June 21, 2026 · 8 min read · Industry

Sustainable travel: which signals you can actually act on.

Carbon-aware booking is a real market trend. It is also full of theatre — eco-rated badges from sources nobody can audit, vague "green hotel" filters that map to a marketing form. Four signals actually change behaviour. Use those.

The bad signals first

"Eco-rated 4 leaves." Where does the rating come from? Many travel aggregators source it from the property itself or a non-authoritative third-party scheme that requires no audit. The user reads it as certified. It isn't.

"Carbon-neutral flight." Most carbon-neutral flight claims rely on offsets purchased from voluntary carbon markets where the underlying project quality varies enormously. The booking surface flatters the user; the atmosphere does not notice.

"Sustainable destination." Often a marketing label applied by a tourism board with the explicit goal of attracting more visitors. Recursive irony.

The common pattern: signals that come from parties with an interest in the booking happening, not from independent measurement. Filter those out of your product, or you train the user to ignore the real ones.

The four signals worth ranking on

1. Rail-vs-flight routing distance

For city pairs under roughly 1,000km, a train is meaningfully lower in carbon than a flight per passenger-kilometre, and often competitive in door-to-door time. Paris to Amsterdam, Madrid to Barcelona, Tokyo to Osaka, Berlin to Munich, London to Brussels — the math favours rail. Beyond about 1,500km it doesn't.

The signal: when both modes serve the route, surface the rail option first. When the carbon delta is large, say so numerically (kg CO₂ per passenger). Most users will take the train if you put it in front of them; the existing UI defaults push them to the flight.

2. Overtourism risk

Measured visitor density relative to carrying capacity. Venice, Dubrovnik, Santorini, the centre of Barcelona, Kyoto's bamboo grove, Machu Picchu, the Acropolis in midsummer — all visibly past the point where additional visitors degrade the experience for everyone, including the locals.

The /v1/cities response includes overtourism_risk as a bounded score for paid tiers. It's a function of public visitor- arrival statistics, accommodation density, and seasonality, not a subjective rating. Use it to re-rank: when a user searches "Greek islands in August," surface Naxos and Milos before Santorini, with a brief explanation. Most users will accept the swap.

3. Low-season vs peak-season timing

Same destination, different month, dramatically lower environmental load. Reykjavik in October instead of July. Cape Town in March instead of December. Tokyo in February instead of cherry-blossom week. Bangkok in November instead of December.

The signal is two-part: surface the seasonal price differential (which the user already cares about) alongside the overtourism risk differential (which most users will accept once shown). The underlying data is public — visitor-arrival monthlies are published by every major tourism board.

4. Hotel-segment density

The number of hotel beds per square kilometre in the city centre is a proxy for how much over-development has already happened. Cities with extremely high density — central Cancún, the Strip in Las Vegas, central Phuket Patong — are signals that more development will not improve the place. Cities with moderate density are still growing in a healthy way; cities with low density (Reykjavik, much of Buenos Aires, Lisbon outside the central neighbourhoods) often have margin.

Use this to rank "where to stay" within a destination. The underlying counts come from OSM and Foursquare's open Places dataset.

Putting it together

GET /v1/cities?country=IT&tier=growth
[
  {
    "name": "Venice",
    "overtourism_risk": 0.94,
    "season_peak": ["May", "Jun", "Jul", "Aug", "Sep"],
    "rail_connectivity": "high"
  },
  {
    "name": "Bologna",
    "overtourism_risk": 0.31,
    "season_peak": ["Jun", "Jul", "Sep"],
    "rail_connectivity": "high"
  }
]

A re-ranking strategy can take a "northern Italy in August" query and demote Venice in favour of Bologna or Verona, citing the overtourism delta. The user gets a comparable experience, the destination gets relief. The signals are all derived from auditable public sources.

Honest scope

We don't ship per-property carbon footprints. The data quality from properties is not good enough to justify ranking on. We don't ship flight-level emissions estimates either; ICAO's calculator and the Travelport / Skyscanner integrations are closer to authoritative. We ship the city-level signals where the data is solid enough to bet a product on.