It's self-evident that TripAdvisor excels at product development. Since its start in February 2000, the company has become what is probably the world’s most-visited travel platform, reporting more than 300 million unique visitors a month.
In recent years, it has launched transactional products at an accelerated pace, moving from pop-up window referrals to hotel metasearch mixed with instant booking.
The Newton, Mass., company has also expanded into transactional products for air, restaurants, and tours-and-activities.
Everyone in the industry has heard of the company's mantra "Speed wins." But we wanted to find out more about what sets TripAdvisor apart from its rivals when it comes to product management -- and what's next for the company's products.
One person who has been well-placed to see this evolution is Adam Medros. Eleven years ago, Medros began work at TripAdvisor as a senior product manager.
Now he oversees all of global product and is a key figure relied on by the inner circle of CEO Steve Kaufer and CMO Barbara Messing.
Tnooz caught up with Medros while he was attending this week's Phocuswright Conference. This interview has been been edited for brevity and clarity.
Tnooz: TripAdvisor has a lot of products ... but a few get outsized attention.
Medros: Yeah. Like, Instant Book, which is transformative. The launch with Booking.com inventory is proceeding really nicely.
Yet when you look at TripAdvisor and how it continues to evolve, some themes stand out beyond the products that tend to get the most attention.
Our content base is accelerating. We announced at the third-quarter earnings call that we're doing 190 pieces of content a minute, up from 160 earlier this summer.
Restaurants is a huge part of that. I would argue we're the largest restaurant review site in the world, by a large margin.
Yet people still think of TripAdvisor as a hotel review site.
Tnooz: You’ve been at TripAdvisor for 11 years.... What has changed over time in product development?
Medros: We've had these evolutionary debates, where it's like, "Well, we make money in hotels, so that's what we focus on...."
The debate was always like, "Well, we don't make any money in restaurants, we don't make any money in attractions, so how much effort, how much time are we going to put into those products?"
Two big catalysts changed that. One was mobile….
Let me tell a story. In 2006, I remember being at a meeting where we were talking about mobile and TripAdvisor.
Someone said, "Well, when are we going to build mobile features for TripAdvisor?"
You got to remember, 2006 is pre-iPhone. People were like, "All right, give me a scenario where mobile and TripAdvisor make sense, because we're a hotel research company."
People would say, "Well, imagine I land in Tokyo and I don't have a place to stay."
And other people would say, “Who does that? Nobody does that.”
All these years later, you've got an entire industry around that. You’ve got HotelTonight, and you've got all these last-minute players, as well as TripAdvisor, in that space, enabling that.
So, one of our themes today is getting better at transactional. Not just Instant Book but if you really look at it: we bought LaFourchette last year, we bought Viator last year.
You're going to start to see a deeper integration on both of them. On mobile, on desktop, around transacting for a restaurant and transacting for an attraction.
We've already begun adding instant reservation for LaFourchette. Both of those sectors bring huge opportunities….
Attractions is still at that very nascent stage, where mobile is starting to be a tool where consumers can say, "Hey, I can show up in a city and I can book a tour later today, and I can get museum tickets." I think that market still has a ton to grow, lots of connectivity.
Tnooz: Interesting analogy.... What about personalization?
Medros: We continue to build around our community and our content. I put it that way because in the past, we talked a lot about personalization. I've been at TRIP for 11 years. Before that, I was at Amazon and Nordstrom.com. I've been around personalization for a long time.
There's one version of personalization, the Amazon version of personalization, which is like, "You're going to like this because you bought this."
It's hard to do, it doesn't work that great in travel, because who you are as a traveler changes. This trip, I'm a different traveler when I attend a conference in Fort Lauderdale than when I go with my family to Fort Lauderdale.
You can personalize for both of those, you can create personas for both of those, but it's just harder to do.
Instead, what personalization in our mind can be is that it can be about giving you a little bit better content and a little bit better clue around, "Is this the right hotel for me?"
It might be, "Have your friends been there?" from a social network connection standpoint.
It might be like, "Hey, we think you're going to like hotel X because you've previously reviewed or liked these other hotels.”
One of the features that is rolling out onto TripAdvisor now is extracting out of our content key phrases, and then giving those to you in a way that lets you really parse the content quickly and decide, "Is this the right hotel for me?"
Think of it as like hotel summarization, but clearly that also plays out into our entire corpus of content: restaurants, attractions. That corpus, that content base, is just growing massively.
Tnooz: So mobile has been transformative....
Medros: As I said, mobile is huge. ... 225 million or so downloads of our app, a massive installed base of users across our devices, people bringing their phones with them in market, and now you take this element of transactional, this element of personalization, this element of content, and you've suddenly got this thing that unlocks vacation like it's never been done before.
Lots of good thematic things coming out of those three areas.
