Long-term plans are just different

A plan for today is more straightforward than one for next year. Uncertainty grows exponentially over time. The closer we are to a task, the clearer it becomes.

That makes short-term plans remarkably different from their long-term brothers.

Yet most teams seem to treat long-term plans as if they were just an extension. They’ll write User Stories, put an estimate on them and map those Stories over future iterations. “We plan to do that Sharepoint integration in Sprint 18“. These preliminary, flexible roadmaps are harmless until someone takes them as a commitment.

Most teams are well-versed in short-term planning. They get together every other week and discuss what they will build. While accidents can happen and short-term plans might change, the near future is pretty stable.

Long-term plans are a different beast altogether. We can’t be sure what the market will look like six months from now. We build software in a feedback-driven way and don’t know what we will learn over time. The distant future is somewhat chaotic.

Planning past a few weeks requires different tools.

Short-term plans are the realm of estimates. “We believe it will take our team the next three weeks to build this feature.” Even with a margin for error and the odd sick day, mature teams can be reasonably confident in their short-term estimates.

Estimates may rot over time, but when they’re fresh, they are pretty tasty.

Long-term plans shouldn’t rely on decaying estimates. What takes our team three weeks to build today might take a week in the far future. Or a month. In some cases, it might require a complete overhaul of a part of the codebase that doesn’t exist yet. Assumptions we make about our team today might not hold in the future. Our superstar data scientist who could build it in a week might be working for the competition by then.

Instead of thinking in features, long-term plans require us to think in more general categories of problems or goals. “We’ll focus on Improved Calendar Experience in the first half of August.

By not going into details, we keep the scope flexible while still being able to make long-term promises to stakeholders. When August comes around, our short-term planning will be the answer to the question:

What is the best Calendar Experience we can build in the coming two weeks?

The power of this approach is that it allows us to make long-term promises. Our salespeople can go out and tell customers that the clunky calendar interface will definitely be better by September. We can stick to those promises without going into crunch mode.

We can build feedback-driven products with a vision by keeping our short-term plan detailed and our long-term plan high-level.