How to Build Audience Cohorts for RetailMedia Campaigns

‍

Most retail media campaigns start the same way. Someone opens the network's audience menu, picks "category buyers," maybe layers on "competitor buyers," and hits launch.

That's not targeting. That's a starting point.

Retail media's biggest advantage is that it's built on real purchase behavior. But that advantage only pays off if you organize shoppers into cohorts that each have a specific job in your plan. One giant audience gets one average message and one blended result that tells you almost nothing.

Here's how to build cohorts that actually earn their budget.

‍

What a cohort really is

‍

A cohort is a group of shoppers who share a behavior that matters to your business, and who should be treated differently because of it.

The key word is "treated." If two groups would get the same message, the same bid and the same goal, they don't need to be separate cohorts. If they'd get different treatment, they do.

‍

Start with your growth question

‍

Before you touch the audience builder, answer one question: where is this brand's next sale most likely to come from at this retailer?

For a new product, the answer is trial from category buyers. For a mature brand losing share, it might be winning back lapsed buyers. For a premium brand, it might be trading shoppers up from the value tier. Your growth question decides which cohorts matter most and how budget should be split.

‍

The core cohorts most brands need

‍

New-to-category. Shoppers who don't buy the category yet but show adjacent behavior. Expensive to convert, but essential for long-term growth in categories that need to recruit new households.

Category buyers, not brand buyers. Your bread and butter for growth. They already buy the category; they just pick someone else.

Competitive loyalists. Shoppers heavily committed to a specific competitor. Worth targeting when you have a clear reason to switch, like a new flavor, a better price per unit or a stronger claim.

Lapsed brand buyers. People who bought you before and drifted. Often the most efficient cohort to reactivate.

Current brand buyers. Your loyal base. Usually best served with low-cost reminders, cross-sell and new item news, not heavy trial spend.

Cross-sell cohorts. Buyers of complementary products or your other brands. Great for portfolio brands and occasion-based selling.

‍

Get specific with the definitions

‍

Vague definitions create vague results. For each cohort, decide on the lookback window (bought in the last 3 months? 12 months?), the frequency threshold (one purchase or repeat buyers?) and the exclusions (do current brand buyers get removed from your trial cohort?).

Write it down. Use the same definitions across campaigns and, as much as each network allows, across retailers. That's what lets you compare results over time.

‍

Match message, placement and goal to each cohort

‍

Each cohort should get its own creative angle and success metric.

Category buyers might see a "try something new" message on search and product page placements, judged on new-to-brand sales. Competitive loyalists might see a head-to-head reason to switch, judged on conversion and incremental sales. Lapsed buyers might get a "new and improved" or limited-time offer, judged on reactivation. Loyal buyers might get a simple reminder timed to their purchase cycle, judged on efficiency, with tight budget caps.

When every cohort is judged on the same ROAS target, the loyal cohort always "wins" and gets more budget, because those shoppers were going to buy anyway. Different jobs need different scorecards.

‍

Mind the size

‍

The more you slice, the smaller each cohort gets. Some networks have minimum audience sizes, and tiny cohorts can produce noisy, unreliable results.

If a cohort is too small at one retailer, merge it with a similar one or broaden the lookback window. Precision is great until there's nobody left to reach.

‍

Refresh and learn

‍

Shoppers move between cohorts constantly. A category buyer becomes a trial buyer, then a repeat buyer, then sometimes a lapsed buyer. Refresh cohorts regularly and track how many shoppers move in the direction you want.

Over time, compare cohort performance across retailers and seasons. You'll learn which cohorts deserve more investment and which are just eating budget.

‍

A quick scenario

‍

A yogurt brand has been running one retail media campaign against "yogurt buyers" at a major grocer. Reported ROAS is solid, but share isn't moving.

The team rebuilds the plan around four cohorts: category buyers who haven't bought the brand in 12 months, buyers of one specific competitor's Greek yogurt line, lapsed brand buyers and current loyal buyers. Loyal buyers move to a small, always-on reminder budget. Most spend shifts to the first two cohorts with a taste-and-protein message. The reported ROAS drops a little, but new-to-brand sales climb, and that's the number that was actually holding back share.

‍

The bottom line

‍

Cohorts turn retail media from a single blunt instrument into a set of precise tools. Start with your growth question, define cohorts clearly, give each one its own message and scorecard, and keep refreshing.

Five Eighty builds and manages retail media cohort strategies across dozens of networks. If your campaigns are still aimed at "category buyers" and hoping for the best, let's sharpen the plan.