Feedback loops

Each level learns from the level below and from the customer.

01What it's for

The system works because every level learns from the level below and from the customer. Team increments feed the Sprint Review, the ART increment feeds the System Demo, and customer evidence feeds Portfolio Sync. Every level runs the same five events (Refine, Plan, Sync, Review and Retrospect), scaling Scrum's pillars of transparency, inspection and adaptation.

The cadence sets the rhythm: day and two weeks for teams, week, four weeks and quarter for the ART and the portfolio.

02What it turns into under business as usual

The flow becomes one-way and top-down: strategy, plan, execution, report. Feedback travels upward only as status, and decisions at the top are not changed by what teams learn.

03How you notice

Portfolio decisions don't change when the evidence does.
The loops have no owner and feed no decision.
What is measured is output, not outcome.

04How to make it work

Give every loop evidence, an owner and the decision it feeds.

Decide as late as responsibly possible, at the lowest level where the best information is.

Escalate only what cannot be decided there, and send the decision back down with the authority to act on it.

05Feedback in refinement

Feedback loops do not start once something is built. The refinement flow on the left of the map, from Epic to Feature to Story Refinement, is a feedback loop too, and the cheapest one: feedback on an idea costs far less than feedback on an increment.

So refine with your teams, stakeholders and customers as early as possible, and let their real feedback shape the work before it competes for capacity. Not refinement on their behalf followed by a handover.

Facilitate it with just enough structure: a Wardley map to see how the value chain is evolving and where to invest, an impact map to connect the goal to actors and the behaviour change you need, and a story map to slice the work into small steps towards the MVP.