couch-table-airbnb-helpshift-1

Jessica Stocks joined Airbnb’s burgeoning Customer Experience team 5 years ago.

For Jessica, who is the Workforce Manager for North America, this growth has presented unique challenges, exciting opportunities, valuable lessons, and hilarious customer interactions. As she told us, “We hear from all kinds.” We visited her at the Airbnb San Francisco office, for a carnival-like tour, a delicious lunch with unlimited cold brew, and a chat about Customer Experience at Airbnb over the past five years.

HS: You’ve been at Airbnb for 5 years–can you describe the changes that the company and your position have undergone in that period?

Jessica Stocks: So I started in CX before they had any sort of specialization, when Airbnb was not quite a hundred people and I was able to work remotely at that time, and you would just handle anything under the sun, there was no division for payments, for trust and safety, for those calls like “what is Airbnb?” You would just take all of it. But quickly after that we started specializing.

HS: At the Customer Support Meetup, someone from Airbnb said that when growth is green, support is red. How do you handle the massive amount of growth you’ve undergone in the past five years?

JS: The amount of growth is hard to account for, especially in a traditionally seasonal industry. The arrival pattern for our volume doesn’t actually change much throughout the year because of the growth that we’ve experienced. But as we become more stable, that seasonality is going to look much more dramatic. And then you’ve got to consider how do you ramp that up, how do you ramp that down, what amount of time do you need for training to make support specialists ready for that summer season. What do you do in the off-season with those teams?

We’ve gone through a couple of different philosophies to handle growth. There was a period in CX’s trajectory where we tried to remain mean and lean, and not mean like angry or anything but just a really tight headcount, really trying to improve our processes and workflows to just make sure that we could get a ton of work done without flooding our department with people. That drive at efficiency was one period. Now, I think we are headed in a direction of not necessarily mean and lean, but really looking at making our operations more efficient and effective.

We’ve mostly worked with call centers so that we can ramp up and down during on and off seasons. The great thing about that is, our seasonality actually blends very well with other companies, like retail, which starts ramping as we’re ramping down. So we have relied on support there, but we don’t want to ramp up and down internally. The founders said many years ago that no matter what our percentage headcount is at our contact partner sites vs internally, we want to have at least some amount of staff internally guaranteed, in case things need to be taken in-house or in case there’s a breakdown somewhere. We need a certain level of redundancy. And especially because of how quickly we iterate on our own product, sometimes the easiest way to get that feedback and understand what changes are good and bad is through the customers support team interacting directly with product teams.

HS: Right, how do your product and CX teams work together? What are your channels of communication?

JS: So I think it’s still a work in progress. We’ve had so much growth and change that it’s been hard to keep up and implement processes. I think we’re coming into that period where we’re actually setting up processes and know how to go about the system in an organized approach, instead of me just knowing a particular person on the product team and being like “Hey Shirley, what’s going on?” or over lunch, chatting with someone and realizing “Oh, I wish we’d known that experiment was happening.” I think that we have gotten by with that for a long time just because the company was so tight and so close.

Also, although we have gone through some pretty significant product launches, especially on the mobile side, they were big enough that they weren’t happening all the time, the product wasn’t overhauling weekly. And things like that, or bugs, we have established ways to send that feedback really quickly. We have Slack channels galore, and we had a whole team that was paying attention to particular bugs and feedback, and they’re dedicated to different product groups so it’s all going through particular pipelines.

I think where we’ll start moving, and this is a WFM (Workforce Management) function in my view, is that we’ll be able to start planning ahead. So when a product team says “Hey, we would like, in a week or two, to redesign the layout of the mobile app”, we will be able to forecast how much more contact we will get, and whether that contact will drop off, so long as the product works. And then from there we should be able to say “ok this was within tolerance”, or “oh, we’re going to have to hire more people because this was a permanent change to the product and it’s going to produce new attitudes and new treatments.”

Something that came up during the meetup panel discussion was our Resolution Center. We said there needs to be a place for hosts and guests to exchange money with each other later – like “oh I forgot a sweater in the listing, Host can you mail it to me? I’ll throw you five bucks.” Things like that that come up and were requiring a lot CX effort and support. The Resolution Center, when that launched, was a massive product overhaul, it dramatically changed CX’s experience and the amount of contact we were getting. But in addition, it also changed user behaviors: they came to know this product, they started using it before contacting CX. So we have to anticipate that the amount of calls we’re going to get to troubleshoot the Resolution Center are a fraction of the calls you would get for Resolution Center-type cases if it didn’t exist. I think we’re getting to the place where we can work with product to anticipate what they’ll be doing and know what those changes are going to mean for customer experience.

HS: Most of your forecasting is focused around seasonality and product changes, then?

Really it’s not, it’s based on usage of the site and app. The areas where we need to improve and head next is anticipating product changes and seasonality. Especially seasonality by language. So as a good example, North America’s volume and actually most of the world’s volume that we receive trends up June, July, August. However, high season for Japan is February. Chinese New Year is another great example. It’s the exact opposite. What we would be doing next is looking at how that seasonality or how product changes can produce different impacts on different channels and regions. A mobile product change, for instance, would have a significantly different impact on our chat volume than on our phone volume because mobile users are more likely to head to chat and to want to use chat.

HS: What is your demographic like?

JS: Fun fact: The fastest-growing group of Airbnb hosts in the U.S. is seniors (hosts 60 or older).

HS: Have you implemented initiatives to improve knowledge so that people don’t call?

JS: That’s been a giant effort of ours for the past year and a half. Improving access to information upfront to make contacts useful and purposeful and as precise as possible, so that when we get that phone call we know precisely what it’s about and we know it’s something that cannot be supported autonomously on the site.

HS: What metrics do you use to measure your success?

JS: We’re really playing with that right now. As the WFM mind in the building, I have very specific ideas on what it should be and shouldn’t be and why. For instance, we do pay attention very close attention to closed tickets per day.

We are very heavily focused on NPS because we want to ensure that we’re accessible, that we’re providing the support and care and concern for those sensitive moments of truth. So NPS is really a guiding light for us, more so than a lot of other organizations. But I think we are still looking for ways to improve our in-depth quality analysis and expectations and standards which will probably eventually lead to expanding beyond NPS.

Published September 29, 2016
block background image