Prototyping At Booking.Com: Lessons & Failures

Melody David, John Iseghohi at UXDX Community: Amsterdam. Video: https://youtu.be/jKIILiq-f18

Readable transcript: edited from the recording's captions for readability (fillers and false starts removed, punctuation and section headings added). Wording is the speaker's own. Timestamps are positions in the video. Names marked [?] could not be verified against the audio.

How we work at Booking.com

[00:00:10] Melody: Hi everyone. Today we're going to talk you through our prototyping process for one of our accommodation partners, but I'd also explain the context of Booking as well and how we work, similar to Rory and I think Ruben earlier. I'll just explain a little bit about our process at Booking. Before I launch into Minimum Viable Pain, which is what I like to call MVPs, I also want to introduce myself and explain a little bit about my role.

[00:00:41] Melody: I'm Melody, I'm a senior UX copywriter in our tech department. I've been working in tech for about two and a half years, and when I first joined Booking I actually joined as a junior copywriter. The first product team that I worked on was a team that was trying to solve problems for our customers in terms of pricing issues. One of the biggest pain points that we discovered was actually the final price, people not being clear about the final price they're going to pay for their accommodation.

[00:01:09] Melody: Now, you'd think that would be really easy for us to solve, but it wasn't, due to the way that we traditionally built our platform and also the way that we worked with global tax legislation and all the complexity of different laws in different countries. It wasn't always clear to guests if there were extra charges included, if they were going to have to pay some sort of fees. Of course, as you can imagine, unexpected charges for customers cause customer service calls and also a lot of friction between guests and partners. So we spent a lot of time in that team prototyping solutions to fix those issues.

[00:01:44] Melody: Trying to fix those issues for a company at Booking.com can be quite challenging, and here's why. We're one of the leading travel platforms in the world, so we have customers from all over the world. This means that our customers travel to 147,000 destinations and they book over 1.5 million room nights with us every day. That's just to give you that idea of the traffic that we have and the different types of people that we're dealing with on a daily basis. So it's quite a lot.

[00:02:12] Melody: That scale is also reflected internally, when it comes to our employees and the people working for our company. We have about seventeen and a half thousand across two hundred offices worldwide, so go figure. In terms of our product teams in our tech department that's based here in Amsterdam, that's our HQ if you didn't know, there are about two hundred product teams across tech. They're working on everything from solving complex user problems to also trying to innovate for the future and think ahead.

[00:02:43] Melody: Right now the way we're set up as product teams is pretty simple, similar to I think what Rory said in terms of moving away from this linear way of doing functions. We actually work in cross-functional teams, so we will have a product manager, we'll have a UX designer, we'll also have a copywriter, and of course many, many developers. We like to work in a very agile way. We have two-week sprints, we do the whole prioritized backlog and kicking off with a sprint meeting.

[00:03:06] Melody: But we also really, really like to test everything, so we're really data-driven. Some of you probably already know that. We don't really like to build things unless we have very strong indicators to do so. One of the ways in which we do that is through A/B testing, similar to what Ruben said. We have a very strong experiment culture at Booking.com. We like to launch things on our users unsuspecting and see how they react to things.

[00:03:31] Melody: But we also like to do our research. Research is really, really important, and as a UX org that's pretty much a lot of our job. We have various different methods of doing that. That could be anything from observing a user interacting with our product, to actually sending them a survey or questionnaire, or interrupting their flow when they're on our site. We actually like to work with user researchers, so we have dedicated user researchers throughout the department who help us choose the appropriate methodology depending on what the user problem is that we're trying to solve, or what the business goal is.

[00:04:04] Melody: The entire product team is involved in user research, so it's not something that is expected just for the researcher to do or the product manager to do. Everybody's supposed to get hands on and really understand their topic and understand who they're building for and who they're solving for. We also like to look at previous data. As I said, we're very data-driven, so we want to see what's happened before, what's worked before, what didn't work before, what other people are doing. Of course we look at industry trends as well.

[00:04:31] Melody: And lastly but not least, we like to speak to people in our local offices in the markets where we operate. We need to understand local realities. We also need to speak to our customer service people all the time, because they get the brunt of a lot of the complex problems that get passed through them. That means that we end up spending a lot of time going on research trips, also working in our usability lab in Amsterdam.

