How do we tackle a piece of investigation in an incremental manner? It’s much easier when we’re delivering a tangible thing—we stand up at our show and tell, and say “Last time we showed you we could do X, now we can show you we can also do Y.” But for investigation work it’s not so obvious.
The kind of work I’m talking about might be what the GDS model calls the discovery phase, or it might be a technical investigation, or perhaps a research exercise around how to launch a product.
We could timebox the research and choose to make some decision at the end of that. That works if the timebox is short, because we don’t lose much if the results are inconclusive or unsatisfactory. But what happens if we need more control? If the time needed is lengthy then we’ll want to track progress with the opportunity to steer things as needed. Ultimately we want to deliver value incrementally.
An approach I’ve used successfully in the past is based on the principle of reducing uncertainty. My friends at IndigoBlue call this kind of thing “uncertainty management”. We can start the investigation acknowledging we have a lot of uncertainty, and continue by reducing that week by week. My general approach goes like this…
- First we consider all the questions we want to answer.
- Then looking across all those questions we ask: What is the most pressing question we need to tackle, and what can we do in the next iteration that will best reduce that uncertainty? That gives us a short term to-do list, or the highest priority activity in our backlog.
- Next we work at reducing that uncertainty until it is no longer the most pressing question. It’s important to recognise we’re not trying to answer the question comprehensively. We’re just trying to reduce the uncertainty sufficiently that it’s no longer the most pressing issue.
- At that point we repeat the process.
I think of this process as a version of the Theory of Constraints, which is about operational improvement. That tells us to keep working away at the biggest bottleneck until it’s no longer the biggest bottleneck, and then repeat.
The format of our show and tell is now clearer. We say “Last time we were unsure about X. Now we are much more confident that Y.” We have demonstrated progress, reduced uncertainty, and we are delivering value incrementally. This gives everyone a lot more control in terms of the time we use, and we can (for example) stop the process early if we think we’ve answered our questions sufficiently to move on.