We'll go deeper on restaurants and attractions this year, we'll go deeper, obviously, on instant booking, continue to expand that globally.
Tnooz: It seems almost as if, 11 years ago TripAdvisor, was a publisher, with all of its user-generated content and listings and it has now become more interested in the transactional side.
While at the same time and in the opposite direction, some of the other big travel players that started at the transactional side of things are now they're tip-toeing into the a media/publisher place. Not to the same extent or degree, but there’s a blending.
Medros: Yeah, absolutely. By the way, those companies are some of our best partners. Interesting “coopetition” dynamic to add to the mix.
Tnooz: Frenemies and all that.... What are hiring trends for TripAdvisor? Are you hiring more Adam Medros-type transaction folks now?
Medros: Look, the product management team at TripAdvisor has grown, so I run the product management team. The product team has grown from ... I think when I joined, there were four of us, three of us maybe.
Funny story, David Krauter and I were hired for the same job. They couldn't decide between the two of us, so they hired us both. We started on the same day. David went down one path. He now runs our SmarterTravel group.
Tnooz: Pure publishing all the way.
Medros: Yeah, and ultimately ended up at SmarterTravel. His background was CNET, so he brought a little bit more of a publisher, content background, and my background was more commerce. I spent ...
Tnooz: You had worked more on the retailing/merchandising side at Nordstrom and Amazon before TRIP?
Medros: Yes. My background was more that commerce side, and so I spent my time, my first few years at TRIP, on product and then search marketing, and then back to product.
To your question, it's a mixture. I think we've always hired product managers who are a balance between thinking about the business needs of the company and who are able to work with technology.
Our product managers drive our road maps and work really closely with our designers and our engineers and our marketing folks to bring it all together.
Tnooz: But TripAdvisor's historically not been a merchandiser.
Medros: Historically, it has not been an e-commerce site. We have to be really cognizant of the fact that the reason why we have so much traffic is ultimately because we help consumers find the right thing experience.
Afterwards, when we ask them, "Hey, was the hotel as you expected? Did TripAdvisor help you?" overwhelmingly — 93% is the number from the Phocuswright study we did a few years ago — consumers say, "The hotel was as described."
At the end of the day, our core offering is about trust and about community and understanding that the community's going to help you have a better trip.
It's also why reviews are overwhelmingly positive on TripAdvisor. People pay it forward in the sense of, "I write reviews because the community helped me have a great trip, avoid the bad things, not to come and complain about the bad things."
As we move towards more transactional modes, I think we're going to have to add some bench strength and some thinking around how to merchandise and how to think about transactional flows that our partners and our friends in Expedia and Priceline have had for years.
Now, the good news is that we're not shy about learning from the best, and evaluating the things that they do to make our own product better.
We've got a culture around iteration and testing, so at any given time we're running 30, 40 experiments on the site.
Some of those are UI experiments, some of those are commerce experiments.
We're not afraid to put some revenue at risk to figure out what the right answer is.
Yeah, we're going to need, from a talent perspective, to add to that base, and to continue to evolve what product folks are capable of doing.
That was a really long way of answering yes, our hiring is evolving to reflect our changing product mix.
Tnooz: It seems clichéd and suspect to me as an outside layman when an executive of any company says, "We have a culture that isn’t afraid of failure.”
How does that work operationally? Is there a way you allow for failure, incentive-wise? Is there something in terms of how you model behavior? Is it something about how the structured process for reviewing the results of testing a new product, and say, "Well, we have a hypothesis, we tested it, it didn't work, but no one gets fired.“ Sorry: this is a vague set of questions.
Medros: Let me tell you one anecdote that I always found really powerful. It's not a specific example, but it really speaks to the culture around us.
I remember a few years ago, I think we were coming back from a management offsite. Steve and myself and I think Barbara Messing, who is our CMO, were in the car together, driving back, and we were talking about an upcoming company meeting, a quarterly meeting.
Steve was asking, "Hey, what are we going to put up as the things we tried that failed?"
It was a slide that we sometimes would put up as ... By the way, product managers hated that slide, because they don't like seeing their name up in lights. "Hey, Adam was the product manager on feature X, and boy, it was a failure."
Steve said, he asked me, "What are you going to put up on the slide? What failed?"
I'm just like, "I can't think of anything that is worth putting up on the slide that was really a failure."
His response was, "Well, then, we're just not pushing hard enough on trying stuff. There's got to be failures."
Not in a, "I hope you guys fail," but in a, "Maybe we're just not pushing the envelope hard enough."
A couple things we do…. You've heard Steve’s “Speed wins” mantra. It's not just about that. We have this whole culture around transparency.
Product managers (PMs), when they're writing specs, they send them out to a particular company email address.
