Clunkers: learning by breaking

The power of a clear goal, and an unclear path to get there

You may think that what happens to a bike when you ride it downhill is relatively predictable - that you could easily design a bike to do that job? However, you face choices. What would you optimise for? Durability, comfort, speed?

The story of the Clunkers is well known - "the mountain bike evolved from junkyard dogs to purebred rock- and rut-eating rigs with full suspension and prices that can reach well into five figures". A whole market exists, whole sports categories, Olympic medals and more, for something that was driven by the market, not by bike companies.

That evolution, of course, was driven by a focus on being fastest from top to bottom, and learning faster than the others how to do that. Ride it, break it, fix it where it broke and see what breaks next time. Different ideas led to different solutions, all tested against which was fastest. It is a challenge that remains open today. There is no doubt that designing an unbreakable bike a priori would have led to a machine that was too slow to be fun, or perhaps to get up the hill in the first place. So, the need for speed was important. The goal was not just to get down the hill.

It's the same way in Formula 1 - all the in silico design still has to cope with the real world and how things break. Using 'where did it break?' as a guide to improve things is a great way to learn more than the others, with ‘make it faster’ as your goal. Hence the multiple sources of data, and ways of interpreting them. Another source of asymmetry is challenging the team with 'what if?' and breaking not just things but prior assumptions. A Formula 1 engine will not start from cold - it is essentially 'seized' until it is mechanically warmed-up, such are the precision tolerances, engineered to perform optimally in a race. Someone had to question if an engine had to be able to start from 'cold' (what does the rulebook say, what are the consequences of pre-warming an engine in the real world?) to have focused on their goal of minimising engine tolerances, to gain that performance advantage... Of course, like in pharma, ‘faster’ is rarely a simple construct. Fast in a straight line would be of no value to an F1 team if it means being slow in a corner. Fast over one lap may not be the same as fastest over 62.

The history of prototyping is built on learning by doing - learning things that are otherwise unpredictable until they hit the real world. Trial, error and new ideas. You’ll behave differently if you’re riding down the hill looking to learn something than if you’re looking to prove something. A phase II program set up to learn will look different than one that sets out to prove, too, but that is more rare in pharma’s more linear approach.

Prototyping also helps understand what is possible - or what is wanted. This is a remarkable opportunity in pharma, where 'unmet need' is rarely understood by either party (patient or pharma): 'what do you want?' - 'well, what can I have?' One way out of that loop is to prototype... To learn by showing (and failing, but building better). In the presence of a prototype, we both know what we want more or, or less of, or another idea entirely might occur...

Speed to the insight is what brings asymmetric advantage. Even when you connect an 'a' and a 'b' (or a top of a hill and a bottom of a hill), the best way to find out what is fastest is to try more than one route.

That applies to 'first to market' or 'best to market' positions in pharma too. Planning to learn is as critical at the top of the slope as speed of response to, or finding a solution for, a breakage half-way down.

Nothing comes out perfect the first time.

All the thinking in the world won’t automatically generate a solution that has ‘thought of everything’. That is why IDEAs need prototypes.

The MIT professor and author Michael Schrage introduces the idea of prototypes with a wonderful concept, called “The Magic Mirror”. This mirror allows you to see yourself 5kg heavier, or 5kg lighter, with a beard, with red hair, in ten years if maintaining / changing a current lifestyle, and any other command you give it. The important thing is not just that it’s clever – it is how much you would use the mirror, how often – how many different visualisations would you run past it? (Would you ask your partner to use it, too? Would you let your partner see you 5kg lighter..?) And, as importantly, what would you do with the results? Would you act upon them?

Each version of ‘you’ would be a prototype – something offered up for evaluation. We always illustrate prototypes further by providing an additional example: you ask us to develop a great coffee cup. Whatever we offer up, you are bound to have additional ideas to critique / improve ‘our’ coffee cup: it needs to be transportable, fit in a cup holder, look attractive on a coffee table, keep coffee hot for 2 hours, be recyclable, etc. What just happened is that you refined your brief to us in a way that would have been difficult had you just been thinking through ‘what you want’ in your next coffee cup. We also went away with ideas about how to make a differentiated, and ‘better’ coffee cup.

Think about the iPod. When it started as a square-ish white mp3 player, the idea of an App Store that would drive revenues, and the adoption of an iPhone, was a long way from Apple’s mind. However, over time, as the tech developed, Apple kept asking users what they might want. Unlike most other companies, however, Apple prefers to ‘show and see’ rather than expect the audience to do all the hard work – ‘if it did this, would that be cool?’

So it is with product IDEAs – the path-to-market strategies for a new product. The only way possible to find out what customers might want from your product, or to see how they might use it, is to prototype. Laying out several ways to ‘see’ your molecule is 180 degrees away from just laying out your product profile (base or optimistic) and then asking them what they think. Think around the problems with that approach – they can only imagine within a certain frame of reference (the data that you have collected and shown) how good the product is at one thing. They certainly don’t know enough about it to think about what else it might do… Whereas, one hopes, you do.

The intention behind prototyping is not to pick ‘a winner’, but to “improvise with the unanticipated in ways that create new value” – once several prototypes have been developed, investigated with the audience (internal first, as a proxy, then external) and the strategic ‘So What’s’ have been examined, newer, better IDEAs can be built.

For example, if a physician comes back and says ‘all that QoL is nice, but I’d really be using this to drive ultimate efficacy’, or ‘the patch isn’t a way to manage GI tolerability, it’s for patients who are non-compliant’, or ‘that’s really interesting – the idea that depression and pain are often inter-related and interdependent, I hadn’t seen it before, but I recognise that’, you then have a brief to look further and to harness these observations to develop the product that way. You can understand all of the strategic ‘So What’s’ of each of those potential directions, and then choose a path with much greater conviction, and much greater differentiation, than before.

‘Not picking a winner’ is hard for a lot of people. The phrase that comes back most often from a prototype IDEA (internally) is ‘I don’t like’… In our workshops, ‘I don’t like’ has its own $1 swearbox, unless it is accompanied by a ‘because…’ The ‘because’ gives us something we can improve. Prototypes are never offered up and defended – they are for play, for manhandling and criticism, because the ‘because’ gives insight that couldn’t be gained any other way.

Ideation likes the simplicity of a goal as simple as ‘fastest from top to bottom’. But being open to seeing other, new goals is the real opportunity-seeking behaviour.