Four Answers to One Question
The question came from an investor on a video call, and it was not a hard one. How many creators are you working with right now?
The founder started to answer, then stopped. Both people could hear him clicking between tabs. Fourteen seconds later he said ninety-one. His operations lead, unmuted by accident, said sixty-eight. His affiliate lead typed one hundred and six into the chat.
All three were reading real data. They were answering three different questions. The founder had counted every creator who ever received a sample. Ops had counted creators active in the last thirty days. The affiliate lead had counted unique creators, but summed across four separate exports where seventeen people appeared twice.
This operator ran four TikTok Shop storefronts — one US, one Malaysia, one Thailand, one Philippines — off a single product catalog and a single team in Shenzhen. That structure is common in the Pearl River Delta. It is also the structure that guarantees every number is contested.
This is for sellers running two to six storefronts, usually split by market rather than by brand, with a shared catalog and a shared creator team. If that describes you, the problem below is not a reporting inconvenience. It is the reason budget conversations stall.
What This Covers
This covers consolidating creator-level data across stores so that one question has one answer. It does not cover multi-store inventory sync, and it assumes each store is legitimate and separately registered. If any two of your stores violate platform rules, consolidating their data is the least of your problems — read why one store’s violation can spread first, since cross-store penalties do happen.
I will also state the limit clearly at the end. Rollups produce confident numbers that can hide real problems. Done well they are still better than four contested ones, but they are not truth.
Why Multi-Store Creator Data Refuses to Add Up
Adding two stores is not twice the work. It is a different kind of work, because four structural things change at once.
The first is that the same person is not one row. A creator can work with your Malaysia shop and your Thailand shop. In each export they look like two new people, sometimes with a slightly different display name.
The second is timing misalignment. Settlement periods differ by market. When you add-to-date GMV across four shops, you are adding four partially-settled periods that do not end on the same day.
The third is currency. Each shop reports in local currency. Converting at order date, settle date, or month-end rate gives meaningfully different totals, and nobody records which one they used.
The fourth is the one that actually breaks teams: each store develops its own definitions. One manager counts a creator as active if they posted this month. Another counts active as “has replied in 60 days.” Both reasonable, incompatible, and invisible until someone tries to combine them.
Decide What “One Creator” Means First
Before touching a spreadsheet, settle three definitions in writing. Most consolidation projects fail because this step gets skipped and every later argument traces back to it.
Define created: the creator exists in any store’s export and has not been explicitly removed. Define active: pick one window and hold it across stores — posted content with attributed orders in the last thirty days is the most usable, because it excludes both dormant and purely aspirational relationships. Define counted once: across stores, this creator is one row, with a per-store breakdown attached.
These definitions are not objectively correct. They are conventions, and the value comes entirely from everyone holding the same ones. Write them into the top of whatever file holds the rollup.
The Rollup Structure Finance Will Accept
Three levels, kept separate. Mixing them is what makes rollups untrustworthy.
The first level is source data per store. One tab per store, exported on the same schedule, never edited by hand. If someone needs to correct something, they correct it upstream or note it beside the data, not in it.
The second level is the name map. This ties one creator who shows up in several stores to one single row. It starts as a lookup keyed on creator handle. Someone keeps it up by hand at first. This work is what makes the rest hold up, and most teams skip it. It is also the exact step DAMI automates — one creator record matched across every shop you run, so the map stops depending on whoever maintained the spreadsheet.
The third level is the reporting view. Read-only, derived from the first two, and the only thing anyone quotes in a meeting. Nothing gets typed here.
| Level | Contents | Edited by hand? | Who sees it |
|---|---|---|---|
| Source per store | Raw exports, one tab per shop | Never | Ops only |
| Name map | Creator handle → one creator + shops | Sometimes, with a note | Ops only |
| Reporting view | Rolled-up metrics, read-only derived | Never | Everyone |
The discipline that matters most is that a number quoted in a meeting must trace back to a source tab in under two minutes. If it cannot, do not quote it.
Handling Cross-Store Conflicts
Once one creator appears in two stores, conflicts appear that no single-store seller ever faces.
The first is rate conflict. You offer 12% in Malaysia and 18% in Thailand, and the creator works both. Eventually they notice. Sometimes they ask for the higher rate everywhere, which is fair on their side and expensive on yours.
Decide your policy before the first conversation rather than during it. Either rates are market-specific and stated as such in every agreement, or they are standardised globally. Attempting a case-by-case blend produces inconsistent treatment that spreads through creator networks faster than any written term.