Why copy is part of the prototype

[00:04:50] Melody: In terms of the importance of research, yes, we've highlighted that, but I think what is really important is all the different roles that everybody plays, and I'm going to speak specifically about my role now. What we do is we speak with the designers and we try to work together on the flow as a team. First we try to identify the user problem, define the user problem, and once we've defined that and we start to understand exactly what the problem is, then we say okay, let's now start working on solutions.

[00:05:20] Melody: That normally is led by the designer, but the whole team can ideate and give ideas. So it's not just for the UX designer and UX copywriter to go away in a room and say okay, this is the solution. We will take feedback, and there will be this rough mock-up stage brought back to the team, everyone will chip in, and then we work together on the flow to say okay, actually this action and this copy makes sense together.

[00:05:43] Melody: The reason why we place such an emphasis on copy is just because we know that bad copy, or placeholder copy like lorem ipsum which isn't real, can confuse users during user testing and it throws them off. User testing isn't cheap, so we want to make sure that we use our time efficiently and we also want to make sure that we get the most out of that session. So what we tend to do is work together on the flow and write real copy. That way we can actually get feedback on the words, if we want to iterate on the design and flow based on what these users tell us during that session.

[00:06:16] Melody: Another reason why we use real copy as well is because we want to make sure that the process is set up in a way that's unbiased. For example, we want to make sure that we work on the interview script with the researcher, and that the researcher understands not only how do we ask unbiased questions, but also how do we get the customer to fully understand the prototype and flow. If that was done separately by the researcher and separately by the copywriter, there might be some friction in that interaction there.

[00:06:41] Melody: And also localization. We have our site in 43 different languages, which means that we have to write in a way that is easy to localize. If we weren't doing that as copywriters and someone else was writing that without that in mind, they might write something that's not actually translatable or not actually scalable to other countries and cultures and contexts. So I hope that's given you a little bit of context about how we work, and John is going to introduce himself and then talk a little bit about this case study.

Introducing WebDirect

[00:07:13] John: Okay, thank you very much, Melody. Hi everyone, my name is John Iseghohi. Yeah, most people actually find it difficult to pronounce that name. I come from Lagos, Nigeria. I moved to Amsterdam twelve months ago.

[00:07:33] John: Yeah, thanks. I prepared for this talk with the speaker notes, but now I have five minutes, which is fine. So I'm just going to try to engage with you and then make it as practical as possible, because we're actually going to look through one of the case studies, one of the product reviews, and then talk through it, and then eventually I probably want you to engage with me.

[00:07:59] John: So I moved to Amsterdam 12 months ago, and prior to then I had worked with some three different companies and in different industries as a UX designer, specifically advocating for users, which means I've been designing for a very long time. One of the things that I actually learned is how to align both the user goal and the business goal to develop, or to design, a good product. Joining Booking, I discovered that hey, you need to really be big on data and understand what data means to your design. Whatever ideas you're coming up with, you have to really be data-driven, and you have to get insights from whatever you are doing and make sure you deliver a good design.

[00:08:44] John: So I will be talking about WebDirect, but before we actually go deep into this case study to understand how we designed the seamless onboarding experience for new users, I would love to talk to you about what WebDirect is. Everybody knows that Booking.com is that platform where you can book an accommodation, whether you're going for a city break[?] or just any kind of accommodation, any kind of holiday you're going for.

[00:09:12] John: But most people don't know that Booking actually has a team that focuses on building products for partners, meaning accommodation partners, B2B. WebDirect is actually one of the products that we've built. As my colleague mentioned, we try to innovate internally, we absolutely really do innovate internally, to be able to make sure that we're building products that our partners will actually want to use. WebDirect is one of the products that we've built, and this actually came up as part of the research we did.

[00:09:43] John: Eventually, building a website or having a website is as easy as anything, with Wix, Squarespace, so many platforms for you to have a website. Which means eventually some of our partners might want to one day decide that, you know what, I want to have my own website and I'm leaving Booking.com, I don't want anything to do with Booking.com. So we decided to come up with this product that can actually help them build their websites, their property websites, in less than one minute.

