If you own a Black-owned trades, professional services, or wellness business, there's a cost baked into your Google and Facebook advertising that a demographically unremarkable competitor simply doesn't pay. Call it the race tax. It isn't a line item anywhere, and no platform will ever show it to you as a number, but it's real, it's measurable, and it comes from two facts colliding: people tend to trust and buy from businesses that look like them, and the ad platforms you have to use won't let you target the audience most likely to extend you that trust.
This is a breakdown of where that cost actually comes from, so you can see the mechanism clearly before we get into how to work around it.
The central mechanism
Where the race tax enters the advertising funnel
The added cost appears when a human trust pattern meets a platform targeting restriction.
Consumer behavior
People tend to trust businesses that feel familiar
Platform policy
Google and Meta prohibit direct race-based targeting
The highest-trust audience becomes harder and more expensive to reach
More discovery spend, slower learning, and a higher acquisition cost
The Trust Gap Starts Before the Ad Even Runs
The tendency to trust people who share your identity has a name in social science: homophily, or in-group bias. It's one of the most consistently replicated findings in the research on trust and cooperation, and it shows up specifically in purchasing behavior. A study on consumer bias and minority founders found that consumers impose a real transaction cost on minority-led businesses, in some cases expecting a price discount just to overcome the trust gap, a penalty that has nothing to do with the quality of the product or service. Research on Black consumers specifically has found that perceived racial congruity, meaning whether a brand, spokesperson, or business owner is seen as sharing the customer's identity, measurably affects how credible and trustworthy that business appears before a single word of a pitch is heard.
This same dynamic shows up downstream in the signals that drive search and paid advertising. Research on Airbnb found that Black hosts priced identical listings roughly 12% lower than non-Black hosts, a direct financial concession to close a trust gap that had nothing to do with the properties themselves. Research on Yelp's "Black-owned" business tag found that once a business was identified as Black-owned, its average rating dropped, and a separate national study found businesses in Black-majority zip codes received measurably lower ratings and far fewer reviews than comparable businesses elsewhere, despite no underlying difference in quality. All of it points to the same conclusion: trust in a purchasing decision is not race-neutral, and it costs Black-owned businesses more to earn it.
Evidence across marketplaces
Trust signals change when identity becomes visible
Different studies, platforms, and measures point toward the same underlying purchasing pattern.
Airbnb
12% lower
Comparable listing prices from Black hosts
Yelp
Ratings declined
After Black ownership became visible
National research
Weaker review signals
Lower ratings and fewer reviews in Black-majority areas
Normally, the answer to a trust gap in advertising is targeting. Reach the audience most predisposed to trust you, prove yourself there, and expand outward. That's the standard playbook. It's also the exact playbook Black-owned businesses are locked out of.
You Can't Target the Audience Most Likely to Trust You
Meta and Google do not let advertisers target ads by race, and Meta's history with this specific issue is well documented. Facebook allowed advertisers to target and exclude users by "ethnic affinity" until a 2016 ProPublica investigation showed the feature could be used to keep housing ads away from Black and Hispanic users. After a 2019 discrimination settlement tied to housing, employment, and credit ads, Facebook restricted targeting further, and in August 2020 it eliminated "multicultural affinity" targeting entirely, the last option that let advertisers reach users based on an inferred connection to African American, Hispanic, or Asian American culture. Google has never offered race or ethnicity as a targeting dimension at all.
How targeting changed
The path from ethnic-affinity targeting to today’s restrictions
2016
“Ethnic affinity” exposed
ProPublica showed that Facebook targeting could be used to exclude Black and Hispanic users.
2019
Restrictions expanded
A discrimination settlement involving housing, employment, and credit ads led to tighter controls.
2020
Multicultural affinity removed
Meta eliminated the final targeting option based on inferred cultural affinity.
Today
No direct race targeting
Race and ethnicity cannot be selected as direct targeting criteria on Google or Meta.
The practical result is straightforward: you cannot tell Google or Meta to show your ad to Black consumers, even though the research above says that's statistically your highest-trust, highest-converting audience. Every other advertiser gets a version of this same limitation, but for a business whose highest-trust audience happens to align with the general market, the restriction costs them nothing. It only becomes a real constraint when your highest-trust audience is a demographic slice the platform won't let you dial in on directly.
The Population Math Makes It Worse
Even if targeting weren't restricted, the pool you'd be targeting is small. Black Americans make up roughly 14% of the U.S. population, around 48 million people nationally, and a much smaller share than that in most individual metro areas outside a handful of Southern and urban markets. A competitor whose highest-trust audience overlaps with the general population is, by default, targeting something close to the full addressable market. A Black-owned business chasing its highest-trust audience is chasing a fraction of that market, and it can't even chase it directly.
