Head office writes the standard. Whether it survives contact with twenty sites is a different question, and usually an unanswerable one until an audit or a mystery shopper answers it for you.
Roll out once
Build a routine as a template and attach it to every location's schedule. A change to the standard is one edit, not twenty conversations and a PDF in an email that half the sites never open.
Assign to the team, tag the site
Work assigned to a team reaches whoever is actually on shift, and each task carries the location it belongs to, so every site's copy of the routine stays separately trackable. Staff changes, cover and promotions then stop silently orphaning schedules — which is the usual reason a routine quietly stops happening at one site while nobody notices, because a task assigned to someone who left still looks assigned.
See where it's holding
Completion rates broken down by site show the shape of the problem: which locations are running the standard, which are slipping, and on which specific tasks. That turns "everyone needs to try harder" into "these three sites need help with the close".
Evidence without the travel
Requiring a photo or a reading on the items that matter means a regional manager can verify from their desk. Reviewing evidence from twenty sites takes an hour; visiting twenty sites takes a fortnight.
A rollout that survives the second month
The common failure is going too wide too fast. Twenty sites given eleven new routines in the same week will run about three of them, and the ones they drop will be the ones nobody was watching. What works better:
- One routine, every site. Usually the close, because it has the most at stake and the least supervision.
- Two weeks of just that, with completion visible daily and someone actually looking at it.
- Fix the routine, not the sites. If four locations are all skipping the same item, the item is wrong, badly worded, or asks for something impossible at 06:30.
- Then add the next routine, once the first one is running without being chased.
The second step is the one teams skip, and it's the one that decides whether this becomes how the business runs or another system people work around.
What head office should watch
Not a league table. Ranking sites against each other mostly measures how busy they are. The useful signals are movement and pattern:
- A site that has slipped against its own record: something changed there
- An item failing everywhere: the standard is wrong, not the staff
- Routines completed in suspiciously little time (the checklist is too long) to be done honestly, so it isn't being
- Sites with no failures at all, ever (nothing that can't fail is measuring) anything
Local variation where it's real
Sites differ: different equipment, different licences, different opening hours. Templates give you the shared standard; per-location schedules let a site carry the extra checks it genuinely needs without forking the whole routine.
The template trap
Templates are the mechanism that makes a shared standard possible, and the thing most likely to be abandoned by accident. It happens like this: a site needs one extra check, someone copies the template instead of extending it, and within a year there are nine variants nobody can reconcile. The next change to the standard then has to be applied nine times, so it gets applied to three, and the standard has stopped existing without anyone deciding to abandon it.
Worth agreeing early: one shared template per routine, changed centrally, with genuinely site-specific checks kept as separate schedules alongside it rather than forked into a local copy.
Where new sites start
A site opening next month is the easiest place to get this right, because there is no existing habit to displace.
Attach the standard schedules before the site opens, not after, and the routines are simply how that location has always worked. Retrofitting them onto a site that has run its own way for three years is a different and much harder conversation, so take the easy one first while the pattern is still forming.
For groups opening regularly, this turns into a short onboarding routine of its own: the set of schedules every new site gets, which removes the question of whether anyone remembered to set up the quarterly checks.