Pivoting is an enormous comfort blanket, that warms you with the reassurance you that it’s better to do something – anything – than persist in some perceived wrongheadedness. But… who says that you’re wrong?
Recently, I was lucky enough to have a conversation with an extremely experienced UK security camera principal, who generously let me see a little ‘behind the curtain’ to how things work in his security camera focused organization. He also advised me that though what I was doing was good, he thought I might have slipped a fraction behind the UK market and that I should now make a fairly significant pivot to attack quite a different segment to stay ‘in play’.
I agonized over this for a couple of days: should I follow the numbers I had on my desk and just keep on going, or should I drop what I had been doing and follow his (genuinely very informed) take on the UK market? Persist or pivot? Stick or twist? Should I stay or should I go?
And then it dawned on me that the kind of high-level ‘extreme pivoting’ Mike Maples talks about only makes sense if you haven’t really engaged with your customers from the start. Because if you haven’t built your startup around what your customers repeatedly tell you they need, then you basically deserve to fail: so extreme pivoting is just another way of saying “we failed miserably but still had enough VC money in the bank for another roll of the dice“.
Hence it seems to me that Eric Ries-style pivoting (learning from low-level exposure to customer feedback) is almost antithetical to Mike Maples-style extreme pivoting (learning just after it’s too late from a Game-Over failure case, but having enough credibility or cash to try again). Yes, they’re both “learning”: but Ries tries to learn before the failure (to avoid it), while Maples almost seems to be advocating a Nietschian learning after the failure (so as to emerge, phoenix-like, from the flames) – that which does not kill you makes you strong, etc.
In many ways, they’re both wrong – Ries arguably promotes learning too early (customers often don’t really know what they want, so how much can you sensibly interpret what they tell you when you show them an early version?), while Maples arguably promotes learning too late (failure can be a equally lousy teacher too, if you don’t honestly know why you failed).
In the end, the security camera principal guy was absolutely right, but he’s fixated on competing in his particular corner of the UK market and I decided that I’m looking at quite a different (and much larger) picture. But all the same, I researched and researched until I found a way of adapting our existing security cameras designs cheaply and quickly to allow them to compete in broadly the way he proposed (albeit in a completely different manner).
Demo’ing early product to customers is a kind of statistical sampling, in that you aim to take on board what they tell you to make a better product in the next iteration – but it’s pretty obvious this is replete with potential sampling errors. How did you select your customers? Have you selected enough of them to form a worthwhile dataset? Are they being honest with you? How did you decide what to ask them? How can you be sure that they’re responding in the spirit you think they’re responding in? How are you ensuring that you’re not suffering from confirmation biases (etc)?
In short, just because you can pivot doesn’t mean you should pivot. Something to think about, anyway.
Leave a Reply