The population constraint
About 14 out of every 100 Americans are Black
The available audience can be substantially smaller inside an individual service area.
14%
Approximately 48 million people nationally
A national share does not equal a local addressable market. In many metro areas, the reachable high-trust audience represents a much smaller pool.
Paying Twice: How Proxy Targeting and Lookalike Audiences Add Cost
Locked out of direct targeting, advertisers fall back on proxies: interest categories, behavioral signals, geographic concentration, page engagement, anything that correlates with the audience they actually want without naming it explicitly. Proxies are noisier and less precise than direct targeting by definition, which means more of the budget goes toward impressions and clicks that were never going to convert, just to accumulate enough real customer data to work with.
Same tools, different starting points
The targeting restriction does not cost every advertiser equally
The difference is how closely broad targeting overlaps with each advertiser’s highest-trust buyers.
General-market advertiser
Broad addressable audience
High overlap with highest-trust buyers
Efficient early campaign learning
Black-owned advertiser
Smaller highest-trust audience
No direct targeting option
Proxy targeting and slower learning
That real customer data becomes the seed for a lookalike or similar audience, the tool both Meta and Google offer to expand beyond an advertiser's existing customer base by finding new users who resemble it. Lookalike modeling is only as good as the seed data feeding it, and building a seed audience large and clean enough to model well costs real spend before it produces a single qualified lead. This is where the funding gap facing Black-owned businesses compounds the problem directly: automated bidding and lookalike systems both need volume to calibrate, and a smaller ad budget takes longer to generate the data those systems need to stabilize. A business with less capital to put behind that seeding phase pays more, for longer, before its cost per lead settles anywhere close to a demographically unremarkable competitor's.
Run the comparison side by side. One advertiser's highest-trust audience is close enough to the general population that broad targeting finds it efficiently from day one. The other has to spend to find a thin slice of the market through imprecise proxies, spend again to build a seed audience large enough to model, and wait longer for the algorithm to stabilize on a smaller budget. Same platforms, same bidding tools, same skill in building the campaign. Different starting cost.
Paying twice
One cost finds the audience. The next teaches the algorithm.
Both stages consume budget before the campaign reaches a stable, efficient cost per lead.
Cost 1: Finding the audience
Proxy signals
Interests, geography, behavior, and engagement
Noisier delivery
More impressions reach people who were unlikely to convert
Discovery spend
Budget identifies real customers through trial and error
Cost 2: Training the algorithm
Initial customers
Real conversion data becomes the starting signal
Seed audience
The platform needs enough clean examples to model
Learning spend
More conversions are needed before delivery stabilizes
More spend before cost per lead stabilizes
The Gap Compounds Once the Ad Is Seen
None of this even accounts for what happens after the ad is shown. Click-through rate is the biggest driver of Google's Quality Score, and Quality Score sets cost per click directly, advertisers at the low end of the range can pay several times more per click than a top-scoring competitor for the identical keyword and position. If the same trust bias that makes targeting harder also makes a consumer less likely to click once they do see the ad, whether because of a business name, a thinner review profile, or any other identity signal, that lower click-through rate pushes Quality Score down and cost per click up on top of the targeting cost already described. The two effects stack: it costs more to reach the right person, and once reached, the odds they act are lower for the same underlying reason. Both SEO and PPC carry this same trust penalty, which is why the cost per lead gap shows up in both channels rather than one.
Quality Score pressure
Weaker click signals move cost in the wrong direction
Advertisers at the low end of the Quality Score range can pay several times more per click, so this visual shows the direction of the cost difference without implying a universal multiplier.
Stronger Quality Score
Lower relative CPC
Stronger expected click-through signal
Weaker Quality Score
Higher relative CPC
Weaker expected click-through signal
We Build Campaigns Knowing This Going In
EraBright is a Black-owned agency, and we understand this dynamic from both sides of it, as a business navigating it in our own advertising, and as the team running SEO, Local SEO, and Google Ads for Black-owned trades, dental, legal, and wellness businesses navigating it in theirs.
That shapes how we build a campaign from the first decision, not the last one. It's the difference between setting a targeting strategy against generic benchmarks and building one that accounts for the fact that the audience most likely to trust a given business isn't one Google or Meta will let you target directly, and has to be found through slower, more deliberate proxy and seed-audience work instead. It's the difference between treating review generation as a nice-to-have and treating it as one of the few trust signals a business fully controls once the ad or the listing is actually seen. We don't build campaigns assuming a level playing field, because the mechanics of these platforms say there isn't one, and we don't think a business owner should have to find that out through an inflated ad bill.
