Seatfrog Auctions Empty First-Class Train Seats to Economy Passengers

01 /The Concept

Seatfrog is built on a simple observation: 570 million first-class train seats go empty every year across the UK and Europe, and the rail industry is quietly hemorrhaging revenue because of it.

The app lets passengers buy tickets from 16 rail operators covering 3,400 destinations across the UK and Europe. Ticket exchange is built in too – travelers can swap a booked ticket for a different departure time for as little as £2.50, right up to 10 minutes before the train leaves, skipping the cancel-wait-rebuy cycle.

The real play, though, is the first-class upgrade auction. Passengers who've already bought a standard ticket can set a maximum amount they'd be willing to pay for a first-class seat. Shortly before departure, Seatfrog checks how many premium seats remain unsold, then runs an automated auction among interested passengers – bidding up on their behalf to their stated ceiling. Winners pay only what it takes to beat the next highest bidder, not their maximum.

The results speak for themselves. Since founding in 2018, Seatfrog claims to have served users across 3 billion journeys and saved them £43 million on tickets. The auction mechanic has proven so sticky that 70% of new users arrive through word-of-mouth referrals – and revenue grew 9x in 2022 versus the prior year. The company closed a £6M funding round bringing total investment to $23.2M.

02 /Why It Matters

The idle-inventory problem is far wider than train travel. US hotel occupancy rates peaked around 70% even in strong years – meaning roughly 30% of rooms generated zero revenue. Airline seat utilization hovered around 80% pre-pandemic and has slid back toward 70% since. Every unsold seat or empty room represents money that will never be recovered.

A [related review](/review/pribyl-iz-niotkuda) covered Caravelo, which helps airlines squeeze more revenue from exactly this dynamic – through bundle optimization, frequent-flyer subscriptions, and auctions on premium legroom seats near emergency exits. Both companies are attacking the same structural inefficiency from different angles.

The conventional solution – permanent price cuts – creates its own problems. Demand is notoriously hard to predict: today's empty room is tomorrow's sold-out night. Indiscriminate discounting destroys margin on capacity that would have sold anyway.

The auction model resolves this cleanly. Price discovery happens in real time, the floor stays protected, and buyers self-select based on willingness to pay. The same logic could work anywhere physical capacity is fixed: theater seats, restaurant tables, even online masterclasses where an artificial scarcity of 50 spots can yield far more revenue than unlimited $11 tickets or unlimited $1,100 tickets – because an auction surfaces what the audience actually values.

03 /Opportunities

The broad direction here is dynamic pricing platforms – tools that help businesses optimize revenue by moving away from rigid price lists toward flexible, demand-responsive models.

In the transport layer, Seatfrog and Caravelo are the clearest examples. For B2B sales – where negotiations replace auctions – Subskribe ([covered here](/review/mezhdu-zhadnostju-i-konkurenciej)) handles enterprise pricing conversations. Orb models flexible tariffs for cloud services. Pricemoov takes a more comprehensive platform approach.

Nibble ([covered previously](/review/nazovite-svoju-cenu)) brings an AI negotiation chatbot to e-commerce, letting shoppers haggle on price in real time. InDrive built its entire ride-hailing identity around the same haggling mechanic.

Engage3 ([covered previously](/review/byt-ili-kazatsya)) takes a subtler approach in grocery retail: shoppers calibrate their sense of a store's price level by checking a handful of anchor items, then essentially stop comparing on everything else. That lets supermarkets attract traffic with competitive anchor prices while quietly improving margins on the rest of the basket.

The toolset for implementing dynamic pricing is mature: English and Dutch auction models, surge pricing algorithms, AI-driven demand forecasting, and LLM-powered negotiation bots capable of handling unlimited concurrent conversations. What's lagging is adoption across the industries that could benefit most. The opportunity is to identify a specific vertical still running on static pricing – one where there's measurable idle capacity and predictable demand variance – and build the simplest possible pricing layer for it.