How playing Inappropriate Guess Who? taught us the truth of customer analytics

This article was originally published February 4th, 2020

We play a game that I’ve realised is a perfect analogue for the dreadful state of current customer analytics. Inappropriate Guess Who? is played with the standard Guess Who? set but you aren’t allowed to ask any of the normal factual questions. No asking about beards, or glasses or even gender. In Inappropriate Guess Who? your questions are all subjective and generally filthy.

• Does this person look like they are okay with lockdown because it means they no longer have to do their turn in the tea run?
• Does this person appear to be at the final stages of a beautiful act of sexual congress?
• Would you trust this person with your kids?

At NRF we occupied one end of a bar and challenged all comers, up stepped Tim Ford, boss of the quite brilliant pricesynergy.io business.

‘Has this person got any hair?’

‘Tim you can’t ask about verifiable features.’

‘Okay, erm… have they just had a really painful shit.’

‘Christ. But yes Tim. They definitely have.’

And thus the game continues until either player is left with but a single person and can guess the identity of their opponent’s character.

‘Is it Bernard? It’s Bernard isn’t it? I flipping know it is.’

‘It’s Alice, Tim.’

That’s the awesome thing about Inappropriate Guess Who? nobody ever gets the right answer. It’s stupid and the fun is in the depravity of the questioning. Inappropriate Guess Who? is all about profiling by outward appearance, bias and cultural prejudice. My guess based on these things is often way off from the other person’s similarly biased judgement. It should always fail because profiling and prejudice are personalised failures of thought. They are failures of dialogue, empathy and culture and they belong to my issues, where yours may be different.

The failure inherent in Inappropriate Guess Who? is also the failure inherent in almost all current customer analytics; you’re trying to guess a likely future behaviour based on biased correlations, and you’re segmenting based on criteria that in real life are absurd. You’re also trying to ascribe a fixed point to that person’s buying character when the reality is that how a person shops changes with mood, need and context.

Just like the subjective judgement of Inappropriate Guess Who? current customer analytics are an attempt to make order out of chaos, a chaos of conflicting signals and confusions made to appear logical. The truth is that in retail especially, we’ve been let down by the current state of customer analytics.

What’s needed is a new generation that stop trying to chase the individual’s fixed buying state but that instead consider context and emotion. If you know what groups of shoppers are trying to do and how subsets of them are feeling while they are trying to do it, then you can build a set of experiences and service propositions that meet those without personalised bias.

That need for new solutions is why Uncrowd exists; we’ve created a customer analytics platform, and a whole philosophy that is a generational step forward precisely because it isn’t single-view of a fixed customer. We’re so ahead that one person at NRF exclaimed ‘NOPE! What you’re saying isn’t possible!’ and then we showed them Friction/Reward Indexing using real data on their stores that we’d gathered earlier in the Pennsylvania market… their response then? ‘This is magical, you’ve done it.’

I guess what I’m saying is that just because NPS values a 9 from Fred West exactly as much as it values a 9 from the Pope and assumes they both then do the same thing post scoring, you don’t have to.