How Voyscout works

How it works - in 4 steps

1. Set up your profile. Tell us where you fly from, how many of you travel, where you want to go and on what budget. We scan exactly for that - not generically.
2. Real prices from day one. We sweep flights and packages on our own schedule (flights daily to once every two weeks depending on the destination, packages daily) and convert each offer into the true cost for your profile - for a family that's airport transfer, food and bags, for a solo traveller the bare fare without pointless add-ons. You see the first real prices within 24 h of the first scan.
3. We build the price history. Each tracked destination gets its own price series. Measurement by measurement it gets clearer what is genuinely cheap for it and what only looks cheap.
4. The verdict matures with data. On a new route we first honestly say WATCH - we're collecting data. Judgement then matures in stages: WAIT after ~a week of daily scans, BUY only after roughly two, because a cheap price on a short series is mostly luck. We count recorded prices, not days (WAIT from roughly 8, BUY from roughly 14), so on routes we scan less often than daily it takes correspondingly longer. A mature WATCH honestly stays where today's price sits mid-history with no clear edge. The rule is transparent: a verdict says where today's price sits in that route's own price history (percentile) and which way it's moving (trend) - and we show the confidence level in each offer's detail. And always: a verdict is statistics over observed prices, not a prophecy - we never fake confidence the data doesn't support.

In a nutshell (as of 10 July 2026):

• Cheapest direct return flight to Bulgaria for a family of 2 + 2: 5,450 CZK (Katowice → Varna, 21 Aug - 4 Sep).
• Median of all 81 scanned date and airport combinations to Bulgaria: 28,682 CZK per family.
• The same week (depart 31 July, return 14 August) cost 13,717 CZK from Katowice, 32,465 CZK from Prague and 47,916 CZK from Vienna.
• Cheapest all-inclusive catalogue price, 7 nights, family of 2 + 2 (August-September 2026): Turkey Side ~27,600 CZK vs Egypt Hurghada ~38,400 CZK.

Indicative observations from our own scans - details in How we watch flight prices and Egypt or Turkey?.

These are numbers from one measurement, not from today. The series have kept running since - current prices and their history are on the price pages.

What does Voyscout do?

Voyscout continuously scans the prices of DIY flights and all-inclusive packages - for your profile: where you fly from, how many of you travel, and where you're headed. From what it collects it builds each route's own price history and ranks offers by real value, not the lowest catalogue number.

Where does the data come from, and how often do we scan?

We scan on our own schedule (flights daily to once every two weeks depending on the destination, packages daily), not in real time. The point isn't to catch every second-by-second move, but to see how a price develops over time - whether today's offer is cheap against that route's own history. Prices come straight from the providers; for packages we also verify the cheapest offer on its detail page, not just the listing, so the price a verdict stands on is actually bookable.

What do the BUY, WAIT and WATCH verdicts mean?

Every watched offer gets one of three verdicts:

• BUY - the price is at its own historical floor, the trend isn't rising, and there's enough history for it to mean something.
• WAIT - the price sits in the pricier part of its own history, or is trending up.
• WATCH - too little data or too noisy a signal for a confident call. We'd rather keep watching than rush a "buy".

A verdict is statistics over observed price history - not a guarantee the price won't drop further. Shaky data (stale, unverified, a different travelling party) never becomes a confident BUY. We break it down in How we read price history.

How often did a BUY verdict go wrong?

We replay our verdicts over our own price history. For every day the rule said BUY on a flight series, we check whether the price of that same series dropped by more than 3% within the next 14 days. If it did, waiting would have been cheaper - and we count that BUY as wrong.

After a BUY verdict the price dropped within 14 days in 52% of cases (238 of 461 graded days, 31 flight series, as of Nov 4, 2026, pooled across parties 2+2, 1+0). For comparison: after a WAIT verdict it dropped in 79% of cases (935 of 1183; as of Nov 4, 2026, parties 2+2, 1+0).

The last price seen in that window, against the verdict day, had a median change of +1.8% after BUY and -1.3% after WAIT (as of Nov 4, 2026, parties 2+2, 1+0).

The cheapest price of a series is not always the same flight, though - departure and return dates shift from day to day. Following only the exact dates that were cheapest on the verdict day, the price dropped after BUY in 34% of cases (146 of 426; as of Nov 4, 2026, parties 2+2, 1+0).

Now the verdicts that really went out by email. Over the last 90 days we sent a flight BUY verdict 4 times to readers outside our own accounts (1 distinct readers, as of Nov 4, 2026). 3 of them are 14 days old and could be compared with the same series (pooled across parties 2+2, 1+0) - that is under our floor (30 graded from at least 5 readers across 5 series), so we are not publishing the share.

The figures pool series for different travelling parties (adults+children, as of Nov 4, 2026): 48× 2+2, 35× 1+0.

As of Nov 4, 2026, data from Jun 2, 2026 - Nov 4, 2026, recomputed weekly over 83 flight series of parties 2+2, 1+0 with the same verdict rule the app uses. Flights only: a package verdict judges one specific verified offer, and the price history does not store that yet. A BUY forced by the end of the watched window is not graded, because it does not say "this is the floor" but "the window is closing and today beats what is left". A day whose series does not yet have 14 days of observations after it is not graded.

Why doesn't a new profile or destination get a verdict right away?

An honest verdict needs a comparison - and you can only compare against the route's own price history. A new route doesn't have one yet, so at first we show WATCH and collect data; real prices, however, you see from the very first scan. The verdict matures in stages: WAIT after ~a week of daily scans, BUY only after roughly two (a cheap price on a short series is mostly luck; the exception is the last few days before the travel window closes). We count recorded prices, not days (WAIT from roughly 8, BUY from roughly 14), so on routes we scan less often than daily it takes correspondingly longer. A mid-history price honestly keeps WATCH. A mature verdict then says where today's price sits in the route's history (percentile) and which way it's moving (trend). We'd rather wait than guess.

What is the real price, and why not the catalogue price?

The catalogue price is only half the story. "Real" adds airport transfer, luggage and board - the net cost for your profile and your home airport. That's what we rank by, so the "cheapest" listing isn't necessarily the cheapest holiday you'll actually pay for. More in Real price: what we count into it, and why.

What honesty rules do we keep?

We keep a few rules, even when they don't earn us anything:

• Prices marked "indicative" aren't verified - always check with the provider.
• Paid placements are always labelled and never enter the leaderboard.
• Links to vendors may carry our affiliate code, but it never sways the verdict - WAIT earns us nothing, and we still say it.
• Want the same data in your own AI assistant? We publish an MCP server with a setup guide - observations only, no verdicts.