Retargeting without being creepy
Following people around the internet with the same ad for three months is why people hate advertising. Here's the version that works.
Retargeting shows ads to people who already visited your site. Done well, it's among the most efficient spend available. Done badly, it's the thing that makes people install ad blockers.
The difference is mostly restraint.
Why it works at all
Someone who visited your site has already shown interest. They're not a stranger — they had a reason to be there.
Most people don't act on the first visit. They're comparing, or busy, or not quite ready. Retargeting keeps you visible during that gap.
Because the audience is small and pre-qualified, cost per conversion is usually far better than cold advertising. It's also why retargeting numbers can look misleadingly good — some of those people would have come back anyway.
The mistakes that make it awful
No frequency cap. The default is often unlimited. Someone glances at your site and sees your ad forty times in a week. That's not persuasion, it's harassment, and it actively damages how they feel about you.
Cap it. Three to five impressions per week is plenty.
No time limit. Someone who visited four months ago has moved on. Membership windows should match your actual buying cycle — a week or two for quick decisions, up to a couple of months for considered purchases.
Same ad forever. If they didn't respond the first five times, the sixth won't work either. Different message, different angle.
Retargeting people who already converted. Showing "get a free quote" ads to someone who became a customer last week is a visible sign nobody is paying attention. Exclude converters — always.
Retargeting bounces. Someone who landed and left in three seconds was in the wrong place. Build your audience from people who spent real time or viewed multiple pages.
The version that works
Segment by what they looked at. Someone who read the drain cleaning page should see drain cleaning ads, not a generic company ad. You already know what they wanted — use it.
Segment by depth. Someone who visited the contact page and didn't submit is very different from someone who read one article. The first deserves a direct offer; the second deserves something softer.
Change the message. They've seen your pitch. Give them the next thing: a specific proof point, an objection handled, a reason to act now. Retargeting is a second conversation, not a repeat of the first.
Cap it, and let people age out. Both protect the relationship and neither costs you anything real.
The honest limitations
Attribution flatters it. Retargeting takes credit for people already returning. If your platform says retargeting drove 40 percent of conversions, some of that would have happened anyway. The only way to know is to turn it off for a period and see what changes.
Audiences are shrinking. Privacy features, blockers and cookie restrictions have made retargeting audiences smaller and less accurate. It still works — it just doesn't reach everyone it used to.
It needs traffic. With two hundred visitors a month, your retargeting audience is too small to run meaningfully. Fix traffic first.
A reasonable setup
Exclude converters. Build an audience of people who spent over thirty seconds or viewed two pages. Set a thirty-day window and a five-per-week cap. Run two or three different creatives addressing different objections. Check monthly whether the audience is large enough to bother with.
That's most of the value, with none of the behaviour that makes people resent you.
Want to know how your own site measures up?
Run a free audit. Our crawler checks your live pages for the technical and on-page issues covered in articles like this one, and tells you what to fix first. Takes about a minute.