Great question — so for me strategy has many layers. It starts at the top but every department, team, etc also have a strategy which needs to align and compliment each other. As you go down the ladder your time frame starts to become shorter.

John Cutler has a great way of framing things, he uses what he calls 1–3s. This is thinking in 1–3 hours, 1–3 days, 1–3 weeks, months, years, decades.

If you apply this thinking about how this “propergates” across the org — your org strategy sits at that 1–3 years/1–3 decades realm. Each department sits at the 1–3months/1–3 years realm and likewise with your products. Then as things break down further you end up with your user stories and day to day tasks.

I’m actually not a fan of cascading or “top-down”, I prefer alignment and “lose-coupling”.

What I mean by that is that you don’t create the org strategy then take pieces of that strategy and tell each department that’s what they need to do, and etc. Rather when you go down the aligned-autonomy path each area looks to decide on their own way of how they’re going to compliment each other (yes there’s a sequencing that needs to occur but I’ve seen orgs where they did both concurrently and then met in the middle and refined both ends, getting input from both ends — sounds messy but can work well!)

I’m actually going through a round of building out strategies right now with a client. And to give you a picture of how we’ve been doing it over the past couple of weeks.

1. Org strategy refinement, not a huge amount to change there because it’s a long timeframe and most still stood relevant. This strategy doc is lightweight and simply in the form of a one pager with top level OKRs and guiding pillars for the organisation.
2. Each department comes up with their own strategy ‘one pager’ — time frame is 1–2 years and this is how they are going to help achieve the org strategy. Again top level OKRs which for some are leading indicators to the org OKRs.
3. Each department ‘one pager’ then breaks down into about 6 different pages as they refine and articulate each of their strategic goals in more detail. This starts to cover not just the “where we will play” and “how we’ll know we’ve won” but it starts to articulate the “how we will win” part of the question. This is now when strategy starts to merge into more details specifics — I.e. the “Big bets” that you’re intending on making to help realise that strategy.
4. Next is each Product Strategy. Depending on the product it may align and compliment at the top — the org strategy — or it will align to a department. Each Product Strategy is also a ‘one pager’, much shorter in timeframe <1 year. It includes OKRs and your roadmap. Again this is getting to the next level of detail. Breaking down the “big bets” into smaller pieces of the puzzle that now sit on a roadmap. If you think about what I talked about in the article this would be where you start to define your outcomes and have an opporutnity backlog aligned to it.

Hope that makes sense and answers your question :)

I also wrote this last year which is on a similar topic which might also help illustrate things:

Alignment through OKR’s and Hypotheses