Most agencies that say "we tried ads and they didn't work" are describing something that happened after the click, not before it. What follows: what paid ads can honestly be held responsible for, and where that responsibility ends.
A paid ad is accountable up to one moment: someone raises their hand. The speed of the reply, the questions you ask, the follow-up three weeks later when they finally pick dates — that is your operation, not Meta's. If you judge your ads on bookings, you are judging your own follow-up and calling it the ad's fault.
The booking loop spans months
So here is what you should expect from a paid account: a reasonably stable price per new conversation, a dial you can turn up or down, and a fast read on which destination or offer actually pulls. What you should not expect is a clean line from click to booking inside the same month. A McKinsey and Skift Research report finds the booking path has stretched from a funnel into a loop, with travelers now hitting roughly 65 touchpoints before booking, up from 45 in 2018, and 78% still researching after they book.
That study counts touchpoints, not calendar days, so it does not tell you how long the loop takes. My own read is the simpler one: a path with that many stops is not a path that starts and ends inside one billing cycle, and the click you paid for and the deposit you bank will usually be reported in different months. That is a bookkeeping problem, not a channel problem.
Skift's reporting on creator content describes the same shape in a different corner of travel marketing — the video goes viral and the booking comes later, which is why brands have had to build ways to measure the lift at all. Paid social behaves the same way in my experience, and that lag is not a reason to distrust the channel. It is a reason to stop reading week-one ROAS as a verdict.
Leads expire in your inbox
The other half of the clock has been measured better outside travel than inside it. Harvard Business Review published its research on this under a title that is the whole finding: online sales leads have a short life. The inquiry your ad bought is perishable inventory, and it expires in your inbox, not in the ad account.
Skift has argued that travel marketing's next KPI is economic impact rather than impressions. For an agency with four people and no analytics team, the small version of that is two numbers kept apart: what you pay for one new conversation, and what share of conversations become bookings. The first one grades the ads. The second one grades you. Collapse them into a single ROAS figure and you will eventually switch off the only part that was working.
If you judge your ads on bookings, you are judging your own follow-up and calling it the ad's fault.
Split the number before you touch the budget
Take the last 30 days and write two figures on the same line: total ad spend divided by new inquiries, and bookings divided by those same inquiries. Then sort your last 20 inquiries by minutes-to-first-reply and check whether the ones answered inside an hour closed at a different rate than the rest. If they did, your cost per booking is a response problem and raising or cutting the budget will not move it either way.
One growing travel agency we work with went from "5× monthly sales — 8 to 40+ in 90 days" — and almost none of that came from a new audience. It came from the half of the funnel that starts after the click: https://brandrstudio.com/our-work/travel-agency-ai-sales See the case →
Get this every week. One issue, one idea, sources you can check. No pitch.