Anybody in the company can be on that alias and read the messages that PMs write that are basically: “Here's what I'm going to test. Here's the before, here's the after. Here are my goals."
The most common feedback that PMs get on their specs is not, "Why are we doing this?" but, "Your goals aren't clear, they're not aggressive enough. Do we care about that outcome? Should we test that? Is that worthy of getting time?"
Then every week, all the specs that get sent out, product managers present at a meeting called Product Review.
Product Review is, again, an open meeting, anybody in the company can come, and it's not an approval meeting. It's a peer-review meeting.
The goal is to really suss out things that people didn't think about, or to highlight these conflicts that are inherent.
We got multiple products and business units and goals competing for space and real estate and users' attention. We've got to have a clearinghouse of how to sort out the, "Well, we test this."
Generally speaking, we'll test almost anything.
For instance, one of the teams that works for me is a team called revenue optimization. They're just all about AB-testing revenue ideas. "Hey, does this idea make the company more money? Does it make our partners more money, therefore making us more money?"
They run, I would say ... In any given week, they're responsible for eight to 10 of the 30 experiments running on the site at any given time.
In their goals, in their quarterly goals, they have a revenue goal, they have goals around trying to change certain behaviors. But they also have a goal about pure experimentation.
"How many experiments did you guys run?" Every time we ask them, "Hey, is that really the right answer? Should we have that goal?" they push back.
The leader of that team pushes back, and she's like, "No, I want my team to be thinking about not just what we test and the outcomes of what we test, but the pace."
That pacing and that expectation of like, "We're going to test lots of little things," because when you hit on one of those little things, it's like pulling a thread from a ball of yarn.
You start to unwind this thing of like, "Hey, there's an interesting thesis here, now let's go deep on that thesis and figure out where does it take us?"
Sometimes it takes you nowhere, and sometimes it actually turns into something that is not just a local maximum. Then how do you break out of local maximums and get to new step-change points.
Tnooz: So you encourage incremental improvements.
Medros: Well….this culture of A/B testing that's grown up over the last 10 years, has a downside in that you can get this massive incrementalism.
That can become a bad thing if you don't ever pause and deliberately make an effort to find the step moments and the step changes.
Tnooz: What I'm trying to suss out there ... It seems like there's a dialogue where the product team can push back and say, "Why are we testing this thing? Why are we doing that?" Is that a lever to fight against incrementalism?
Medros: No, it's actually really easy to fall into the incrementalism. It's easier to test something that takes three days than it does seven days.
Probably, from a leadership standpoint, the thing that's hardest in building a culture around A/B testing, is to actually, occasionally, at product review or in our own discussions, to occasionally ask the “So what?” question.
"So what if that works? If that metric goes up by three percent, that's your goal for this project. Let’s say it will go up by three percent if we do the test. Do we care? What if it goes up only two and a half percent? Are you going to stop working on it?"
Sometimes a PM might answer, "Well, three percent's a placeholder, I don't know what the number's going to be." What are you shooting for? What moves the needle? What's going to be an inflection point?
It's OK if you don't hit the inflection point on the first shot. You probably won't. Are you working towards that, or are you working towards a series of tweaks, a series of minor changes?
For PMs, that's a mentality they've got to build, which is understanding how to balance that incrementalism with, "Hey, this points toward something that might be a step change. Now let's go deep on the step change."
Tnooz: OK. When would you guess product review has gotten imbued in the culture there?
Medros: This product review process ... When I first started in 2004, I think we were small enough that you just didn't even need it. You just e-mailed the whole company.
Tnooz: Didn't have to be formalized.
Medros: Right. "Here's what I'm building." I still remember one of my first projects. You'd walk over to one of the engineers, you'd say, "Hey." Gene was his name. "Hey, Eugene, here's what I'm thinking about building."
He's like, "OK, that's not going to work and let me tell you why. This system doesn't work that way, but here's what we could do."
We'd talk about it at his cube, he'd be like, "Yeah, come back in a couple hours. I'll give you a test link." Build it, test it, launch it, and away we go.
As the company got bigger and a little more structured, maybe around 2006, 2007, we started saying, "All right, so we need to formalize a little bit, like written specs, shared with folks in a more formal way, product review meeting weekly just to align all these different goals and all these different pieces."
It continues to this day. The biggest challenge, as we've grown, is the sheer volume of stuff that's happening every week.
From a tech and development standpoint, the amazing thing is that we've done weekly releases forever. I can't remember a time we've ever missed a weekly release.
There've been some weeks where we've said, "No release this week because it's frozen," like over the holidays or something. For the most part, releases every week.
We're actually starting to accelerate that. Recently we’ve been doing two releases a day, and we’ve been looking into getting to the point where we could release features all the time, continuously, in a way that just gives us an amazing ability to multi-factor test things at the same time.
