The audience is already ideal
When a user posts about your product, the audience arrives pre-filtered: the poster has proven they're your customer, and the people who follow them tend to be like them. Ad platforms spend billions approximating a filter the follow graph already is.
Ad networks spend billions building complex machinery to guess who your customers are. The social graph does it for free. Messages naturally drop into the hands of peers who share the poster’s exact problems, tools, and scale.
Who posts? The users who value it most
An earned tier's posts come from people who use the product enough to want more of it — that's what earning means. Nobody writes a public post about a tool they opened once. So the population of posters is a double filter over your user base: real users, selected again for the ones getting enough value to say so with their own name attached. Posting is a signal nobody can fake or buy: it costs the poster something to give, so it means something. Your ideal customer isn't a profile you wrote in a planning doc; it's whoever keeps posting.
Who reads it? People like the poster
Homophily — the tendency of people to connect with people like themselves. Named in sociology in the 1950s, and the reason a follow graph is already segmented.
Follow graphs are built out of shared work, shared problems, and shared taste — a founder's replies are full of founders, a designer's timeline is designers. So a post from your customer lands, by construction, on a cluster of people doing similar work with similar problems: the population most likely to contain your next customer. No interest categories, no lookalike modeling, no pixel — the graph did the segmentation years before you showed up.
What ad targeting is actually for
Lookalike audiences are an industry-scale attempt to reconstruct exactly this: start from your customers, find people statistically like them, pay to appear in front of them. The reconstruction is expensive because the platform owns the graph and rents it back. A customer's own post skips the reconstruction — it travels the real graph, carried by a name the audience already chose to follow, in a voice they already trust more than an ad-shaped anything.
Why the post beats the ad even at equal reach
Two posts can reach the same hundred people and do different work. The ad is discounted on sight; disclosure rules exist precisely because a recommendation from a person carries weight an ad doesn't — which is why the earned tier requires the disclosure rather than hiding it, and why no reward turns on conversions: the post stays a post. The reader sees a real person they follow, saying something true about a tool they use, with the relationship on the label.
What this means for the dials
It reframes what you're buying when you price a post. The reward isn't for reach in general; it's for reach into a cluster that took your ad budget years to approximate. And it's why the reply bonus counts distinct replies rather than views: replies from the poster's graph are your ICP talking about your product in public, which is the asset the whole program exists to produce.