A Romanian fitness creator with 180,000 followers looks like an obvious pick for a Romanian supplement brand. Then the campaign runs, the engagement looks healthy, and almost nothing converts. The post-mortem finds the problem: roughly two thirds of that audience is in Italy and Spain — Romanian speakers who emigrated, still following creators from home, and in no position to order from a Bucharest warehouse.
Nobody lied. The creator is Romanian. The content is in Romanian. The follower count is real and the engagement is real. The brief simply matched on the wrong variable.
Two numbers that get treated as one
Creator location is where the person is based, or says they are. It comes from a profile bio, a geotag, a language setting, or a self-declared country in a media kit.
Audience market is where the followers actually are. It is a distribution, not a value: 61% Romania, 14% Italy, 9% Spain, 6% Germany, the rest scattered.
Most discovery tools let you filter by country. Very few make clear which of these two the filter means, and the default is usually creator location, because it is far cheaper to determine. One profile field versus an inference across tens of thousands of accounts.
For a US brand searching US creators, the distinction rarely bites. The domestic audience is so large that creator and audience location mostly coincide. The gap opens in exactly the markets Lyren is built for.
Why they diverge in smaller markets
Diaspora. Romania has one of the largest emigrant populations in Europe relative to its size. A creator producing Romanian-language content has a natural, large, and highly engaged audience outside Romania. That audience is genuinely valuable — to a money-transfer service, an airline, a diaspora-facing retailer. It is worth very little to a brand whose product ships domestically or whose stores are in Cluj.
Language reach beyond borders. Romanian is spoken across the border in Moldova. German content reaches Austria and Switzerland. This is not diaspora; it is a language market that does not match a legal or logistics market.
The algorithm does not respect borders. Short-form recommendation systems optimise for watch time, not geography. A creator can go viral in a country they have never visited, and the follower base that arrives never leaves.
Expat and relocated creators. Someone who moved to Berlin three years ago may still hold an audience that is overwhelmingly back home, or may have rebuilt an audience that is overwhelmingly local. The bio says Berlin either way.
Why platform-reported location is not the answer either
Even when a tool does look at audience location, the underlying signal is weaker than the confident percentage in the interface suggests.
- Creator-facing analytics report audience country, but only the creator sees them. Anything a third-party tool shows you was inferred, not read.
- Inference typically works from a sample of followers or commenters, then extrapolates. Sampling error on a long tail is substantial.
- Follower profiles are sparse. Many accounts have no location, no bio, and no posts.
- Language is a proxy for country, and a bad one in precisely the diaspora case that matters most.
None of that makes the estimate useless. It makes it an estimate, and it should be presented as one. Treat a reported "68% Romania" as a range, not a fact, and be more suspicious of round numbers and of small accounts where the sample is thin.
Checking the gap yourself
Ten minutes per creator, no tooling required. Do it on your final shortlist, not on the long list — it belongs at the scoring step of the research workflow, after the mechanical filters have already cut the list down.
- Read the comments, not the count. Open the last five posts and read fifty comments. What language are they in? For a Romanian creator, a wall of Italian is the finding.
- Check who the creator collaborates with. Sponsored posts reveal which brands already believe in this audience. If the paid partnerships are with international money-transfer or telecom brands, someone has already concluded this is a diaspora audience.
- Look at the timing of engagement. Comments clustering in a window that matches Central European evening rather than Romanian evening is a one-hour signal, weak on its own but confirmatory.
- Ask for the screenshot. On a real shortlist, request the creator's own audience-country breakdown from their native analytics. Serious creators have it. A refusal, or a "roughly all Romania", is itself information.
- Compare across their platforms. A creator whose TikTok is pan-regional and whose Instagram is domestic is common. You may be buying the wrong platform rather than the wrong creator.
When a mismatch is exactly what you want
This is not a rule that foreign audiences are bad. It is a rule that the audience should match the brief.
- Tourism boards want the opposite: a Romanian creator whose audience is largely German is precisely who sells a Transylvania trip to Germans.
- Diaspora-facing services — remittances, telecom, shipping, legal services — need that Italian and Spanish share.
- Export brands testing a new market want creators whose audience already sits in the target country, whoever they are.
- Recruitment campaigns for return migration want the diaspora deliberately.
In every one of those cases the brief should say so, and the filter should be set accordingly. The failure is not foreign audience. The failure is not knowing. It is the single sharpest version of a wider point about running a brief in the Romanian market, where creator location and audience market come apart more often than they line up.
Worth separating from a related question: an audience in the wrong country is not the same problem as an audience that is not real. A creator can pass this check and still fail the authenticity checks, and the reverse is more common than people expect — a completely genuine audience, in a country that cannot buy the product.
What to put in the brief
Replace "Romanian influencers, fitness, 50k+" with something that can actually be checked:
Fitness and nutrition creators whose audience is at least 60% in Romania, 20,000 to 150,000 followers, engagement at or above the niche median, active on Instagram or TikTok in the last 30 days. Creator base of operations is not a requirement.
That last sentence is the one that changes results. It opens the shortlist to a creator living in Vienna whose audience is 80% Romanian, and closes it to one living in Bucharest whose audience is not. Both of those are the correct outcome, and neither is what a creator-location filter would have given you.
How Lyren handles it
Lyren treats audience market as the primary discovery filter and stores creator location separately, as an optional secondary one. That choice is why the two are never silently collapsed into a single "country" dropdown.
The inference is exactly that — an inference, from weighted signals rather than a single field — and it is shown as a distribution rather than a verdict, so you can see when a creator is 55/45 and treat them with the appropriate caution. When the signal is thin, the honest answer is that the signal is thin.