It's really incredible to have watched it grow and to see these product teams who are just trying to parse apart, "My tests running against your tests, how do they do, where's the effect of and the signal out of the noise and all of this."
We are always learning how to test more effectively.
Tnooz: If there's a product manager who’s, say, a few rungs of the ladder down from you in their career — at any company, not just TripAdvisor — and they're sort of saying to themselves, "I wish I knew how Adam Medros is consistently recognizes what’s truly important to prioritize in building their skills and in advancing their team’s goals.” What advice could you give someone looking to advance in a similar trajectory to the top performance you’ve shown in your career?
Medros: That’s a tough question to answer on a whole bunch of levels. The thing that ... One of the things that I always felt was important to me was the idea of building a narrative.
I think of our product road map as just a narrative. That narrative is just always evolving. It's not that we've got some ... Maybe Steve does ... But the rest of us don’t have some giant master plan that lets us see three years in the future. In fact, quite the opposite.
The thing TripAdvisor's always been good at, and the product organization that TripAdvisor's always been good at, is taking the here-and-now.
We start with, "What are some of the themes that we're focused on in the company, and then how do those themes play out into a narrative around what our users want, where the industry's going, where the opportunity is, and where our strengths are?"
Look at Instant Book. The narrative of Instant Book is, in retrospect, not hard to see.
We went from a site where we help people read reviews and ultimately decide what the right hotel for them was, or if this hotel was any good for them, and then click off to a partner and book it.
Then that evolved from a pop-up model to a meta-model. That was a pure user experience evolution, as the Web evolved from a pop-up model to a meta-model.
We probably stuck with the pop-up model too long, but it worked really well. Our partners liked it.
Sure, people would say to me, "God, you guys pop up a lot of windows." I'm like, "But we pop them up on the hotel you're looking for to help you find a great price."
Tnooz: It was relevant.
Medros: Yeah, that's right. “I hate it, but I love it.”
OK, so metasearch came along. The narrative switched to a better inline user experience. Price comparison, all those things.
Then, as mobile grew and evolved, what we heard from consumers was, "This meta thing's fine, useful. Price comparison's really important to us, that's why we use you."
Again, "You help me find the right hotel, you help me make sure I get a great price, but yeah, can't I just book it on you guys? I don't want to go to this other site, where it's not mobile-optimized," or "I don't want to go to this other site that I don't know where I'm putting in my information and giving out my e-mail and all these other things.
At the end of the day, consumers ... Instant Book came about because we looked at the friction in actually booking the hotel, and listened to our consumers who said, "I just want to book with you guys," and we said, "OK, we can facilitate that.
We can do it in a way that's friendly to our partners, doesn't put us in competition with our partners, delivers great leads to suppliers and great leads to our partners, and the consumer's happy."
Back to your question. That ability to paint a narrative and then to think about, "All right, given how we operate, given this iterative approach, what's the balance between big, swing-for-the-fences, thematic things we're going to try this year, and incremental improvements to get there?"
An evolving product manager, a good product manager, is able to ... Look, you got to be able to get stuff done, work with engineers, paint a QA, do analytics.
Tnooz: Right.
Medros: Then, as product managers grow in their career, what they actually do more and more of is they tell stories, they build narratives around where we're going and how we get there.
Tnooz: And communicating them downward, and then upward, is that right?
Medros: Yeah, and across. Just making the case for, "Here's the vision, here's how we get there. Hey everyone, get on board."
Then all those things that you talk about any good person's going to be a good ... I think product managers also have to be this nice balance between a good listener, but they're not trying to do everything by committee. They can't do everything by committee.
Tnooz: Right, it's not the same.
Medros: Yeah, they got to listen to all their stakeholders, and they got to be able to play back. "Hey, Sean, what I heard you say you need is X. OK, I'm not going to get to that in D1, I'm not going to get to that in D2. Here's why. Here's what we might get to in B3. OK, you've got a concern about that? I got my own set of things I'm trying to balance. If we're in conflict, let's go escalate that and figure out how to resolve it."
They're good negotiators and communicators as well.
Tnooz: Is there a book that has helped you and might help others?
Medros: Yeah. Marty Cagan wrote a book called ... Marty Cagan is an ex-eBay guy. I don't know if you know him.
Tnooz: I don't, but many Tnooz readers will.
Medros: He's got a blog called "Silicon Valley Product Group," SVPG.
Every product manager should be reading Marty's newsletter every week. He wrote a book called "Inspired," and it's a collection of a bunch of his blog posts about product management.
It's really ... I don't agree with absolutely everything. Marty believes more in prototyping than I do, but I think that's very company- and culture-specific.
The essence of what he talks about is how to build products, and how product managers function. It's as good as it gets. Marty's great.
NB:Illustration of an A/B comparison test via Shutterstock: