When ‘Best Practice’ isn’t Best

Like any other journalist, I get a torrent of press releases every day. Certain words and phrases feature prominently. Too prominently, too often. This post is not specifically about those words and phrases, it’s about how these words and phrases have come to be so numbingly banal. The phrase is ‘best practice’, and this post is a cautionary tale of when ‘best practice’ actually leads to failed practice, and why.

It seems as though ‘best practice’ would, by definition, always be the best way to do something. But sometimes it’s not. This is because some problems aren’t conducive to the creation of one ‘best practice’, but instead need to constantly adapt to a changing set of background conditions.

There are, broadly, three types of problems in this world: those with static background conditions; those with changing background conditions; and those where the background conditions change in response to your actions. ‘Best practice’ only works well for the first of these problems, less so for the second, and often fails entirely in the third.

Consider the problem of finding the quickest way to walk from point A to point B (assuming pedestrian traffic is always light). Once you’ve discovered the quickest way, that route doesn’t change. That can be a ‘best practice’. Easy.

Now consider the quickest way to walk from point A to point B in a suburb that is constantly undergoing road works. Each day these road works occur in a different location, and you don’t know ahead of time where they’ll appear. But when they do, they block off that street, forcing you to take a detour. Some days your ‘best practice’ route is clear of road works. Other days its blocked. On the days its blocked, you don’t find out until you’re already half way down the street, then you find you have to backtrack and go around, which takes longer than if you’d taken a different route.

In this situation, finding the ‘best practice’ is more difficult, but not impossible. You could, over time, build a picture of the different routes and the probability that each is blocked at some point by road works. Then you could build a decision matrix that represents all these values and find the route that is, on average, most likely to get you to point B the quickest.

This is the domain of decision theory, which deals with how to make decisions given background conditions that are fixed or that change of their own accord (meaning: the route you take doesn’t influence whether road works are more or less likely down that route).

But now consider the third type of problem, where the background conditions change in response to your actions. Say there are thousands of people attempting to get from point A to point B at the same time as you. If you all followed the steps above, you’d all agree on the quickest route – which would rapidly get clogged with pedestrians. So, anticipating this, you decide to take the second quickest route in the hope of avoiding the traffic. But lo, everyone else thought the same thing, and that secondary route is now clogged while the primary route is empty.

In this situation, the route you take influences the background conditions for you and for other people by contributing to potential congestion. In this situation, you need to consider how others will respond to your actions and how that will affect your choices. This is the domain of game theory, the more complex (and far more interesting) cousin of decision theory.

In this situation there will be no one, simple, ‘best practice’ that will remain the same over time. Even if there was a way to encourage everyone walking from point A to point B to take turns taking a variety of (longer and shorter) routes, thus lightening the load for your all (the ‘Pareto optimal’ solution, to use game theory jargon), there’d be temptation (to ‘defect’) for some to exploit this predictability and sneak into a shorter route, to the detriment of the whole system. If enough people defected, you’d reach an impasse where everyone is taking longer to get to point B (the ‘Nash equilibrium’).

Now to press releases. I would suggest that the problem of choosing the most effective language in a press release is a problem of the third kind, not the first two. Your press release needs to stand out. Yet if you use the same ‘stand-out’ language as everyone else (i.e. ‘leader’, ‘unique’, ‘best’, ‘top’, ‘premier’, ‘fastest’ etc), then your release no longer stands out.

‘Best practice’ might dictate that using this language has the most impact, but that’s only when this language is used in isolation. However, in reality, when everyone else is using that language, then the effect sinks and the reader switches off – or even gets irritated.

So the language that you use should not remain static, but should adapt to the ever-changing background landscape. You don’t want to go too wacky – that’d be like taking an overly long route – but your effect is lost by jumping on the same overused buzzwords. Finding that ideal solution is a real challenge – finding the language that will deliver the message and stand out without being dull or tired will take constant effort. But if you don’t, then the entire industry spirals down into the nasty Nash equilibrium state, and then everyone loses. I’d say that’s where we are today.

In ‘games’ like this, there is no one ‘best practice’. You need to actually be smart to succeed, not just good at following rules. So next time someone cites a ‘best practice’ to you, consider the nature of the problem, and consider whether this ‘best practice’ will actually make you an innovative thought-leader or a drone crushed under the weight of vacuous buzzwords.