How the toggle cycle works
A feature toggle lets you choose when to adopt new or changed behaviour. While a toggle is optional, you decide when to switch it on, test it and prepare your users. At its expiry date the choice disappears: the new behaviour becomes standard, and the toggle can no longer be switched off.
In recent years, a typical release has seen around 70 to 80 toggles expire. Most of them don't touch the way your company uses M3. The work is finding the few that do, in time.
Why reading the release notes is not enough
Release notes describe what a toggle changes in general terms. They can't tell you whether the situation it changes actually occurs in your company: whether you use that order type, that allocation method, that kind of delivery. That depends on your configuration and on your data.
- A toggle that changes how replenishment deliveries are prioritised matters only if you have replenishment deliveries.
- A toggle that changes a credit check matters only for customers near their limit.
- A toggle in a process you don't run can be safely ignored.
A practical approach, toggle by toggle
- 1List the toggles that expire with the coming release, with their expiry dates.
- 2For each one, write down in plain words what it changes: which check, default or step behaves differently.
- 3Find where in your processes that change lands, and which teams work there.
- 4Check your own data: does the situation the toggle changes occur? How often, and in which records?
- 5Give every toggle a verdict: it affects you, it doesn't occur in your data today, or you can't tell yet.
- 6For the toggles that affect you, decide what to do before the expiry: test, adjust a setting, inform the team.
"Not found in our data" is only true on the day you checked. Your data keeps changing, so record the date with every verdict and check again shortly before the expiry.
Test what matters, not everything
Retesting every process twice a year is expensive and rarely thorough. With a verdict per toggle you can focus: test the processes where an affecting toggle lands, using records that are in the situation it changes. That is where surprises would come from.
Plan it as a recurring cycle
Releases come twice a year, every year. Teams that handle them calmly treat release readiness as a routine with an owner, a checklist and a fixed moment before each expiry, not as a project that starts when users report that something behaves differently.
How M3CloudSlam helpsInfor release readiness →Release Toggle Impact gives a verdict per toggle, checked against your own data.