A garment manufacturer outside Ho Chi Minh City sells through TikTok Shop into Vietnam and keeps between ten and fifty creators active at any moment. Its affiliate team is two people. Every month the same thing happens: plan seats empty out, someone opens the creator database, filters by apparel and follower count, and copies a fresh list into a spreadsheet.
The next month the spreadsheet is stale and the process repeats. Twelve months in, nobody can answer a basic question — which of last year’s fifty apparel creators actually produced orders, and are they still in the list?
This article is about replacing that monthly rebuild with a circle: a saved set of filters plus a source pool that keeps matching new qualifying creators, and that the whole team can reuse instead of reinventing. It assumes you are running targeted plans and doing your own outreach. It does not cover affiliate managers who work from an open plan, where the platform matches creators on its own without you choosing anyone.
One caveat at the start. A circle solves repetition and institutional memory. It does not solve taste. If nobody on your team can tell a good apparel creator from a bad one, automating the search just produces a bad list faster.
Why Lists Go Stale Faster Than You Think
A static creator list decays for reasons that have nothing to do with anyone’s effort.
The first is creator behaviour. People stop posting, switch categories, pivot to a different platform, or get an account restricted. A list built in March can easily contain twenty percent dead accounts by June, and those dead entries are invisible until you try to contact them.
The second is your own business. The manufacturer above added a plus-size line in the middle of the year. The apparel filter that served them perfectly for six months had no concept of size-inclusive content, so every new match was wrong in a way nobody noticed for two full cycles.
The third is staff turnover, and it is the expensive one. When the two-person affiliate team changes over, the private knowledge of “these fifteen creators reply quickly, those twenty never do” walks out with whoever leaves. The replacement starts from a blank filter.
A Circle Is Filters Plus a Source
A creator circle — sometimes called a saved segment or a reusable pool — is two things working together.
The filter set is the criteria: category, platform behaviour, follower band, content type preference. The platform’s follower bands run 0-10k, 10k-20k, 20k-25k, 25k-50k, 50k-100k and 100k+, and content format can be filtered to shoppable video, shoppable live, or no stated preference. Getting these right matters more than anything else in this article.
The source is where qualifying creators come from, and it should include people you already have a relationship with. Strong circles pull from every creator who has previously requested a sample or completed an order for you, not only from cold database searches. Those are people who have already signalled intent, which changes response rates substantially.
| Source | What it’s good for | What it gets wrong |
|---|---|---|
| Cold database search | Volume and breadth when entering a new category | No prior interest; expect the lowest reply rate |
| Past sample requesters | Warm intent and proven product interest | Small pool; dries up if you stop sampling |
| Previously fulfilled creators | Known output quality and reliable contact | Biased toward whoever you happened to pick |
| Shared team pool | Preserves the team’s judgement over time | Only as good as the labelling discipline behind it |
That fourth row is the one that compounds. How SMBs Filter the Right Creators From a Creator Database goes deeper on building the filter criteria themselves, which is where most of the value actually sits.
Setting Filters That Exclude
The mistake nearly every team makes at this stage is building filters that *include*. Category equals apparel, followers between 10k and 100k, content preference shoppable video. That returns thousands of people, most of whom are useless to you.
Useful filters exclude. The filter set that works for a Vietnam apparel brand with two affiliate staff looks closer to this:
Exclude anyone whose last post is older than fourteen days. This single condition removes more dead weight than anything else. It is a recency proxy for “this person is still working.”
Exclude anyone outside your shipping footprint. A creator whose audience is overwhelmingly outside Vietnam will generate views you cannot convert, because you cannot ship economically to where their viewers are.
Exclude follower bands you cannot service. If your sample budget supports fifty creators a month, filtering to 100k+ produces aspirational matches that will ignore a message from an unfamiliar brand. The 10k-50k bands are usually where a 10-50 creator programme actually converts.
