A behind-the-scenes look at an e-commerce startup's successful ad campaign on Facebook, powered by Facebook's opaque AI ad-targeting algorithm
The platform is so good at “microtargeting” that many small e-commerce companies barely even bother advertising anywhere else. Tweets: @bluegrasscap and @stevendilla Tweets: Bluegrass Capital / @bluegrasscap : “In the start-up-industrial complex, its like a systematic transfer of money from VC firms to start-ups to Facebook” http://www.nytimes.com/... Steven Dilla / @stevendilla : As of last year, on the top one million websites in the world: Google had trackers on 76%, Facebook on 23%. http://www.nytimes.com/...
Context & Ripple Effects
This 2017 profile lands mid-arc in the story of how digital advertising consolidated. Stratechery had already mapped the Facebook-Google dominance of digital ads a year earlier, and this piece adds the demand-side mechanism: an AI targeting system so effective that small e-commerce companies barely advertise anywhere else.
The quotes frame the economics — Bluegrass Capital's 'systematic transfer of money from VC firms to start-ups to Facebook' — while Steven Dilla's tracker data (Google on 76% of top sites, Facebook on 23%) shows the reach gap that underpins it. Later reporting would stress-test the black box itself: the NYT's own targeted-ad experiment exposed what advertisers think they know about users, and unsealed court documents revealed Facebook managers questioning whether their own targeting claims misled advertisers.
First-order effects
- Small e-commerce companies concentrate their ad budgets almost entirely on Facebook, making the platform's opaque algorithm their de facto growth engine and single point of dependency.
Second-order effects
- VC-backed startups' fundraising effectively subsidizes Facebook's revenue, as investor capital cycles through startup ad spend into the platform — a dynamic competitors like Google must counter with equivalent automated targeting tools.
Third-order effects
- The same microtargeting machinery migrates beyond commerce into politics — Trump's campaign planned to exploit FB's automatic ad-spend distribution again in 2020 — while advertiser dependence on an unaccountable algorithm builds the case for the scrutiny and market diversification (Amazon, Apple) that later reshaped the ad landscape.
The trend: Digital advertising is consolidating around platforms whose AI targeting is simultaneously the product, the moat, and the black box that advertisers cannot audit.