How AirBnB Utilizes Twitter for Customer Support

While traveling in Central America, I became a heavy AirBnB user. Back home I used AirBnB occasionally, but while traveling especially in costly places, it proves the cheaper and often better alternative to a hotel or hostel. During my time abroad, I had one negative experience with an accommodation that looked totally different than listed on AirBnB. After reaching out to AirBnB via Twitter, I was pleasantly surprised by a great reply within two minutes. Two emails and 15 minutes later, I already received a proof of refund for the price paid. This drew my attention towards how AirbnB utilizes Twitter for customer support.

First off, I am actually not that picky with places while traveling. It does not need to be fancy for me, so I found the “rustic” garden house appealing. When I saw I was supposed to move into a shed the size of a bed plus three feet give or take, topped off with damp pillows, I concluded it was a bit too “rusty” for me. Now, that you have an idea as to why I reached out to airbnbhelp let’s get into some data.

Twitter for customer support means tweets, a lot of tweets

As I usually do while analyzing social media with quintly, I started with a Key Metrics Table to get a bird's-eye view of the numbers. This table also helps to put numbers into perspective with different profiles being analyzed. Period of time analyzed: Jan 1 - Feb 20.

Not surprisingly, AirBnB itself has significantly more followers than their support channel. We can already interpret this as a good sign as there seem to be less problems than interested people. Comparing this relation with the German Telekom and their support channel (63k vs. 55k), you can see that the difference here is significantly less noticeable, which might indicate more problem-facing customers. Looking at the numbers of Tweets airbnbhelp has been sent this year already: you can tell that a handful of people are working to keep their customers happy.

Different channel, different purpose

The breakdown of interactions makes the different purposes of the channel noticeable. Whereas Airbnb’s main channel receives 70% likes, 28.1% retweets and just 1.8% replies, replies make 77% of the support channel. The dashboard used for this analysis shows as well how the Airbnb main channel dwarfs the support channel in terms of likes and retweets. All this is not negative at all, just a sign of different, well-set priorities and goals.

Supreme discipline “Response Times”

The aforementioned differently-set purposes of channels also becomes apparent when looking at the response times. On the left, you can see that Airbnb does not reply to users and through this, educates their community to reach out to their dedicated help channel for assistance.

Looking at our superb customer satisfaction team at quintly, they continuously strive to keep our response times fast. This is a really important KPI in order to retain happy customers. By the way: Our team can be really proud of the fact that we achieved a two minutes’ (!) response time on average.

With that said, looking at the stacked column chart you can see one of most important numbers for customer support teams. Here we can see that airbnbhelp manages to reply to almost 40% of all inquiries in under 2 hours. Another 40% are replied within 2-8, which make 80% of all inquiries are replied in under 8 hours. For the individual customer, waiting for 8 hours might be frustrating, but looking at the size of the company and the number of questions and complaints received, this is a great achievement. Additionally, it could be interpreted negatively that almost 10% of all tweets are not answered, but this could have different reasons, too. Problems might have been sorted out in a different way, or some of these tweets were just spam. In order to answer this question more precisely, we would need to look closer at the questions table, containing the individual tweets.

Benchmarking AirBnB’s Twitter performance

To put this into perspective, we compared the numbers analyzed to a similar service which could substitute AirBnB – Booking.com. At quintly, we deeply believe in the importance of competitive benchmarking, so we first understand the own numbers, and then add meaningfulness by comparing to the competition. In the analyzed case, Booking.com does not have exactly the same business model nor do they have a channel that is purely dedicated to help. Nevertheless, it’s interesting to see how many questions they receive and how these inquires are handled.

booking.com being here the “bigger channel” with more followers (booking.com has 112k and airbnb 35k followers), received significantly fewer questions than @airbnbhelp (754 vs. 2,352). These purely quantitative insights have set the base for the comparison and we can now dig deeper into response times.

Earlier we compared AirBnB and their help channel and identified a clear strategy in both: One channel is for helping the user and the other one is for other external communication.

The chart shows that booking.com leaves 75% of their received 754 questions from Jan 1 to Feb 20 unanswered. AirBnB answers almost 93% of the received 2,352. Kudos, AirBnB!

The channel seems to fulfill its purpose and clients receive, on average, a quick answer. In contrast, booking.com seems to lack managing to reply to questions asked. Important to note here is that booking.com is not a help channel but one channel to fit all purposes. Airbnbhelp’s only purpose is to keep customers happy with fast solutions, like my refund of the “garden shed”.

Twitter for customer support – A wrap-up

The analysis showed that AirBnB manages to do that in over 90% of the cases, while booking.com has a lot of potential for optimizing their communication with their customers and potential users.

Having said that, this use case does not only exemplify the great possibilities companies have on Twitter for customer support while engaging with their customers, but also should point at something more strategic. Having a channel that is clearly dedicated to one goal makes measurement very straightforward - the KPIs are clear. For Airbnb’s help channel it’s Response Time, and for AirBnB it’s Interactions per Tweet. For booking.com this is pretty unclear and thus hard to measure.

Perhaps the analyzed 75% are not as bad as it sounds as all important tweets have been answered. Perhaps their amount of interactions is not too bad as well. But as the word “perhaps” indicates, you can never be sure. AirBnB knows exactly: “In the last 2 months we answered 40% of all tweets under 2 hours, which is great but the goal until end of Q2 is to increase this number to 55%”. Goal set, KPI clear.

So the clear takeaways are: Twitter can be a great network for fast and clear communication with your customers. Secondly, a channel with one clear focus helps to measure your Twitter performance correctly and accordingly adjust strategies.