Use this article to choose how to roll out Ingrid: go live in every market at once (big bang), or pilot one market and roll the other markets out in waves (phased).
This article is for retailers planning an Ingrid implementation across more than one market. Both Ingrid rollouts are valid choices, and neither is the safe default. The right rollout follows your resourcing, your risk appetite and your testing capacity. You form a view together with your Ingrid implementation consultant during the planning workshop.
Big bang rollout: all markets at once
In a big bang Ingrid rollout, you have one go-live across every market on one date: one team, one cutover. The table below shows how a big bang rollout compares on time to value, resourcing and risk:
Time to value | Fastest. Every market is live from day one. |
Resourcing | Heavy peak. Every market is tested and supported in parallel. |
Risk profile | Concentrated. Launch issues hit your whole footprint at once. |
Best when | Your markets are similar, your team is strong, and the launch window is tight. |
Phased rollout: market by market
In a phased Ingrid rollout, you pilot one market, then roll the other markets out in waves, applying what each wave teaches you. The table below shows how a phased rollout compares on time to value, resourcing and risk:
Time to value | Slower in full, but your pilot market is live early. |
Resourcing | Smoother. The same team moves from wave to wave. |
Risk profile | Contained. Issues stay inside the pilot market before they scale. |
Best when | Your markets differ, or your team prefers to learn before scaling. |
Testing capacity often decides between big bang and phased
In practice, the choice between a big bang and a phased Ingrid rollout is usually settled by testing capacity rather than by strategy.
Many teams test their priority markets first, launch those markets, and let the live run inform the next wave. If your testing is resourced for two markets rather than nine, that has effectively chosen a phased rollout for you, and it is better to recognize that during planning than to discover it during a test cycle.
Related
Implementation: the five phases of an Ingrid implementation