[00:10:12] John: So what is WebDirect? WebDirect is actually a software that helps our partners to connect their already existing information and property information from Booking.com and sync this into a website in less than 60 seconds. Very easy.

Launch, and why it didn't work

[00:10:29] John: Sounds really good, and anybody would want to just jump on it and say, you know what, let me buy this product. Yeah, we built this product, did all the research, everything works really nice from our team, we're excited about it and thinking that everything would just go as planned, but trust me, it didn't work that way. Partners already are used to, I just want to list my property on Booking.com and start getting bookings straight up.

[00:10:53] John: So getting them to actually set up a website, which means they may have to start doing some form of website management, make sure that the content is up to date, even if we've synced all the content from Booking.com into these websites, content including guest information, availability, cancellation and all these things, we've already done all that for them, but we still get the problem whereby retention to the product was really not where we actually wanted it to be. Growth rate was not there, and even the experience with the product, because we tried to build the product as lean as possible, minimum viable product as it is, for them to quickly set up their website, but some of this eventually did not actually deliver on the things we wanted.

[00:11:38] John: So we started exploring, we actually went deep into research. We've built this product, so what is the problem? Why are our partners complaining? Why can't we get them to actually stay in the product and eventually use it to get their first direct booking?

[00:11:56] John: So we dug deep and discovered that partners actually wanted some form of hand-holding, because they're already used to that. Once you set up your website on Airbnb or any other OTA, somebody either calls you or sends you an email and says, hey, thank you for signing up on Booking.com, this is what you need to do, and eventually you do those things and then you're ready, up and going. But for this new product, because it's a small team, the onboarding process was not there, because they really wanted to move and move really fast.

[00:12:30] John: Eventually we discovered that partners actually needed some hand-holding, so we built an onboarding process into the product which would help them actually discover the magic moments of the product. The first step, after a new partner actually signs up, meaning the old partners have already gone through the terrible stages, we needed to come up with something fresh, a new idea to help them into the product. For new partners we're now building a three-step process into the onboarding flow. Follow with me.

[00:13:01] John: Once you click on sign up for this product, you see the USP: your website will be translated into three languages upfront, you can start receiving payments almost immediately with 50 currencies on your website immediately, you don't have to get a developer to do all the setup for you, everything happens within one minute. Good.

[00:13:20] John: But now here comes the problem. We try to explain to them what we've done. Step one, hey, we've synced all your information from Booking.com into this website for you. Step two, this is a preview of your website, meaning you see your hotel name, you see your property information, you see everything. But there lies the problem: they don't like the preview of the website you give to them. So you now see where the problem keeps coming. I mean, it's an MVP. But no, they started complaining, and eventually we start getting calls: I paid for this product, I paid for this website builder, and I mean, I don't like it, I want to change, I want to change the fonts, my brand colour is red, you're giving me blue.

[00:14:00] John: So 50% of the users that came into the product actually dropped off[?]. Meaning, the MVP, what we actually wanted to gain here is to increase retention rates, get partners to actually start telling us what they really want. So now we know, from just building something as simple as this, that partners want some form of customization. They want some form of, I want my brand to look this way, not this way.

The second MVP: customization

[00:14:31] John: So now we know, so we iterated on this. This was our MVP, and eventually did another one, which is just for us to do two things. One of the learnings from the first MVP was we noticed most of the partners that actually signed up for this product already have their websites, but they still bought a website builder from Booking.com because it just feels like it's Booking, yes. So we needed to understand how many of our new customers are coming that already have their websites, and also segment them, so we can actually do something interesting with that data.

[00:15:05] John: And also introduce website customization in step two. Initially, we didn't put that for the first MVP, it was just plain. Now the second MVP, we introduced customization. So customization means that you can select a theme, you can select fonts, you can select colour, you can play around, all these really nice features. Take note, it's an MVP. Take note again, it's a small team, not the entire big tech product team. It's a really small team just building this, so you can see how we're trying to just learn and just make things as lean as possible and exposing it to millions of users.

