What you will take away
- Start from the value chain, not from the system inventory.
- Score initiatives by cost of delay, not by how loud the complaints are.
- The data layer is usually the real blocker behind three other projects.
Why transformation programmes stall
They stall because they are scoped as a programme. A three-year plan to modernise everything produces eighteen months of design, a reorganisation, and a delivery date that arrives after the assumptions have expired. Meanwhile the warehouse team is still rekeying orders from email into a spreadsheet.
The alternative is not lower ambition. It is a different unit of work: a sequence of quarter-sized changes, each useful on its own and each making the next one cheaper.
Step one: map the value chain, not the systems
Draw how work actually moves from a customer request to a delivered outcome and cash collected. Mark every point where a human moves data between systems, where a decision waits on a report, and where an error is discovered late. That map is where the money is, and it rarely matches the org chart.
Do this by watching, not by asking. The documented process and the real process diverge in exactly the places that matter, because the workarounds are where people are absorbing the cost of the system.
Step two: score by cost of delay
For each candidate, estimate what one more quarter of the current state costs in hours, errors, lost revenue or risk exposure, then divide by a rough delivery estimate. The ranking that falls out is usually different from the one produced by whoever complains most persuasively.
- What does not doing this cost per quarter, in money or hours?
- How many other initiatives are blocked until this is done?
- What is the confidence in the estimate, and what would raise it?
- Can it ship in one quarter, or does it need to be cut smaller?
The right first project is usually the boring one that three other projects are waiting on.
Step three: fix the data layer early
In most of the assessments we run, the blocker is not the user interface. It is that customer identity lives in four systems with no shared key, so any integration, report or automation starts with a reconciliation problem. That work is unglamorous and it unlocks everything downstream.
Agree the system of record for each core entity, build the integration path once, and stop new systems from creating a fifth copy. This is the piece that makes the following two years cheaper.
Step four: ship, measure, then re-plan
Set the baseline metric before the first change - hours per order, error rate, days to close, whatever the value map identified. Ship the first slice within a quarter, measure against the baseline, and publish the result including the parts that did not work.
Then re-plan. A roadmap that is not revised after each delivery is a document, not a plan. The organisations that get through transformation successfully are the ones that treat the sequence as a hypothesis they keep testing.
OrbixEdge Engineering
The engineers who build and run the systems we write about - platform, cloud and product teams shipping for clients in retail, health, logistics and fintech.
