Pay for better photos of the supply side's listings

This tactic means paying for professional photos of the listings your suppliers post. Buyers see better inventory, and the supplier does none of the work. The cited case is Airbnb. Third-party write-ups say its founders photographed New York listings themselves and that revenue in the city doubled within the month. The effort later grew into a program with more than 2,000 photographers.

How it works

On most marketplaces, buyers judge a listing by its photos before they read anything else. Suppliers may under-invest in photos, so the marketplace can pay for them itself. The marketplace can close that gap by paying for the photos itself. The quality of the whole catalog then rises, not just one listing.

1. Pick a market where the supply looks weak. Airbnb focused on New York, which the platform had entered in 2009 after starting in San Francisco.
2. Shoot the first batch yourself. According to a growth study of Airbnb, the founders rented a $5,000 camera and went door to door taking photos. An essay on doing things that don't scale describes the same door-to-door trips as recruiting new users and helping existing hosts improve their listings.
3. Check the result quickly. The same growth study reports that New York revenue had doubled by the end of the month. A blog post on how photography shaped Airbnb says monthly revenue doubled after the pilot.
4. Turn it into a program. The formal photography program launched in summer 2010 with 20 contracted photographers. By 2012 it had more than 2,000 freelance photographers covering 13,000 listings on six continents.
5. Track the effect per listing. Enhanced listings were reported to be two and a half times more likely to be booked. Researchers are also cited as finding that high-quality photos added an average of $2,455 in yearly revenue per unit.

What it takes

Our estimate: about 40 hours of work, and about 30 days before you can read a signal. Most of the hours go into choosing eligible listings, getting supplier consent, booking a few freelance photographers or shooting yourself, and uploading and tagging the new images. A small pilot of a few dozen listings fits in that budget. The signal takes about a month because bookings and transactions on most marketplaces arrive slowly per listing. You need enough of them in both groups to see a difference that is not noise.

What to measure

Booking rate per listing is the main outcome. Compare photographed listings with a randomized holdout, not with the rest of the catalog. A good reading is a clear, sustained gap in favor of the treated group. A bad reading is a gap that appears in week one and then fades.

Revenue per listing over 30 days tells you whether more bookings also mean more money. Better photos can attract more bookings at lower prices, or support higher prices.

Search click-through rate is the earliest signal. If photos are the lever, buyers should click photographed listings more often within days. If click-through does not move, a later booking lift probably has another cause.

Photo cost against added gross revenue per listing decides whether this can scale. Work out your payback period per listing. If the added revenue takes too long to cover the shoot, the tactic only makes sense for your highest-value supply.

Refund or complaint rate is the guardrail. A rise means the photos promise more than the listing delivers.

When it backfires

• Photos that oversell. Wide lenses and heavy editing can lift bookings and then raise complaints and refunds. Watch the guardrail metric, not just conversion.
• Shooting listings with no demand problem. If buyers are scarce in a market, better photos only spread the same demand across different listings.
• Cost per listing that never pays back. A program on the scale of thousands of photographers is a large fixed commitment. Scale only after unit economics hold in a pilot.
• Reading a lift with no control. The New York figure is a city total over one month with no stated baseline. Seasonality, new supply or press could produce the same pattern in your data.
• Suppliers who leave anyway. Photos do not fix poor hosts or sellers. Track whether photographed suppliers stay active, using a cohort view.

Run it as an experiment

Hypothesis: professional photos raise the booking rate of eligible listings in one market over 30 days, without raising complaints.

Comparison: we recommend a randomized split of eligible listings in one market, with half photographed and half held out. Use the A/B test sample size calculator to check that you have enough listings and bookings. If you cannot randomize, run a before and after with a fixed 30-day window on both sides and say so in the write-up.

What to log: listing ID, group, shoot date, photo cost, daily impressions, clicks, inquiries, bookings, price, revenue, refunds and complaints.

Decision rule: we suggest scaling if the treated group beats the holdout on booking rate and revenue per listing, the photo cost pays back within your target window, and complaints do not rise. Otherwise, stop or narrow to high-value listings. Our guide on how to run a growth experiment covers the write-up, and the growth experiment backlog template helps you rank this against other tests.

How strong the evidence is

The grade is practitioner report. The numbers come from third-party write-ups, and none of them was published by Airbnb itself. The Airbnb case carries the grade alone. The doubling claim appears in two secondhand accounts that differ: one says revenue in the city doubled, the other says monthly revenue doubled. Neither gives a baseline figure or a control group. The two and a half times booking figure does not say what the enhanced listings were compared with. The per-unit revenue finding from researchers reaches us only through a summary, not the study itself. The evidence shows a plausible early win and a program that grew. It does not show how much of the growth came from photos rather than the founders' wider work with hosts in New York.