[00:15:41] John: So customization was introduced. You keep playing around, and once you play around, they play around, really eventually we noticed things actually moved the way we wanted. We noticed people eventually, and eventually we can now start recommending to partners that already have their websites what they can do, which is, if you have a website already, meaning we now get you to change your DNS to point to the website we created for you and just use this, which was good.

[00:16:20] John: So for the second MVP, this actually went live anyway, and as you know we learnt a lot from this. Good numbers. But yeah, the goal of having a website as an accommodation partner is really for you to get revenue, not just that, but that means you need to do some form of work. You need to write articles, you need to set up SEO, you need to do so many things which we can't do for you. So we still need them to do some work, but we need to educate them about this.

[00:16:51] John: So then comes the first MVP: we were able to sell customization, we were able to show them what. Second MVP, we were able to actually let them customize. Should we go for the third MVP to let them start doing some form of SEO optimization on photos and all that? We need to check the effort, if we actually have a bigger team that would be able to invest time, man-hours, into that. We need to also check, would it be profitable for us, or should we just keep it at this point? Anyway, we actually solved the same problem, but that's a talk for another session.

Walking through the flow

[00:17:26] John: So we moved, boom. Now I'm just going to show you like a live recording of the second MVP so you can actually have a look. So I did a small recording. Now, for step one, as you can see it's very simple: what's your property, do you have a website. So this is actually a live user with this, and I've got a recording[?]. So I'm going to play this and then you see.

[00:18:00] John: Cool. So step two, customize your website look and feel. Users actually spend most of their time here. Before now, before we introduced this three-step process, this never existed. All they needed to do was they paid, they bought the product, and then they get lost.

[00:18:20] John: Just to give you more context about a day in the life of some of our partners, they receive over eighty emails on a daily basis, emails about their accommodation, cancellations, so many emails, not just from Booking.com but from other OTAs. So you can imagine the emails about this particular product will be sitting somewhere in between unread emails on a daily basis, which means they never get to see these products at the end of the day, because they have information about revenue that they need to take care of. So we decided to remove sending them emails or anything. Once you sign up for the product we just push you straight into this onboarding process, and then the magic moments, which is this.

[00:19:02] John: You see, they just start looking for fonts that actually align with their brand guidelines. But what did we do? We didn't actually have to put in all the fonts in the world here. We wanted to know what fonts are they even interested in, so just have like 50 fonts. And then eventually, if their fonts are not here, maybe we start getting calls: hey, I noticed you don't have my brand font. Then how many calls did we get about that? That would determine if we need to do some form of integration here, which I'm sure everyone knows, maybe we can integrate Google Fonts, many giants, with this, but that would require lots of effort.

[00:19:39] John: Cool. So eventually you go to that stage, that's the third step, and then you land into the product. Amazing, right? Really simple product, easy to use. But trust me, oops. Good. So that's the live recording of the MVP that we built.

Learnings

[00:20:04] John: Cool, quickly then we move to learnings. Like every product, everyone has learnings. There are so many things I must have learned while building this product. Now we actually learned that most of the partners already had their websites, which is true, because we did research. We know that our partners actually want to have their websites, but they have already listed on Booking.com, and we want to be the one that will provide a website for them, not Wix, not Squarespace. Just keep them in the ecosystem, which is nice.

[00:20:34] John: They expect everything to work once they click the button saying yes, because once you put a listing on Booking.com everything works, just sit back and relax, start getting your bookings. But for this new product it isn't like that, you need to do some form of work. So you can see how the partners are already trained to just enjoy bookings, and now we are telling them that, you know what, you need to do some work if you want to enjoy this product.

[00:20:56] John: So partners wanted more than just themes, fonts and colours. Those are the three things we introduced into the MVP. There is a concept: I have something around my property, I want to write about it and put it on my website, but we've not built that into the product yet. But we were able to learn that before, people want all that, and so many things, but these are just a few that I was able to put together for this presentation.

[00:21:28] John: Cool. And as you know, at Booking.com we're hiring. If you're interested please go to careers.booking.com.

[00:21:37] Melody: I have to say that. No I don't. But if you are interested in connecting with John or I, then we have our social media information there, Twitter and LinkedIn, and we'll be around for the rest of the day. But thank you for your time.