Expedia Group, Inc.
Q3 2020 Earnings Call Transcript

Published: 4 Nov 2020

[Operator:]

Good afternoon. My name is Christine, and I'll be your conference operator today. At this time, I would like to welcome everyone to the Expedia Group Third Quarter 2020 Conference Call. All lines have been placed on mute to prevent any background noise. After the speakers’ remarks, there will a question-and-answer session. Thank you. I'll now turn the call over to Michael Senno, Vice President, Investor Relations and Treasurer.

[Michael Senno:]

Thanks, Christine. Good afternoon, and welcome to Expedia Group's financial results conference call for the third quarter ended September 30, 2020. I'm pleased to be joined on the call today by our CEO, Peter Kern; and CFO, Eric Hart.

[Peter Kern:]

Thank you, Michael. And good afternoon, everybody. I hope you're well. We appreciate we are not the most newsworthy thing going on right now in the country, but thank you for spending a little time with us, and hopefully we can help give you some more color on our quarter. I'll start by saying, as I said before, that travel generally performs in this period as everyone has expected. And in general closings, openings, worries over the pandemic, have the - the exact influence you expect. But we were pleased to see in the third quarter that we’re stabilizing travel trends and significant improvements on our cost structure, we were able to post markedly improved performance. I'm going to get into some of the internal workings, including margin, expanding on it, which actually Eric will cover in his remarks. But first, I just want to make a few comments on the recent trends in travel. What we saw over the third quarter was basically a stabilization, as I said. July was down in lodging gross bookings net of cancellations. This was a function of two things. You may recall there was a slight increase in COVID cases being reported then, and that had some impacts on travel, along with Vrbo for us, which had a very, very strong June. And that strength subsided a bit. It still was strong in July. So July again, lodging gross bookings net of cancellations was up about 65%. And in the months that ensued, that got to the low to mid-50% range down. So a reasonable improvement, a steady improvement, and that obviously helped with our third quarter results.

[Eric Hart:]

Thanks Peter. And thanks, everyone, for joining as well. While we continue to see significant year-over-year declines, in our business in the third quarter, taking into account the impact of COVID is having on travel, our financial performance was better than we expected, with over $300 million in adjusted EBITDA and reaching essentially cash flow neutral in September, for the first time since February. As you know and I've discussed in previous quarters, we are keenly focused on driving margin expansion, which largely falls within three buckets, resetting the fixed cost base, reducing variable costs of revenue, and increasing marketing efficiency. I'm going to touch on, excuse me, each of those individually. So first, on fixed cost bases, as you'll recall, we started the year targeting $300 million to $500 million, in annualized savings. And we were tracking ahead of that amount, as of our last call. Since then, we've identified additional headcount reductions in certain areas. And incremental opportunities across the P&L to drive efficiency, in areas like real estate and software and licensing. We are now targeting $700 million to $750 million, in annualized run rate savings, compared to our 2019 exit rate. And we've already actioned over $550 million. As you project this forward, please keep in mind, we'll have annual increases in the remaining cost base. And also make targeted investments in some areas. But overall we've made great progress on fixed costs. And longer term, our platform operating model will position us, to scale the business, far more efficiently going forward.

[Operator:]

Our first question comes from the line of Lloyd Walmsley from Deutsche Bank. Your line is open.

[Lloyd Walmsley:]

Thanks. Two, if I can. First, I just wanted to make sure I understood you correctly earlier. When you said - I think you said the core fixed cost savings target is now $700 million to $750 million. But then you said you've been able to reduce variable costs, if I heard you right. Which overlaid on 2019 for example, would be an additional $200 million in savings, so approaching $1 billion on kind of a normalized basis. So the question is just did I catch that right? And does it feel like you're getting to the kind of edges of the scope of savings there, given the magnitude we're talking about? And then the second one would just be, as you work to bring all the marketing data from separate brands into kind of a consolidated kind of new data storage for marketing. Any - is it early enough that you – in enough that you can see some early learnings, or is it too soon to say in any sense for the timing on when you guys will be in a place where you feel good about that, that transition. Thanks.

