3 second answers vs 3 month answers

Which is better...?

How long does it take to forecast a market opportunity? Or to estimate a clinical study duration, enrolment and size? How long to estimate a launch price? It can take months: based on a target/ TPP, you commission some research, see what it says and send off an answer. Or, it can take seconds.

Unfortunately pharma mostly adopts the former approach, whereas the second is the more helpful to the group dynamic in early phase.

Because the goal of early phase is to explore, it is infinitely more helpful to collaborate with team members' 'what if' questions, with good answers instantly, instead of more detailed answers slowly. 'What if...?' requires feedback quickly. Slow, steady answers put so much weight on the first hypothesis being fixed that they're useless as soon as they're requested. A 'long range forecast' is wrong as soon as it is asked for. If, instead, what is asked for is 'if we launched into a rapid progressor population, instead of all-comers, would that affect the studies, the label, the price?' The most useful answers are fast ones, especially if they come armed with their own 'yes, but...' or 'yes, and if...'

'Good answers quickly' is a form of asymmetric learning. As soon as you slow it down to a traditional 'detailed answers slowly', you're back to asking the same questions in research as everyone else. Obvious answers, or the same answers as everyone else sees.

Having teams become comfortable with helpfully quick answers means having team members who can estimate, based on experience. If the only resort is to go to drawn-out research, the process is already broken.

What if, for example, you'd like to know the opportunity and risk profiles for 20 or 30 different launch opportunities (which is a reasonable number for most assets we look at)? Would you like to run a traditional exercise across all 30? No. You couldn't afford the time or the budget. Instead, companies then opt to do it for 2 or 3 at most, knowing that they're ignoring more interesting 'what ifs' in search of boring predictability. However, you could make good estimates of most important parameters for 20-30 options in one afternoon, with the right people working together. Which of those approaches is more likely to lead to an interesting choice of launch option? You may have wide confidence intervals on your estimates, but that's OK - you know how to increase your confidence in any of those assumptions. If you opt for the long, drawn out process, you may have great confidence, but only in something that will never happen - the risk was all baked into the selection of that one TPP.