Two Conflict Types Worth Planning For
The second conflict is content cannibalisation. The same creator posts similar videos for two of your shops in adjacent markets, and the audiences partially overlap. You are now paying twice for reach that partly repeats, and your own two storefronts compete in the same comments.
The mitigation is straightforward but requires planning: assign creators primary and secondary markets explicitly, and stagger content calendars so overlapping markets do not receive near-identical videos the same week.
Third is attribution confusion. When a buyer follows a creator from your Thailand shop and purchases in Malaysia, both stores may record something. Knowing your own duplicate count before it appears in an investor conversation is far better than explaining it afterward. If you have not read how invitation and resends actually work, start there, because this is where duplicate-looking records usually come from.
What the Rollup Actually Reveals
A correct rollup shows three things no single store can show you.
The first is true creator concentration. You may believe you work with ninety creators. The rollup may show forty-three unique people, with twenty-two appearing in two or more stores. That changes how you read every per-creator average you have been quoting.
The second is cross-market performer transfer. Some creators who succeed in Malaysia also succeed in Thailand. Most do not. Knowing which pattern holds lets you expand creator relationships deliberately instead of hoping. In practice transfer rates are low — often under 30% — which argues for market-specific recruiting rather than replicating your best list everywhere.
The third is genuine cost per incremental sale. Once duplicates are removed, per-creator cost across your whole programme is often 20 to 40% higher than any single store suggested, because you were dividing spend by double-counted creators.
Where Rollups Create False Confidence
This is the part worth being cautious about.
A clean rollup makes every number feel authoritative, including the ones built on definitions you chose arbitrarily. “Active in thirty days” produces one count. “Sixty days” produces a completely different one. Neither is wrong, and the rollup does not mark which you used.
Rollups also launder timing problems into single figures. Combining four partially-settled periods into one GMV number looks rigorous. It is not. Label every combined revenue figure with its settlement status.
Finally, a rollup cannot tell you whether activity in two markets represents two opportunities or one opportunity counted twice. That judgement needs someone looking at actual creator overlap, not just numbers.
Making It Stick
The rollup’s real enemy is maintenance, not design. Most teams build one, update it for three weeks, then quietly return to asking three people in a meeting.
Two habits make it survive. Assign one named owner whose job is updating it, not everyone’s job and therefore nobody’s. And set the update rhythm to match a decision you actually make — if your budget meeting is monthly, update monthly, not weekly.
The Honest Ceiling
Doing this by hand works up to about four shops and a few hundred creators. Past that, the name matching grows faster than its worth. Matching one creator across six shops with mismatched names is hard work by hand.
That ceiling does not mean your rollup failed. It means the job changed. Keeping track of who each creator is across shops belongs in a system, not in one person’s head.
“We rebuilt the master list three times in one quarter,” a Dongguan operator running five storefronts told me. “Every rebuild came out different, and we never knew why. That was when I stopped trusting our own numbers, and it was the most expensive thing that happened all year.” The lesson is not that rollups are bad. It is that losing trust in your own data costs more than the data was worth.
DAMI keeps one view of every creator across all your shops, with shop data and team overview built in. That drops the hand-matching step that breaks most rollups. If your team rebuilds the list by hand each month, see how combined reporting works on the product page.
You can also create your account and run one month of your own four-store data through it before changing how your team works.
Frequently Asked Questions
How many unique creators do we really have across four stores?
Usually far fewer than the sum of your exports. Count every row that maps to the same creator once, then keep per-store detail attached. If the total drops by more than 30%, your previous per-creator averages were all wrong.
Should commission rates be the same across every market?
Pick one policy and state it in every agreement. Either rates are market-specific by design, or standardised globally. Blending case by case produces inconsistent treatment that spreads through creator networks quickly.
Our four stores settle on different schedules. How do we roll that up?
Do not silently add partially-settled periods. Label any combined revenue figure with its settlement status, and reconcile to payment receipts once a month rather than trusting summed GMV.
Do creators who succeed in one market usually succeed in another?
Often not. Cross-market transfer rates below 30% are common, which argues for recruiting market by market rather than replicating your best list everywhere.
When is a manual rollup no longer workable?
Around four stores and a few hundred creators. Past that, resolving one creator across six stores with inconsistent handles becomes unreliable by hand. The mapping should move into a system that holds it.
