The "searching raises the price" myth, and what's actually moving the fare
Clear your cookies, open a private window, never search the same route twice — we've all heard the advice. Here's what's really moving the fare between two searches, and why it isn't that.
Someone will eventually ask why we bother recording fares over weeks when they could just outsmart the search page by browsing in an incognito window. The idea is that airlines track you, notice you circling back to the same flight, and quietly raise the price to push you toward buying before it climbs further.
We don't do that, and we haven't found evidence anyone else reliably does either. But the fare on a route really does move between two searches an hour apart, sometimes by enough to notice, and that's real enough that the folk explanation for it survives every airline denial. Worth being specific about what's actually going on.
What we can rule out
We poll routes on a fixed schedule, from a small number of servers, with no browser fingerprint, no cookies, no account. If personalized price-raising by search behavior were a real mechanism, our own polling would be the easiest target for it — the same request, repeated, from the same infrastructure, indefinitely. It isn't. A route we've checked daily for weeks shows fares moving in both directions, landing inside the distribution we've already recorded for that cell rather than climbing toward whatever the last price happened to be.
That's not a controlled experiment. It's just months of unglamorous polling that would show a search-triggered ratchet if there were one to find.
What's actually happening
Airline fares are sold in buckets, not as one price per flight. A given route and date usually has a handful of seats at the lowest advertised fare, more at the next bucket up, and so on. As the cheap bucket sells — to anyone, not to you specifically — the system reprices the flight to the next one. Search twice and get a higher number the second time, and the more likely explanation is that seats moved in the cheap bucket between your two searches, not that the system remembers you were looking.
Add to that the fact that fares are served by systems that cache and refresh on their own schedule, sometimes pull from different feeds depending on time of day, and get repriced continuously by revenue management software running for reasons that have nothing to do with any individual shopper. A fare that's different an hour later is closer to a stock quote than a price tag on a shelf.
None of that is dramatic, and none of it is personal — which is where the incognito-mode advice gets it backwards. It treats a process that's always running as if it were a reaction to being watched.
Why this matters for a watch, not just a search
This is really the same reasoning behind why we won't score a route without enough history: if fare movement looks noisy from a handful of searches, the fix isn't to search more carefully, it's to stop treating any single search as the signal. One fare is one data point. What tells you whether $412 is worth booking is where it sits against dozens of fares already recorded for that same route, month and booking window — the same idea behind what "26% below typical" means.
Clearing your cookies changes nothing about which fare bucket is currently open on that flight. Watching the route for a few weeks tells you whether the price in front of you right now is unusually good or unusually ordinary — a more useful question than whether searching moved it.
What we'd do differently if the myth were true
If personalized price-raising were real and measurable, it would leave a specific fingerprint: fares climbing for repeat visitors while staying flat or falling for fresh ones, on the same flight, at the same moment. We don't have visibility into anyone else's individual search sessions, so we can't rule that out with total certainty. What we can say is that our own recorded history — the same routes, watched by the same infrastructure, day after day — doesn't look like a system reacting to being checked. It looks like the ordinary churn of seat buckets selling and revenue management doing its job. The full mechanics of how we turn that churn into a score, rather than a search habit, are on how the score works.