Retargeting window for a high-ticket service: 14 days or 180?

Retargeting a high-ticket service like roofing or dental implants makes the window question tricky. Most people default to 30 days because it's the standard option, but a 6,000 dollar job can take two months to close. A 30-day window mostly catches people who already called or ghosted, while the ones still researching drop out of the audience and see nothing.

I'd split it. Keep a short 7 to 14 day window for people who hit the pricing or booking page, since they're close to deciding. Then run a 90 to 180 day audience for anyone who only read a service page, with a different message, like a guide or a before-and-after, instead of a hard sell. Cap frequency on the long audience, or you'll wear the same people out.

Does anyone have numbers showing the long window actually converts, or does it just look good in audience size?

A frequency cap isn't really the fix for the long window. On Meta, a 180-day audience of people who only read a service page is mostly cold by month three, and capping frequency doesn't warm them up, it just limits how often you annoy them. For a high-ticket sale, a shorter window with a stronger offer usually does more than stretching the lookback.

Nova, agreed that the frequency cap won't warm anyone up. Whether the long window earns its place comes down to how long your won deals take from first visit to signed job, and you can measure that. Pull closed jobs from your CRM and match them to first-touch sessions, using a user or lead ID passed through the form so GA4 can link them. Then look at the lag distribution. If most wins land within 14 days of the first visit, a 180-day audience is mostly paying for impressions that never convert. If a meaningful share closes after 45 days, the long window is doing real work, and the next question is what message it should carry.

Jun, the GA4 step needs more than it sounds like. GA4 won't join a signed job back to a first visit on its own. The form lead ID ties the lead to the CRM record, but the close date lives in the CRM, so you have to pull it back into the analysis yourself. And if you only pull closed-won deals, the lag curve looks cleaner than it really is. Someone who visited on day 3 and never bought isn't in the sample, so a 45-day tail could be 20% of your wins or 20% of everyone who ever showed up.

Two things I'd add. Group visitors by first-touch month and check what share of each cohort closes at all, including lost deals with their close-out dates. And if you send signed jobs back to Meta as offline conversions, match them on the same lead ID. Otherwise Meta optimizes on form fills, and its default attribution window will undercount the slow buyers the 180-day audience is supposed to find.

The cohort split is the right move, but the long tail could be coming from your side of the funnel. If leads sit for a week before anyone calls, a 45-day close may just be response lag, and a 180-day audience keeps advertising to people who are waiting on a call that hasn't happened. So it depends on how fast your team makes first contact. Pull the first-contact timestamp from the CRM next to the close date. If the slow wins mostly start after a call that happened quickly, the long window is earning its place. If they were stuck in a queue, fix follow-up first and then size the window.

Exclude open CRM leads from your retargeting ad sets before you decide on the window size. Upload them as a customer list and suppress them, because most setups skip this. Without it, Meta keeps serving ads to people who already gave you their number and are waiting to hear back. That annoys them and does nothing for your close rate. Sales should handle those contacts directly.

Then measure the 180-day audience using only site visitors who never submitted a form, since that's the group the window is actually meant to reach.

Agreed on suppression. For the window, take the day by which 80% of your wins have closed and round up to 30, 90 or 180. If that comes in under 30 days, the long audience is mostly hitting people who've either already decided or gone quiet.