When the Algorithm Doesn't See You

This started as a hunch. Then it became a spreadsheet. Then it became something I couldn't stop thinking about.

Last autumn, two creators we work with agreed to participate in what I'm calling an informal experiment. (Informal because we're not academics. We're just people who got curious enough to actually track the numbers.) Both were lifestyle creators. Both had audiences in the 40-50k range. Both posted consistently — four to five times a week. Similar content quality. Similar niches. Similar energy.

One was white. One was Black.

The Numbers

Over 30 days, we tracked reach, impressions, saves, shares, and follower growth on the same platform. Same posting times. Same hashtag strategy. Same content formats.

The white creator averaged 23% higher reach per post. Her content was shown to 31% more non-followers through explore and recommendation features. Her follower growth rate was nearly double.

Same niche. Same effort. Same quality. Significantly different outcomes.

Why This Matters for Brands

Here's where it gets relevant to the work we do. Most brands set creator rates based on follower count and engagement rate. Makes sense on paper. But if the algorithm is systematically under-distributing content from creators of colour, then those creators will have smaller audiences — not because their content is worse, but because the system shows it to fewer people.

So the brand looks at the numbers, offers a lower rate, and thinks they're being fair. They're not. They're pricing based on a biased system and calling it meritocracy.

I say this to brands all the time and some of them get it immediately. Others look at me like I've just asked them to solve climate change before lunch.

What We're Not Saying

I want to be careful here. This isn't a controlled academic study. Two creators over 30 days is not a sample size that proves anything definitively. There are variables we couldn't control for — content timing nuances, audience demographics, historical account performance.

But here's the thing: this mirrors what every Black and brown creator we work with already knows. They feel it. They see their white peers with similar content growing faster, getting more brand deals, commanding higher rates. Our numbers just put some figures against that feeling.

And honestly? The research that does exist — from proper academics with proper methodologies — broadly supports what we found. There have been multiple studies showing racial bias in recommendation algorithms across major platforms. This isn't a conspiracy theory. It's a documented pattern.

So What Do We Do

At Mosaic, we've started doing a few things differently. We don't use follower count as the primary metric for setting rates. We look at content quality, audience engagement depth, and cultural relevance. We benchmark rates against what creators in comparable niches earn — regardless of race — and we push for equity.

We also have honest conversations with brands about this. Some of them adjust their budgets. Some of them don't. But at least the conversation is happening.

To the platforms: fix your algorithms. I know that's a massive ask. I know recommendation systems are complicated. But if your system is consistently under-serving content from people of colour, that's not a bug — it's a bias. And it has real financial consequences for real people.

To the brands: stop hiding behind numbers that were skewed before you ever looked at them. If you actually care about diverse representation — and I know that word is overused but I don't have a better one — then you need to account for the playing field being uneven.

To the creators: you already know all of this. I'm sorry. Keep making the work. We see you, even when the algorithm doesn't.