[Eric Hart:]

Peter, do you want me to take the first.

[Peter Kern:]

Yeah, I'll take the first you take the second. So I on the first, yes you did hear correct may so the $700 million, $750 million is on the fixed cost basis. It's in relation to the $3 million to $500 million that we spoke of earlier in the year, of the $700 million, $750 million we've actioned $550 million of that, so we are feeling like we are making good progress against it. We of course continue to leave no stone - no stone unturned. And just as we continue to work through it, we get more cost that we've identified, we get confidence and those that we get better line of sight in. And then also, obviously we've actioned a fairly large percentage of it. So yes, it's the $700 million, $750 million on a fixed cost basis, and then the incremental $200 million on the variable cost of sales with spent approaches to your math, approximately a $1 billion in savings. Now you mentioned it, but I just want reiterate as well, that is on the 2019 exit rate if you will, so that – those numbers presume that we’re back in a range that is approximately at the 2019 levels. And then lastly, there will be some offsets for that, over time, you can imagine inflation on contracts or increases and headcount costs, et cetera. So there will be some offset. But we feel pretty confident in the numbers that we're going after and then continue to – the teams are doing a great job. And I guess it was the last part of your question just around, are we reaching the edge of it, and I would say, you know, we're still flipping over rocks. But it's, you know, diminishing returns to a certain extent, so I would say the range that we shared with you is sort of the probability adjusts where we think we will get.

[Eric Hart:]

Yeah. And Lloyd, I'll just jump in on the other question, I would say the answer is no, we haven't. It's not wired together yet in a way that we can see the early wins. I think we are, you know, getting closer, but it's not simply a matter of taking three different groups and putting them together. I've mentioned on previous calls that we've been also consolidating all our data. And, you know, doesn't - you know 17 data into hopefully one soon. And all of these things, you know, are kind of linked together. So there's a lot of great work going on. I think we are getting closer and I'm excited about what we're going to be able to see when we can see the granular level of detail and profitability by, you know by customers, by geos, by everything, which we haven't had the granularity we like. And that has made us, I would say blunter instrument then we’d like to be and we believe there's significant you know, good guys from getting this all wired together and that again is why we want to make short term trade-offs to that important work, to get that done, but we haven't seen any early returns yet. I think will be a little while still.

[Lloyd Walmsley:]

Okay, thank you. Impressive job on the cost side, guys. Thanks.

[Peter Kern:]

Just one other quick add to that and thanks for the question again is. We are still taking the action to achieve those numbers, so just make sure as you're building those, and there's the lens of we need the volume to come back to see particularly on the variable cost side. And we are still actually in the – and as I mentioned, the 550 on the fixed cost side is going to take us still some time to get to the gap of that and the seven to 7 to 750. And the same thing on the cost of revenues as well. It will take us a bit of time to get there. But that's what we're shooting for and what we think we'll get.

[Lloyd Walmsley:]

All right. Yeah, that's helpful. Thank you.

[Operator:]

Your next question comes from line from Naved Khan from Truist Securities. Your line is open.

[Naved Khan:]

Yeah. Thanks a lot. Just a couple, one on the costs side. Eric, you called out the incremental $200 million in variable cost savings, and you named three areas. We tried to think about the relative magnitude of savings in each of these areas? How would you stack them up or number them? And then I had a question on the gross bookings side. So you talked about a plateauing in October versus Q3. Is that overall for the business or just US? How should we think about that? And is that plateauing on gross bookings or on net basis? How should we think about that?

[Eric Hart:]

Yeah. I'll take the first one. And Peter, perhaps you can take the second one, but I'm happy to take it as well. But on the cost of revenue, there's three components. There's the conversation platform and the cost of savings, the credit card and the variable. And I would effectively put them in that order to a certain extent. But I think the one that has the largest opportunity for us is on the conversation platform. We are historically primarily a phone-based contact center-type of conversations with customers and with our supply partners as well. And with the technology that we've developed, we have the ability to move a lot of those either to self-service or virtual agent. And on the virtual agent side, we are able to get more - service more customers, if you will, per agent. And so both of those add leverage to the system. And we've really started the rollout into that. We've gotten good NPS scores and good returns, if you will from doing that. And it's really a process of rolling that out, making sure that we have more and more use cases, we get it embedded into the right flows into the site, into the mobile apps as well. And I think that is likely the largest opportunity. And then credit card costs and variable cloud as well are meaningful, but I would say that those two are a distant second, if you will, after the conversation platform opportunity.