Include content preference deliberately. Filtering to shoppable video excludes creators who only do live selling, which is a real and legitimate choice — just make it explicitly rather than by accident.
The last thing to add is an exclusion list rather than a filter: everyone who has ignored two or more outreach attempts. Nothing else protects your daily invite volume from being spent on people who have already declined you twice.
Letting It Run on a Schedule
A circle only saves real work if the matching and the outreach run without someone remembering to do it. Scheduled tasks handle this: once per week is the common setting for advertisers, with the timing chosen to match when your audience is most likely to check messages.
Volume discipline matters here more than frequency. Sending invites faster than your team can respond to replies is worse than sending fewer, because a reply that goes unanswered for three days tells the creator the brand is not serious. If two people can process maybe thirty conversations a week, then thirty new contacts a week is the ceiling regardless of how many names the circle can produce.
There is also a platform-level limit worth respecting: plan groups are configured at 10, 20, 30, 40, 50 or 100 creators, and invite intervals sit between one and five seconds. Neither is a hard constraint most teams hit, but both shape what “automatic” can reasonably mean per day.
If you are running several stores rather than one, a shared pool prevents the classic failure where each store independently invites the same creator and the brands end up bidding against themselves. TikTok Shop Batch Invite Tool for Vietnam Multi-Store Apparel Brands in 2026 covers that coordination problem specifically.
Where DAMI Fits
DAMI’s creator circle is the piece that makes this repeatable rather than monthly-rebuilt. You define the conditions — category, platform, follower band, content type — and it keeps matching qualifying creators, including ones who have already requested samples or completed orders for your store, so the circle refills without anyone opening a database and filtering again.
The part that matters for a team of two is that the circle is shared, not personal. Both people see the same list, the same filters, and the same history, which is what stops the “nobody knows which creators worked last year” problem from recurring every time somebody goes on leave.
Pair it with a scheduled invite task and the monthly rebuild disappears: the circle produces candidates, the task contacts them on a fixed rhythm, and the two affiliate staff spend their time on replies rather than on list-building.
Keeping it Honest
Three failure modes to watch.
The circle slowly drifts toward your best guess rather than your best data. Every few months, look at which circle members actually generated orders and adjust the filters to match reality. A filter that has not been touched in six months is almost certainly optimised for a version of your business that no longer exists.
Shared pools decay without labelling. If nobody marks why a creator was added or rejected, the shared list becomes an unlabelled pile and the next person ignores it. Two minutes of labelling per creator is what makes the pool worth sharing.
Automatic does not mean unattended. The weekly matching is monitoring-free in the sense that it runs, but somebody still has to read replies within a day or two. The teams that get value from automation are the ones whose response discipline was already good before they automated anything.
Frequently Asked Questions
How many creators should be in a circle?
Bigger than your active plan, because you want slack, but not so big that nobody reviews it. For a 10-50 creator programme somewhere between 150 and 400 steady candidates usually works — enough to refill several groups, small enough to audit monthly.
Should the circle include cold creators or only warm ones?
Both, but keep them identifiable as separate sources. Cold database matches give you volume; past samplers and fulfilled creators give you conversion. Mixing them without tags makes it impossible to tell which is carrying the programme.
How often should the filters be revised?
Whenever your product mix or shipping footprint changes, and at minimum quarterly. Changing categories is the trigger most teams forget — the manufacturer’s plus-size line should have caused a filter review the week it launched.
Will automated matching waste my invite volume?
It will if you match aggressively and exclude weakly. Recency filters and a suppression list for repeat non-responders are what keep automated outreach from spending your daily allowance on dead accounts.
What if every store on my team uses the same circle?
That is the point, provided you coordinate. One shared circle with visibility into who contacted whom prevents two stores inviting the same creator in the same week, which creators notice and which makes both brands look disorganised.
If you want to see whether a shared circle fits the way your team works, the DAMI product page walks through the automation side, and you can register for a DAMI account to test it against one of your existing segments.
