The fastest way to mess up a multi-branch loyalty rollout is to launch all branches on day one. The second-fastest is to launch each branch with a separate database. Here's a 4-step rollout that keeps customer data clean and lets you compare branch performance side-by-side from week one.
Step 1 — Pick the lead branch
One branch goes live first. Pick the busiest one — yes, the busy one, not the slow one. The slow branch will give you 60 sign-ups in 30 days, which is too few to learn anything from. The busy branch gives you 600, which is enough to spot what's broken in the cashier flow before you put it on the other two.
Run the lead branch alone for at least 14 days. Track the metrics from the first 30 days playbook. Don't tell the other branches anything beyond "we're testing something at Branch A — your turn next month."
Step 2 — One workspace, one customer database
Every branch shares the same member list. A customer who signs up at BGC should be recognised at Greenbelt the next day, with the same balance, the same tier, the same earned stamps. Two branches with two databases creates "two of every customer," and it's almost impossible to merge cleanly later.
Practical implication: when you set up branches in the system, they should be branches OF a workspace, not separate workspaces. The branch ID lives on each transaction (so you can attribute revenue), but the customer is one record. If the system you're using doesn't support this, switch before you launch anywhere — fixing the data model post-launch costs more than the migration.
Step 3 — Train each branch on the lead branch's numbers
When you bring on Branch B, do it after Branch A has 30 days of data. The training session is concrete: "These are the actual sign-up rates, the actual cashier scripts, the actual reward redemption pattern from Branch A. Here's what worked. Here's what didn't."
Cashiers learn faster from real shop numbers than from abstract training decks. Bring two cashiers from Branch A to Branch B for the launch shift if logistics allow. They explain the script in real-life terms, the new branch picks it up in 2 hours instead of 2 weeks.
Step 4 — Per-branch QR posters, central data
Each branch should have its own sign-up QR poster pointing at a branch-tagged URL (e.g. scaleplusrewards.com/j/your-shop?branch=bgc). Why: when a customer signs up via the BGC poster, that branch gets credit in your dashboard analytics. The customer record is still in the central database — branch attribution is just a column on the transaction.
This becomes meaningful at month 3. You'll see that one branch's poster outperforms another's by 3×, or that one branch has a strong week-2 return rate while another doesn't. Without per-branch attribution you'd see "the program is working" or "not working" across the whole business and miss the actual signal.
The mistakes we see most often
- Launching all branches simultaneously. Looks coordinated. Actually means none of them get the operational attention they need.
- Creating a separate program per branch "for clarity." Customers travel between branches. They don't want two cards.
- Different rewards per branch. "BGC gives a free latte at 9 stamps, Greenbelt gives a free pastry at 7." Confusion compounds. Pick one program. Run it everywhere.
- Not tagging the cashier. Per-cashier attribution (not just per-branch) lets you spot the one staff member who isn't running the script. Without it, the underperforming branch looks like a branch problem when it's actually a Tuesday-3pm-cashier problem.
What this looks like in real numbers
One Filipino specialty café we work with rolled out across 3 branches over 60 days using exactly this sequence. Snapshot at day 90:
| Branch | Members | Sign-up rate | R2 (week-2 return) |
|---|---|---|---|
| BGC (Branch A — lead) | 1,840 | 34% | 44% |
| Greenbelt (Branch B — day 30) | 1,210 | 31% | 41% |
| Tomas Morato (Branch C — day 60) | 620 | 27% | 38% |
The numbers converge. Branch C, opened last, ended at slightly lower rates — but it took only 30 days to get within 5 points of Branch A's mature rate, instead of the 90 days it took A to reach that rate originally. The lead-branch playbook compresses the learning curve.
This month
- Pick the lead branch. Confirm it's the busy one, not the comfortable one.
- Verify your loyalty system supports one workspace + multiple branches before you sign up.
- Generate a unique per-branch QR for each location.
- Schedule the second branch's launch for 30 days after the first, third for 60 days after the first.
Multi-branch is where most loyalty programs accidentally double their workload. Doing it sequentially with shared data means each launch is easier than the last, instead of harder.