Airbnb’s growth hacking story

Growthackers.com presents some insight into Airbnb’s growth hacking experiences that I didn’t know before. Interesting article that can be read here.

Airbnb, like many successful ‘out of nowhere’ startups began its life by solving a problem that the founders had themselves in 2007 — they wanted to rent out their own San Francisco apartment. They soon found out that they were not just solving their own problem, they were solving a problem that people all over the world had too — a solution that is now worth $10B.

Here are the things that stood out for me about what Airbnb did:

  1. Craigslist growth hack —very early on, anyone who listed on Airbnb could cross-list on Craigslist with one click, Airbnb helped by filling out all Craigslists forms with a ‘bot.’  The hack required some technical gnarl to perform, but it was perfect for this early stage. The team also appeared to look for all listings of vacation properties being listed on Craigslist, and emailed the owners to list also on Airbnb.Yes, it’s spam. But yes, it worked.
  2. They looked specifically at ways to improve early experiences  — we’ve all seen those rental listings where the basket of dirty laundry is still on the bed, and the washing up not done. Who wants to stay there? So instead of showing the owners how to take better photos, the team rented great cameras and took great photos for them. Now the team has a photography program where owners can schedule professional photographers to take photos for them.Yes, it was expensive. But the return on investment was apparent immediately in bookings.
  3. Disrupt the market with better everything — Airbnb had to be more than just better than a hotel. They were also substantially cheaper. And they have overcome trust issues by using social connections and by verifying locations. So when disrupting a huge, stale business, every part of the experience had to be addressed.
  4. Let’s hear it for the heart — Airbnb pushed out a “Wish List” feature that significantly affected engagement  — by allowing customers to “wish” to stay in places they’ve browsed. But it really took off when they changed the icon to add an item to a wish list from a star to a heart — increasing by 30%.Let’s hear it for A/B testing too then!

The article includes a video by a product marketing manager from Airbnb, and this gave more general advice that they learned as they grew, specifically in International growth.  Airbnb is challenged in that they have to both market to travelers, and to hosts. Hosts turn out to be the harder ones — people need to feel comfortable renting out their homes. So Airbnb has invested heavily in bringing in more inventory of rental places.

1. Talk to humans. In supply acquisition efforts, Airbnb used, and worked hard to optimize, Facebook ads. But it turns out they weren’t that great in terms of performance. And they were expensive. So they talked about how they tried something much more tactile.  They decided to ‘get out there’ and talk to real people.  In France, they invested in doing party and information sessions.  They got early adopters to invite friends.  They took out booths at local events.  They ran low cost, local ads. They handed out flyers. They put airbeds out in shopping malls.  And for everyone they connected with, they followed up personally. They asked questions. They figured out what was important to that market.  And they paid extremely close attention to what they spent on all this.

Turned out, their CPA was 5x better than those Facebook ads. Plus they learned a ton at the same time.

But that isn’t scalable, is it? Well, no. But once the market was kick-started, it grew at 2x the rate of the comparison control market. So they advise not to worry about scalability at first. Try and learn first, and maybe scalability is less important.

2. Launch with all you’ve got. In the beginning, we all see that a great article in a huge blog gives you a spike in traffic. But it’s not sustainable.  So the team really focused on throwing all they had at the problem. They pursued an integrated approach in Europe, that incorporated lots of in world guerrilla tactics, such as stunts, parties, flyers, and press. They had low budgets, so they worked on co-marketing with others — such as getting liquor companies to sponsor events, getting partners to co-tweets. They made it happen, whatever way they could.

3. Invest in people who can help you. Airbnb found that they didn’t have so many sources of traffic that they couldn’t spot anomalies.  When they saw in Google Analytics that there was traffic coming from one particular source — such as an influential blogger — they reached out directly. Influencers have been very good for them. They invested in giving those influencers what they want, encouraging them to do more.

4. Study your users… you know more than you think about them. Even if you only have an email address and Google Analytics you can ‘infer’ quite a bit about a user.  You can infer from device types, browser types, locations, and even an email address (like a .EDU). When you can’t afford a complex program such as KissMetrics, this is a helpful place to start.

5. Use email resources — these really help with email tests — Airbnb mentioned Campaign Monitor (which I like a great deal), Mailchimp, Sendgrid, and Litmus.

6. Don’t ditch your data — always be very data driven. Both positive and negative are very valuable. You can learn as much from what doesn’t work as from what does. But don’t overmeasure. Be very focused on what you want to learn.

The speaker spoke about finding the “North Star” key metric, and sticking to it.  Some of the metrics that they favor following are no surprise:

  • CPA (cost per acquisition) – lets you compare different levers and is the most important metric.
  • NPS (Net Promoter Score)
  • Whatever leading indicators is valuable; for them, they find that people who search with specific date ranges convert better than those that don’t. So they test with that as a leading indicator.
  • Plus, all the usual conversion funnel metrics

And finally, surveys. If the data can’t tell you something a survey from Olark can.  Use them to learn.

Thanks again, Growthackers!