that’s much harder to game than just sending out seven identical emails written by AI.
It’s not impossible to game, but it’s a lot harder to game.
And by the time you’ve strung two of three of those together, you can be pretty sure that you’ve got a genuine prospect on your hands.
If the demo leads to an in-person sales call, which leads to your prospect attending an event hosted by your company, you’re way past the opportunity for gaming a metric. That’s almost certainly a real sales opportunity, which you should have a decent chance of converting, because your potential customer wouldn’t have wasted that much of their time in an activity they had no interest in.
So the metrics which are most helpful to track, and the hardest to game, are the state changes from outside your process, not the activity taking place inside it.
But proceed with caution
In the early stages of a process like this, gaming the metrics is still possible. It’s only later on, when your prospect has invested significant time and effort from their side, that it’s almost impossible to game, unless your sales manager has 150 close relatives.
You can, for example, create fake engagement with a sales email.
You can almost certainly do this with some dodgy software, although that’s not something I have any knowledge of.
But you can also do it by putting things into your barrage of emails to encourage people to take some action.
My favourite of the moment is “Reply ‘no’ and I’ll stop sending you emails”
I never reply “no” because the minute I do, the AI bot on the other end knows my email address is real and active. All that’s going to happen here is someone will book my reply as “engagement” against their engagement metric, and the AI is going to get cranked up to 11 because it has convinced itself that I’m a real prospect because I’ve replied.
One day I might do that just to see what happens because they’ll almost certainly ignore me, which means I’ll have a case to take to the ICO, but frankly I’ve got better things to do with my time.
So, until then, I’ll just keep ignoring their emails, and mentally ratchet up their position on the list of organisations I’m never going to buy anything from instead.
If “engagement” is only defined as a reply to an email or a click on a link and nothing else happens from the customer end, that’s not really engagement at all. Someone might have hit reply by accident, or you might have caught them at a weak moment.
Movement is the key from the point of initial engagement onwards. Is your prospect going further down the sales funnel or are they just stuck at the second level without any signs of them ever making progress in the direction of an eventual sale?
Timescales can be quite helpful here.
If a prospect took an action, even if, in theory, it was moving them closer to a sale (eg by clicking a link in one of your emails) but they haven’t done whatever the next thing in your process is in the next 30 days, they need to come out your “Level 2” prospect list and go back to being a “Level 1” prospect.
Of course, it depends a bit on your business and your sales cycle, so 30 days isn’t a hard and fast rule. But you need to define a timescale which forces the sales team to recognise that they need to start from the beginning again. They can’t just carry forward a metric to justify how busy they are if it now looks like a genuine prospect has dropped out the running for some reason.
I’ve spent too many years being told by sales directors that their extensive list of Level 2 prospects are “going to buy any day now” that long ago I stopped regarding metrics like that as very helpful. It’s movement through a process I’m looking for, not how many Level 2 prospects you had on the books at the end of last month.
It might well be interesting enough to know, but the predictive value of knowing the number of Level 2 prospects isn’t nearly as high as knowing that one prospect, in the last 30 days, has downloaded a fact-sheet, met in-person with a sales manager, and is attending one of our sponsored events next week.
Movement is the key. And that’s hard to game – especially beyond the initial engagement – because very few people have the time to waste meeting salespeople to talk about products they have no intention of buying.
It’s not just sales
Although I’ve used making a sale as an example above, the same principles apply to every department.
Every single department in your business is likely to be able to game whatever internal metrics you give them, so they are often of questionable value – even though you’re likely to be paying a group of expensive people to manage them, track them, and report on them.
One example I came across recently was an HR Department trumpeting their success in health and safety because everyone in the company had attended the compulsory company-wide health and safety training.
I don’t know about you, but no matter what I thought about health and safety training, if the HR Department told me I’d be on a disciplinary if I didn’t attend the training, I’d make sure I attended the training, even if only under sufferance.
Would I necessarily remember what the trainer told me, or act on whatever recommendations they shared at the end of the session?
Well, that’s a different matter. Knowing that 100% of the company attended the training is interesting as far as it goes, but it’s likely to be a less-than-perfect indicator of whether there will be fewer accidents at work in the future than there have been in the past.
That’s because this is an internal metric, based entirely within the process itself (ie “threaten everyone with a disciplinary if they don’t attend”). It’s been gamed by the HR Department to claim a victory against their metrics, but there is no evidence of a change of state from outside the system at all.
To be fair to that HR Department, there are some legal benefits from putting everyone through “sheep dip” training like this in the event of an accident in the future. We did everything we could, the Head of HR can tell the Health and Safety Executive Inspector after the factory roof gets blown off in an explosion.
But what you really want is fewer accidents at work, more people wearing their safety boots, nobody obscuring the warning signs in the factory with random pallets of spare parts, or whatever.
It’s entirely possible to have compulsory company-wide training which makes no difference at all to any of those states, all of which would make accidents at work less likely in the future.
When you design metrics and reporting systems, counting “steady state” situations – such as the number of Level 2 prospects you have or the number of people who attended the compulsory health and safety training – is of extremely limited benefit.
Nearly always, all those metrics do is give a get-out clause to someone who can then say, “I did everything I could – look, I hit my target of getting 100% of the staff to attend a compulsory training event”.
Track the metrics from outside the system instead, especially those which indicate a change of state which are, by definition, much harder to game.
Then you might have a set of metrics you can truly run your business on.