How a retargeting audience actually gets built
A tag sits on your site — the Meta pixel, the Google Ads tag, or a GA4 property wired to Google Ads. When someone loads a page it drops an identifier and fires an event. You then write a rule: anyone who hit the pricing page in the last 30 days, or anyone who added to cart and didn't buy.
The platform holds that list, refreshes it as people qualify and expire, and serves your ads to it. You never see individuals — you see a count, often rounded or hidden until it clears a minimum.
Not all lists come from a pixel. Uploaded customer lists, video viewers, Instagram profile visitors and lead form openers work the same way, and they're usually the healthier list, because they don't depend on a browser cookie surviving.
Membership duration: the window is the strategy
Membership duration is how long someone stays on the list after their visit. Most accounts leave it at the default and then wonder why their retargeting reaches people who bought three months ago.
| Window | What it's good for | What goes wrong |
|---|---|---|
| 1–7 days | Cart and checkout abandoners. Highest intent you'll ever buy. | The audience is tiny. On low traffic it may never clear the platform minimum. |
| 8–30 days | The workhorse for considered purchases — pricing page visitors, demo page bouncers. | Fine. This is where most budgets should sit. |
| 31–180 days | Long B2B cycles, high-ticket categories, brand recall. | You're paying to reach people who bought from someone else six weeks ago. Exclude converters. |
| 181–540 days | Winback and renewal, Google Ads only. | Treat it as a separate campaign with different creative. Same ad, ninth month, is just annoying. |
The traffic floor nobody quotes you
This is the part that gets left out of the pitch. Platforms enforce minimum audience sizes before they'll deliver — Google needs on the order of a hundred active users for Display and around a thousand for Search, and Meta needs a base of roughly a hundred people before an audience serves at all. Check the current figure inside your own account, because these move.
Now the arithmetic. A site with 1,000 visitors a month, on a 30-day window, has at best 1,000 people on the list. Strip out cookie loss, ad blockers and declined consent and you're realistically at half that. Split it by page — cart abandoners, pricing viewers, blog readers — and each segment is a few dozen people.
At that size retargeting isn't a channel, it's a rounding error. Put the money into prospecting or search intent until the traffic exists to retarget — usually where SEO or a well-built search campaign earns its place first.
Frequency caps and the point it turns hostile
A small audience and a normal budget produce the same result every time: one person sees your ad eleven times a day. Cap frequency at campaign level — two or three impressions a day is a sensible ceiling — and learn the signature of fatigue: frequency rising, click-through rate falling, cost per thousand impressions climbing.
The fix is more creative, not more budget. Three or four variants in rotation, and a rule that converters drop off the list immediately. Nothing dents a brand quite like chasing a customer around the internet with an ad for the thing they already bought.
Why your lists shrink on their own
Browser policy quietly changed what a retargeting list can be. Safari's tracking prevention caps script-set first-party cookies at seven days, so a 30-day window behaves like a seven-day one for a chunk of Indian iPhone traffic. Firefox blocks third-party cookies outright. Ad blockers stop the pixel firing at all, and consent banners take another slice.
Server-side tagging and conversion APIs recover some of it by sending events from your server rather than the browser. They don't restore the list to what it would have been in 2018, and anyone claiming otherwise is selling implementation. Plan for smaller lists than your organic traffic suggests, and keep your UTM parameters clean so you can measure what you do capture.