[Peter Kern:]

Yes. And then I guess Eric, you can follow on, and I'll do cover…

[Eric Hart:]

Yes. On the volume side, I think this recovery is going to be a series of many stories. And so if you take hotel, for instance, and I could say that that's been flat for the last couple of months. But if you de-average that number, what you actually end up seeing is that there are areas that are flat to declining, and there are areas that are gradually improving. And so using a hotel as the example, Europe has been more challenged. Peter already mentioned that. And they've got even more challenged over the last week or two, just given the increasing lockdowns that are reoccurring. Whereas the US is had a more steady increase over time on the hotel side, we will see what happens with the wave three that's occurring right now, and something that we're looking at very closely. But as we say it's plateaued, it's actually many stories underneath. And I think we're going to see the same thing. We've got international versus domestic. And we wait in some regions of the world in one versus the other. And that's the series of many stories that will make up how we see overall the business is performing.

[Peter Kern:]

Yes. And the only color I'd add to that is our numbers also include bookings from our B2B partners and people in other parts of the area. So it makes the many stories even more vast in terms of our exposure to different geographies, our exposure to different types of business. So I would say thinking of it as a whole is probably the simplest way, but all things being equal with the trend the way they are in the last two weeks, its probably good news that we're weighted towards North America. But there have been times when it's been something else. And we wish we had more China business since domestic China is doing quite well relative to the rest of the world.

[Naved Khan:]

Got it. Thank you, Peter. Thank you, Eric.

[Peter Kern:]

Sure.

[Operator:]

Your next question comes from the line of Mark Mahaney from RBC. Your line is open.

[Mark Mahaney:]

Two questions, please. You mentioned market share, but in North America, do you see any evidence of material market share shifts amongst the different providers, Airbnb, Booking, Expedia? Secondly, you mentioned those three areas of cost efficiency, the fixed, the variable, and then you talked about performance marketing. Can you just drill down a little bit more on performance marking? I know you've had some pretty bold robust plans for bringing performance marketing efficiencies. And then all of a sudden, we had COVID. Is it clear to you that you've got enough evidence that you can really - that there are new found efficiencies in performance marketing? It's still TBD to when we get real robust recovery in travel conditions. Just spend a little bit more time, please, on performance marketing efficiencies. Thanks a lot.

[Eric Hart:]

Peter, are you there? Did we lose you?

[Peter Kern:]

Oh, sorry, sorry, sorry. Unmute button was on. Sorry, Mark. I'll jump in. I was saying, I'll take the second one first, and then you can remind me the first one. I would say on the Performance marketing side, you are correct. We have talked historically about bringing more efficiency out of it, getting our return on marketing better, more precise, et cetera. The problem was I think that was to, to some extent, wishful thinking because we did not have all the data, all the tooling and all the approaches synchronized, if you will. We let our brands compete. That has some dis-synergies that never were quantified, but certainly existed. We did not have the benefit of all the data, because each brand had its own data and on and on and on. So I am positively optimistic that when we have all data flows right and we have all the algorithms rewritten for that and when we have the tooling right, there should be significant upside for us in actually getting – bringing out that return that you've heard about over these many years. I think that is what is required. And it is real, and we will get to it. We have to do it obviously. We haven't done it yet. But to otherwise sort of mess around the idea of taking one brand higher or lower, it was a very blunt set of instruments and I think candidly, it was optimistic to think we could do a lot better. We certainly could have run more profit at the expense of share. That we certainly made some trades for share over a profit historically. And we certainly make trades like that in certain geos and for certain reasons over time. But I think more broadly, the opportunity to just perform better is really unlocked by stringing all of those work together and really making it powerful. So I'm a total believer in it and I think it will happen. But we haven't done it yet, so we have to do it. And then sorry, Mark, your first question? I apologize.

