How Startups like Airbnb Measure Customer Satisfaction

“The ability to learn faster from customers is the essential competitive advantage that startups must possess”, wrote Eric Ries in his book The Lean Startup, the unofficial bible of the startup movement. Measuring customer feedback is extremely important to startups, and in this blogpost we will show you how three of them (TransferWise, Uber and Airbnb) do it.

New to customer satisfaction? Read our other blog post about the subject.


TransferWise is an Estonian startup based in London founded by the people behind Skype. They offer cross-border money transfers and to date raised almost 400 million euros in funding.

TransferWise asks for the Net Promoter Score (NPS) after they process transactions. NPS is a popular way of measuring customer satisfaction among startups. First it is simple and provides a clear target metric to focus on. Secondly it measures growth because it rates word of mouth marketing and recommendations, very useful when you run a startup that wants to grow explosively.

Unsurprisingly the Vice President of Growth of TransferWise, Nilan Peiris, is a strong supporter of NPS, and the company uses it as their most important metric (he has a presentation on the NPS-led growth here).

The company, however, did not always use NPS. When they got started they relied mainly on interviews and open questions. At this time TransferWise’s product was still in development, and in-depth feedback was necessary to develop it into a product ready for wider adoption.

Only after this stage, according to Peiris, NPS becomes the key metric to measure and to push growth. “Only focus on NPS once the product is working” in his words.

By focusing on this single metric, and consistently improving it through actions such as referral programs or social media marketing, TransferWise expanded from a small startup to a company which handles millions of transactions every year.


Airbnb was founded in 2008, since then it expanded into a global company present in 191 countries. In 2017 Airbnb processed 500.000 stays every night globally, and feedback from customers is key to handling this volume of travelers.

Airbnb asks customers to leave behind a review after their trip is over. The review is optional and asks the users a broad number of questions.

It asks for:

  • textual feedback (open question)
  • the general experience, accuracy, cleanliness, check-in, communication, location and value of their stay on a scale of 1 to 5 (Customer Satisfaction Score or CSAT)
  • the NPS of Airbnb as a whole

For many people, going on vacation happens only once or twice per year, which allows Airbnb to ask this many questions without boring their customers. The resulting data serves two purposes: rating individual homeowners who rent out their properties (part of which future customers can see), and providing input for Airbnb itself.

Unique about Airbnb’s approach is that they invest heavily in their research team (listen to this podcast to hear from the director of that team), and part of their task is to play with the above data to see how it can inform business decisions. We get a glimpse of their work in a blogpost written by an Airbnb data scientist about how NPS scores relate to future bookings (spoiler: it does, but only in a limited way).


Uber is a global ride-hailing company that provides millions of rides every day by connecting drivers and riders. The company uses a simple 5-star rating system similar to CSAT. And while there might be good reasons to use this system, Uber’s interpretation of it caused some problems.

Uber uses this data to adapt their service, but also to rate their drivers. In some cases a low rating causes drivers to be deactivated and not being able to drive for the company anymore.

This quickly caused problems, for much of its history the company would deactivate drivers that went below a 4.3 average score during the first 25 trips. Uber interprets only 5-star trips as normal, but users often only give a 5-star rating if the service was exceptional: Uber interpreted the rating differently than their customers.

Frances Frei, a Harvard Business School professor, even called their interpretation of the 5-star rating “close to useless.” Unsurprisingly, Uber recently announced they would update their rating system: when the rider now gives a rating lower than 5 stars, they can select a reason why. This keeps the simplicity of the system, yet adds more information for Uber to work with.

This shows that customer satisfaction ratings always depend on context and how customers perceive them. Where in the case of TransferWise a single metric can be very useful and provide focus, in the case of Uber it did not accurately reflect reality, asking for more information can improve results in such cases.

Want to start measuring customer satisfaction like these startups?

Our Checkmarket platform offers all of the above methods (including NPS and CSAT), together with many more feature.


Net Promoter and NPS are registered service marks, and Net Promoter Score and Net Promoter System are service marks, of Bain & Company, Inc., Satmetrix Systems, Inc. and Fred Reichheld.

Laisser un commentaire

Votre adresse e-mail ne sera pas publiée.