How it works

The useful thing about this site is that you can check it. Here is exactly what we read from each operator, what we work out ourselves, and what we refuse to claim.

1307experiences read in full
5cities, refreshed every day
1,158,511traveller reviews behind them
2026-10-09prices and facts last re-checked

The three facts that decide everything

Every operator on Viator publishes these in structured fields. No booking site lets you filter on any of them, which is the entire reason this site exists.

What else we record

Around those three, we keep everything the operator publishes that helps somebody decide: duration, private or shared, what is included and excluded, cancellation terms and the free-cancellation deadline, confirmation type, languages, meeting arrangements, accessibility and stroller notes, weather-proofing, itinerary stops with their real place names and coordinates (788 activities are placeable on a map), and the operator’s rating with its full star histogram.

Categories, and what an operator’s tick-box is worth

Categories are the operator’s own tags on Viator, not ours. We filter out the ones that are internal bookkeeping rather than a kind of day — merchandising badges, compliance flags, quality scores, even Viator’s own name for retired taxonomy entries — and show what remains: walking tours, food and drink, historical tours, day trips and so on. An experience commonly carries several.

The accessibility and family notes work the same way: they are checkboxes the operator ticks, and some tick them carelessly. A whale-shark swimming tour in Cancún arrived flagged “infants and small children can ride in a pram or stroller”. So two rules apply. We attribute — the site says “operator says stroller OK”, never “stroller-friendly”, because it is their claim and not our verification. And we refuse to filter on the physically impossible: where an activity happens in the water, on a bike, up a cliff or on the back of a horse, the stroller claim is dropped rather than surfaced, because presenting it would be us endorsing something plainly untrue. That removed 66 claims across the catalogue.

How prices work here

Every price is the operator’s advertised from-price, taken from the supplier feed and shown with the date we last checked it. Prices are re-pulled daily. We never compute a quote: where we show what a party of five would pay, the multiplication is shown and labelled an estimate, and children with their own fare band may pay less. Your exact total is always confirmed on Viator before you book.

What we work out ourselves

These are our calculations from the published facts above. They are judgements, not facts, and the reasoning is always on screen.

How we sample, and why it matters

Percentages come from an unfiltered sweep of the experiences Viator features for a city — the inventory a visitor actually sees. We run two further sweeps, one sorted by rating and one filtered to private tours; their results are listed but never counted. Rating-sorted sweeps clip the perfect-five tail of small operators, and private sweeps are disproportionately priced per group — either would flatter our numbers. So the site keeps two populations: what we count from, and the wider inventory you can browse. Every page says which it is using.

Current per-group shares, all from the counted population: London 6.8% · Dublin 13.4% · Cancún 21.2% · Paris 5.4% · Rome 0.7%.

What “unknown” means

An unpublished limit is never read as permission. If an operator does not state a maximum group size, we do not assume your party fits — the activity is shown with the gap labelled rather than silently passed or silently dropped. Counts of “what fits your family” only ever include activities that actually publish the limits we checked.

What we do not claim

Where the data comes from

Experiences come from the Viator Partner API, refreshed daily by a scheduled job that re-pulls, re-runs our data checks, and only publishes if they pass — stale-but-labelled beats fresh-but-wrong. Location names and coordinates come from the same API’s location endpoints. One honest limitation: most meeting points arrive as bare map references we cannot resolve to a name, so we show a meeting point only where the operator’s data actually gives us one.

Family stays

Our accommodation research works differently and by hand: we read each property’s own published room pages and record occupancy in their wording, bed types and counts, bedrooms, bathrooms, room size, kitchen and laundry — 25 properties, 41 room types, 17 of which publish a full bed breakdown. Every record carries a source URL and a verification date, and anything unpublished stays unknown. We never infer a bed layout from an occupancy number: a room that sleeps five could be a king plus three twins, or a queen plus a sofa bed and a cot, and those are very different rooms for a family with teenagers.

Corrections

If something here is wrong, we want to know and we will fix it and say so. Tell us.