[Mark Mahaney:]

Yeah. Do you think that there's clear signs of a market – you brought up market share. Have there been any signs of market share shifts in North America?

[Peter Kern:]

Yeah. I think well, definitely you've seen alternative get an outsized share relative to where it was. And we've been beneficiaries of some of that. Airbnb have been beneficiaries of some of that. So the more concentrated you are towards the things, the use cases that are working well, I think you're going to get share benefits. The – more broadly, I think share has moved around for a lot of little reasons, and you're dealing with low volumes. So I look at markets where share points move around. And then you look at the dollars related to it, and they're tiny. And you say, okay, it doesn't really matter. It could just be a bad weekend or good weekend. But I think nothing we're seeing is – portends to anything for the future for like a long-term effect. I don't think there is anything that suggests anyone's winning or losing in a way that will give them long-term benefit. I think for us with Vrbo, we believe it's really good for the brand and people are getting a lot of exposure to that use case. And people are seeing how nice that vacation can be. I don't think that means that people will never go back to hotels, but I think it helps us relative to Airbnb. So that's good for us. And I think beyond that, we don't expect anything going on in any single geo or overall is really a long-term effect, which is partly why we're not overly exercise if - we're not perfectly tuned as a machine to – on the performance marketing side because none of this is with levels down 50, 60, whatever plus percent in certain markets. It's hard to think that any of that is sustainable or meaningful for the long-term. So that's been our approach.

[Mark Mahaney:]

Okay. Thank you very much.

[Peter Kern:]

Yeah.

[Operator:]

Your next question comes from the line of Eric Sheridan from UBS. Your line is open.

[Eric Sheridan:]

Thanks for taking the question. Maybe two, if I can, following up on some of the topics talked about so far. On Vrbo, with the success you're seeing in terms of the demand and the new customers that are coming into the platform, is there anything to call out in terms of your go-to-market strategy, your marketing strategy, that you might think portend for sort of longer term, better returns on marketing spend or margin structure in Vrbo as a business? And then on your brand strategy and sort of the realignment of the marketing platform, are you also of the same time thinking through whether you need to be in the same number of brands? What I'm saying scope of brands on a multiyear view towards an eye of like managing a mix of sort of growth versus profitability? Thanks, guys.

[Peter Kern:]

Sure. So let me do Vrbo first. Yes. Well, a few things. One, there's a lot of work that's been going on over the last several months on Vrbo and its margins in terms of – Eric mentioned or we've mentioned in the past. Bringing it on to our payments platform, that's been good for us, saves us money with a third-party provider, allowed us to use power of Expedia's payment platform, which itself is getting stronger every day and more efficient. So those kinds of benefits are there. There's work being done on the consumer process, on the fees and other things that – where we think we're under market. So I think we've had opportunity and we found opportunity to withstand margins at Vrbo. But to your bigger question, we definitely think that we did not land the Vrbo rebranding as strongly as we wanted to, that there is opportunity now given that so many people are using it and the use case is so attractive right now and that we are concentrated in whole home, which is like the most attractive part of the alternative market right now that we intend to and are already pushing into a much more aggressive part to broadly push the brand. We've turned the brand over in a few more markets in Europe, just a few weeks ago, and we will keep pushing. But yes, there's an opportunity there, and we think that is a place we want to be more bold even in a down market, and we will continue to be pushing into that. To your second question about how many brands we have, yes, we have a lot. We certainly won't keep any - that we don't think makes sense. But right now, we think the number of brands creates opportunity, as long as we have segmented them and thought about them geographically, et cetera, to their greatest effect. So my big push has been, if we have a brand, frankly, that maybe none of you have heard of, like Wotif in Australia, that is our strongest brand for the moment in Australia, then we need to concentrate on building that brand and not worry about whether it's Expedia or hotels.com or anything else. Likewise, Vrbo has a different brand, a different company that acquired in Australia that we're not going to change that brand, because it's a strong local brand and we should push into that. We have historically had this record of, okay. We have an existing brand or we bought a buy brand. We're also going to push Expedia into that market. We're also going to push Hotels. And that kind of, I would say, took away the value of multi-brand, because it didn't allow us to optimize for each one. So that's why we're doing the segmentation work. That's why we're trying to figure out what the proposition is for each brand. Some brands admittedly, we have historically run for more profit, whereas others we have pushed. We are looking at the whole kind of plate of options and driving the best result by country, by brand that we can. And that may mean, we closed some brands in some countries. That may mean, potentially, we close some brands, we’ll stop. But we have no intention right now to do that, but we are looking at the whole thing and trying to optimize for it all. But we think, in general, having more gives us the opportunity to do more, if we do it smartly.

