According to Digiday, Fanatics restructured its paid media program to optimize for customer lifetime value rather than standard audience segments, documenting a 19% lift in LTV across the shift. The sports merchandise platform stopped buying against demographic or interest clusters and started bidding toward the customers most likely to return and spend over time.

The mechanics: Fanatics fed historical purchase and retention data into its bidding algorithms, then directed platforms to optimize not for click, add-to-cart, or even first purchase, but for the modeled future revenue of each acquired customer. Media dollars moved away from broad sports-fan audiences and toward the signals that historically preceded repeat orders — specific product categories, browsing depth, cart composition, time-on-site patterns. The campaign logic became: pay more to acquire the customer who will come back, pay less for the one-time buyer, regardless of how well either fits a traditional persona.

This worked because Fanatics already had a clean view of its back-end economics. The brand knew which acquisition behaviors correlated with repeat purchase, which product lines drove the highest lifetime spend, and which channels delivered customers who stayed. That existing data infrastructure let the team replace guesswork audience profiles with outcome-based feedback loops. The platform could test, learn, and allocate in real time, steering budget toward the customer cohorts that historically generated the most total revenue, not the largest immediate conversion volume.

The underlying mechanism is portable. When a physical-product brand knows its repeat-purchase rate and average order sequence, it can weight media spend toward the customers who exhibit early signals of long-term value — higher first-order AOV, faster second purchase, engagement with content beyond the transaction, email open behavior, referral activity. The brand stops paying for volume and starts paying for durability. That shift requires clean cohort tracking and a willingness to sacrifice short-term conversion count for better unit economics over six or twelve months.

A small brand runs the same play by tagging customers at acquisition and sorting them by ninety-day behavior. Track which traffic sources, ad creatives, or landing-page variants produce customers who reorder within three months. Feed that data back into the media buy: increase bids on the creative or audience segment that correlates with repeat purchase, decrease or kill spend on the sources that convert once and churn. On Meta or Google, create a custom conversion event that fires on second purchase, then optimize toward that event instead of first checkout. Start with a test budget, measure the LTV delta between cohorts, then reallocate the full media plan once the signal is clean. This does not require a data science team — a spreadsheet, a UTM discipline, and a three-month patience horizon will surface the pattern.

The broader lesson: audience targeting is a proxy for outcomes, useful when you lack outcome data. Once you have enough customer history to know what a valuable customer looks like at the point of acquisition, optimize directly for that profile and let the demographics follow. The 19% lift Fanatics documented is the spread between buying the right persona and buying the right behavior.