Imagine that at any given time there is work happening across a series of time-scales. I use a basic system of 1s and 3s. It looks like this:
Every 1-3 hour bet connects to a 1-3 day bet. Every 1-3 day bet connects to a 1-3 week bet and so on -- all the way up to 1-3 years and 1-3 decades. The ranges are broad enough not to get people hung up with estimates, but narrow enough to reason about. And they are all connected (very important).
Team member should be able to trace their 1-3 hour work chunks all the way up to the company's 1-3 year bets. Without skipping steps.
Rough, Contrived Example:
1-3h: Research how we've been storing transaction types
1-3d: Adding new transaction type autocomplete logic
1-3w: Less error-prone transaction classification
1-3m: Improve worfklow efficiency for bookkeepers (>15 person teams, high transaction volume, multi-currency)
1-3q: Addressing challenges of larger bookkeeping teams
1-3y: Expand sales opportunities with larger, multi-national orgs
We could layer in things like assumptions, facts, measurement, and more persona information. You can get way more detailed. A 1-3w bet will likely have many “sub” bets, etc. The key idea is to explore the story and connections.
It is also a handy slicing forcing function. You can typically get smaller AND bigger (important given that many teams are so focused on “the sprint” that they lose sight of the bigger picture).
In doing this exercise with many teams, I've come to see how often teams bump into a "messy middle" problem. Work in the near term is clear (people come into work and want something to do). Work in the long-term fits on a slide in an executive's deck. Turns out making slides is easy.
It is the messy middle bets/missions in the 1-3 month and 1-3 quarter range that are far less coherent.
Give it a try...