[Operator:]

Your next question comes from the line of Deepak Mathivanan from Barclays. Your line is open.

[Deepak Mathivanan:]

Hi, guys. Thanks for taking the question. Just a couple of ones on Vrbo. So first, can you talk about where you are currently in terms of in a merchandising Vrbo's augmented accommodations inventory on Expedia and hotels.com? Is that something that you're seeing benefits from currently? Is like, this is a perfect time to kind of channel logo inventory at a higher level into your other strong brands, since consumers are looking for that type of inventory? And then second question related to it, how you feel about the inventory levels on Vrbo, particularly given the strong demand that's there for authentic foundations. Are there any markets where there is inventory pressure on Vrbo at this time? Thanks.

[Peter Kern:]

Sure. So a couple of things there. I would say on the latter part, no, there's - we've seen some pressure over the summer a little bit in a few markets, but we haven't generally seen literally not – no inventory. There's a few spots, but in summer, recall, was everybody getting headlong into the only alternative they were comfortable doing. So there was a really acute demand spike there. We do think there's more opportunity to serve them - more opportunity. And I think we will continue to try to grow our inventory. The question is - and obviously anywhere we see demand, we will want to drive inventory. The question longer term is, we’ve historically had deficit, in particular, in cities in studies and those kinds of use cases, and that is a bigger strategic question for us. It's a place where Airbnb obviously has made great hay, although it hasn't helped them during COVID. Likewise booking has done a nice job with city-based inventory, and to your point, a more integrated experience. And so moving on to that question, the product is not great when it comes to surfacing vacation rental opportunities or alternative foundations through our main OTA brands. We've been slow to get there and it's not where we want it to be. It's definitely an opportunity, but we have found that there is a little bit of a conflict from the standpoint of people don't generally go to our alternative – go to our OTA brands for alternative accommodations. It's not what they think of. So even though it's there and present, it hasn't performed perhaps the way you or I might think it would. So I think we've got a combination of work we need to do on that front, which is we can definitely make significant improvement in the product and how the content will surface, how people see alternative accommodations and where they see it and et cetera. But also we need to lean into it, and we need to market particular in places where our big OTAs are – have good reach and brand recognition and Vrbo does not, it may well be more sensible for us to lean into you know, get a vacation home on Expedia as opposed to trying to introduce – called the Vrbo brand. So we are looking at all that. I would say dual work streams where we're trying to materially improve the product experience with alternative accommodations on the main OTAs, and then how do we market that and how do we test to push that into those markets. And I'll just say while that seems like that would be job one, and we'd be all after it because of the appeal of alternative right now. You have to keep in mind that we are re-architecting literally our entire technical infrastructure and platform. And some of these things have to sort of get prioritized. And it's in the works, but it's not the only thing we're working on.

[Deepak Mathivanan:]

That’s help. Thanks so much.

[Peter Kern:]

Yeah. Thank you.

[Operator:]

Your next question comes from the line of Justin Post from Bank of America. Your line is open.

[Justin Post:]

Great. Thank you. Maybe first question on hotel supply, are you seeing any inroads there or good deals on Expedia that you might otherwise get, any reason for consumers to come directly to Expedia versus other sites? And then secondly, thanks for the Vrbo update. I know you said it grew in 3Q. I'm assuming that's bookings and revenues. Can you give us a year-to-date update on Vrbo? