The theory behind KPIs is a good one.
You identify the activities that really matter in your business (that’s the K – for Key – bit).
You decide what results you need to hit your top-line and bottom-line objectives (that’s the P – for Performance – bit).
And you have some way of tracking and measuring those to make sure you’re on target (that’s the I – for Indicator – bit).
Key Performance Indicators make a lot of sense as a concept, I’m sure you’d agree.
But, like a lot of concepts in business, when an otherwise good idea gets mangled inside a corporate setting, their usefulness can decline dramatically.
In many organisations, KPIs just mean a random collection of numbers because, at one time or another, the CEO has asked a question about some data. As a result, that data now appears for ever more on the monthly reporting dashboard, no matter how trivial that bit of information might be.
For a lot of support departments, battling to get a KPI on the monthly report for the board is taken as a sign of how important they are in the hierarchy. Even where that’s taking the definition of the word “key” down a dark alley late at night and punching the living daylights out of it.
I once worked in a large organisation where I was one of six divisional CEOs who brought in all the income and made all the gross margin for the group. Important to know how we were doing, you might have thought.
But the monthly KPI sheet was full of just about everything else the organisation got up to (supply chain, HR, compliance, etc, etc) and completely missed the point that without revenues and gross margin coming in at the required rate, everything else was just rearranging the deckchairs on the Titanic, no matter how interesting the managers of those support functions thought the numbers they insisted on reporting to the board were.
So KPIs are great as a concept. In practice they’re mostly designed badly and implemented poorly. As a result, they often end up not doing the job they are supposed to do inside organisations.
Today, however, I’m going to concentrate on just one of the ways in which KPIs can end up causing more harm than good.
And that’s the question of whether your KPIs inform or excuse.
A week on YouTube
For reasons that needn’t detain us here, I spent a lot of time on YouTube a few weeks ago.
And I noticed something which, at some subconscious level I know all social platforms do, but which hadn’t been demonstrated to me quite so starkly as YouTube did that week.
Normally I only spend 10 minutes on YouTube here and there, so this might have been going on for a while, but I just haven’t been spending enough time on the platform to notice previously.
But in my intensive week of YouTube watching I noticed that my feed was almost entirely full of videos I’d watched before or collections of videos very similar to something I’d watched before.
So, because I’d watched the video for “Back in Black”, I was being served a collection of videos which included “Thunderstruck” and “It’s a Long Way to the Top”.
Because I’d watched a clip from “8 out of 10 Cats”, whole episodes started appearing in my feed, and more clips, and collections of clips, and collections of collections of clips.
Because I’d watched the Mamas and Papas sing “California Dreamin'” (one of the most perfect pop songs ever recorded, in my opinion), my feed was full of all sorts of covers – some horrendously bad, I have to say – of one of my all-time favourite songs.
Now this is social media algorithm-writing 101 – if you find something that gets watched, drop more stuff like that in their feed to keep people hooked on your platform so you can serve them more ads.
In a nutshell, that’s the social media companies’ business model.
It’s deliberately manipulative, of course. But when it comes to serving me music videos, I can probably live with it.
(Although it did teach me that “California Dreamin'” is a much harder song to sing well than most people think it is. Watch 10-12 covers of it on YouTube then watch the Mamas and Papas, if you don’t believe me…)
With one exception
Once upon a time, there was one exception to this social media algorithmically-driven approach.
Or at least a place where it happened a lot less.
And that’s Twitter (as it used to be, not as it is today).
You might disagree about how well they did it, pre-takeover. And certainly in more recent years it was less true than it was, even before the platform turned into the bin fire it is today.
But for a long time, Twitter was about discovery, not about “if you like that, you’ll like this”.
It was a very different conceptual positioning.
The thinking at Twitter seemed to be that if they showed you enough interesting stuff, you’d stay on the platform and they would be able to show you more ads.
They also didn’t penalise including links to sources outside the platform in your posts, which most social media companies do. Twitter figured if you regarded their platform as the place you went to discover interesting stuff, you’d keep coming back even if you temporarily left their platform to read an article in the New York Times.
However, we’re getting diverted here from our main purpose. This isn’t an article on social media business models, although it’s starting to feel like one, so let’s move on…
Back to KPIs
So, what does this mean for your KPIs?
Well, you need to ask yourself whether your KPIs are designed merely to confirm your current view of the world or to give you a different perspective to the one you had before.
The problem with the social media companies’ algorithms is that they are designed to reinforce and validate people’s views, no matter how distasteful or extreme, because that’s how they get you to keep hanging around.
If there are enough people in the world who are crazy enough to share a set of extreme views, the social media companies will move heaven and earth to make sure their social media feeds are full of the crazies who think the same way as they do.
I’m not making a moralistic point here, but just as a matter of fact, humans tend to have their feelings validated by knowing there are lot of other people who think the same way as they do.
Now that might be only one completely crazy person in each of the 200-odd countries in the world, but squeeze that through a social media algorithm and a crazy person will quickly become convinced that the whole world agrees with them, because they’ll see nothing in their feeds apart from people agreeing with them and cheering them on.
KPIs work in a similar, if slightly less insidious way.
If I’m a senior manager in an organisation, and I can convince the organisation to adopt some KPIs which are likely to show my department’s performance in a good light, I’m essentially triggering the same thinking process in the board that the social media platforms try to leverage.
Provided my KPIs are above target each month, the steady drip-drip of “above target” ratings will have the board convinced I’m a genius in no time.
I’ve seen this game played a lot in corporate boardrooms.
That’s how failing organisations somehow limp on until it’s too late to save them. There has been so much collective back-patting inside the organisation because their KPI reporting says that everything is going perfectly – even as the bottom line goes deeper and deeper into the red each month and the cash flow stops flowing into the business and mostly flows out.
In that sense, KPIs are often designed to excuse the performance of individual departments, not to inform the running of the business.
Because everyone around the boardroom table keeps telling one another that Bob in Customer Services is doing a great job because he keeps hitting his KPIs, just like lots of crazy people imagining their views are acceptable because they keep having those views validated by an audience of exclusively like-minded individuals on a social media platform, Bob escapes the in-depth scrutiny he might well need.
But there’s another way
There is another way.
A bit like the way Twitter used to be, you can use your KPI reporting to inform the board about what’s really going on, rather than just to use your monthly reporting to excuse whatever the current level of performance might be.
But to do that, you have to largely ignore whatever you’re being told in the KPI reports by people with a vested interest in them being shown in a good light.
Let’s take a simple example.
Most organisations realise that offering a good level of customer service is a wise strategy if they want to grow and protect future revenue streams.
The way they track that is often by way of the NPS (Net Promoter Score). That’s the concept which is responsible for all those surveys you get after buying something online, asking you to score your experience on a scale of 1 to 10.
And it’s not a completely crazy concept.
It doesn’t take a huge leap of faith to believe that people who rate your customer service highly are more likely to recommend your business to other people they know who are looking for the same products and services as you supply.
So far, so good.
But NPS scores are entirely gameable, and you can provide a pretty poor service whilst still getting halfway decent NPS scores, if the person running your customer service department is so inclined.
As a board, you have a choice.
Do you just pat the head of customer service on the back for turning in a 7.2 this month against a target of 7.0?
Or do you go on a voyage of discovery?
Is it real?
The question I always ask about KPIs is “is it real?”.
I know that, mathematically, it’s likely to be an accurate calculation. Although it’s not completely unknown, very few senior managers in large organisations will tell an outright lie in their board reports. The maths is correct 99.99999% of the time, and on the few occasions it isn’t, that’s because your senior manager doesn’t understand maths, not because they’re being dishonest.
The maths usually isn’t a problem.
Rather, the problem is that the mathematically-correct answer to a calculation which has been gamed to show the person managing it in a good light does not necessarily reflect reality.
If you think I’m being harsh, call up a few major multinational companies’ customer helplines today.
I’m prepared to bet that 90% of those experiences will be terrible.
And I’m also prepared to bet that their internal KPIs will show they have achieved a perfectly acceptable customer service performance this year, however they choose to manage it.
Just that one simple thing – calling up your own call centre as a mystery shopper and not just taking the metrics at face value – means you’ve started to inform yourself about what the real customer service performance is like in your organisation. You’ve started to inform yourself rather than excuse whatever your organisation’s customer experience is by pointing to an all-green KPI dashboard.
You see, if someone is above target on the KPI sheet, they don’t get many questions. The focus switches to people who are performing below target.
There is a reason to do that, of course, and if there’s a problem that needs fixing elsewhere, that almost certainly requires some attention.
But in taking that view, the organisation is potentially underappreciating the magnitude of the problems bubbling under the surface (which, by the way, will usually pop out as a much bigger problem in a few months’ time) in favour of sorting out a more visible problem elsewhere, albeit one of, ultimately, a lesser order of magnitude.
Evidence sources
I consider myself really fortunate to have studied Law at university, although I never became a lawyer, because that’s given me a mindset I’ve often deployed over the years.
And it’s based on my second-year class on the Law of Evidence.
That’s where you learn how to “prove” a case in court…what the burden of proof might be…how compelling a particular piece of evidence might be, and so on.
So every time I see a KPI report, mentally I always treat it like the pleadings from “the other side” in a legal case.
I’m probably not saying to myself “isn’t the customer service team doing well to hit a 7.2 NPS”.
My internal dialogue is much more likely to be “how can I corroborate that claim, and is there any evidence around which might suggest that claim is not as well-founded as my opponent would like to think it is?”
I’m prepared to admit this approach has not been universally popular at places I’ve worked in the past but, while there are some benefits in the short-term for everyone basking in the glory of a 7.2 NPS, if that’s a house built of straw, the quicker you find out and fix the underlying issues before anything goes horribly wrong, the better, as a rule.
So next time you get a KPI sheet or performance dashboard of some sort presented to you, ask yourself this question – how would I be able to corroborate the picture of the organisation being presented to me here?
And then go and do that. Sometimes what you find opens the door to a perspective that you can use to transform your organisation for the better.
A perspective you’d never have come across if you had just taken the KPIs at face value and murmured your congratulations, along with everyone else around the boardroom table, to Bob in Customer Services for doing such a great job and turning in a 7.2 NPS.
When lots of people have convinced themselves that their views are correct because all the see and hear is people agreeing with them, that’s the point of maximum danger. That’s the point you need to find other ways to corroborate whatever you’re being told, because nine times of our ten, you’re probably not getting the full story