Reading a regional rollout
A staged rollout gives a brand data from markets that started at different times. Aggregating them produces a line that blends early markets at maturity with late markets at launch, and the aggregate trend then reflects the rollout schedule rather than the market response.
The correct reading is a cohort reading, and the correct comparison is stage against stage.
What does the aggregate hide?
Whether later markets are performing like earlier ones. An aggregate that keeps rising as markets are added will rise even if every new market is performing worse than the last, because addition dominates.
That is the failure mode: a rollout that is decaying in quality while growing in total, with nobody seeing it until expansion stops.
How should it be read?
On a common axis of time since launch, one line per market or market group, with the entry date noted. Overlaying the curves tests whether later entrants behave like earlier ones.
If the curves overlay, the rollout is consistent. If they fan downward, something is changing — the markets, the execution, or the competitive response.
What legitimately differs between markets?
Market size, competitive structure, seasonality at entry, and the quality of the local execution. Each can explain a lower curve without implying the proposition is weakening.
Which is why the comparison needs the possible explanations tested rather than asserted. A downward fan is a question, not a conclusion.
How does seasonality interfere?
A market launched into a strong season shows a flattering early curve, and one launched into a weak season shows the reverse. Comparing curves from different seasonal entry points without adjustment compares seasons.
Either group markets by entry season or adjust each curve for the market's own seasonal pattern. The second is more work and more honest.
What should be measured at each stage?
Distribution build, velocity per stocking store, and the category response. Those three separate the rollout's execution from the proposition's performance, which is exactly what an aggregate cannot do.
Distribution build in particular is an execution measure. A market with slow build has a field problem rather than a demand problem, and the two need different responses.
When should the rollout be paused?
When later markets consistently underperform earlier ones at the same stage and the difference is not explained. That is the signal the aggregate will not give, and it arrives several markets earlier than the aggregate turn.
Agreeing that criterion before the rollout starts is what makes acting on it possible.
What should be reported to leadership?
The cohort chart, the market count by stage, and a single sentence about whether later markets match earlier ones. That is the whole story and it fits on one page.
Frequently asked questions
How many markets are needed to see a pattern?
Enough that one market's idiosyncrasy does not drive the conclusion, which usually means several per cohort. With very few markets, the honest statement is that the pattern is not yet distinguishable.
Should markets be grouped?
Usually yes, by entry period, because individual market curves are noisy. Grouping trades detail for stability and it is generally the right trade.
What about markets that were chosen first for a reason?
That is the most common confound: first markets are chosen because they are favourable, so a downward fan is partly expected. Stating the selection logic alongside the chart lets a reader discount appropriately.
Does this apply to a channel rollout?
Yes, and to any staged expansion — stores, accounts, regions. The technique is about entry timing rather than about geography.
How long should each cohort be followed?
Until the curve flattens, which for a listing is usually several months. Comparing cohorts before any of them has flattened compares ramps rather than outcomes.
What if execution differed between markets?
Then the comparison measures execution as well as proposition, and the honest report says so. Compliance data per market is what allows the two to be separated.
Should the pause criterion be numeric?
It should be specific enough to act on, which usually means a threshold on velocity at a stated stage. A vague criterion becomes a discussion at exactly the moment a decision is needed.
How does competitive response fit in?
Later markets may face a competitor who has now noticed, which is a legitimate explanation for a lower curve and an important one commercially. It should be checked rather than assumed in either direction.
What is the single chart to build?
Velocity per stocking store by months since launch, one line per entry cohort, with counts. It answers the rollout question directly and it is usually available from data already held.
Should the criterion to pause be shared with the field team?
Yes. A team that knows the threshold understands what the programme is being judged on, and it removes the suspicion that a pause was a budget decision dressed as a performance one. Channel context is collected by NACS.
### What should a brand actually do with this?
Build the cohort chart from the first market rather than when somebody asks for it. A rollout monitored from the start has a comparison available at every stage; one analysed retrospectively has to reconstruct entry dates from memory.
Report it at the same cadence as the rollout decisions. A cohort chart produced quarterly for a monthly decision cycle arrives after the decisions it was meant to inform.
Working with retail data? See NRS Insights data solutions or contact NRS Insights.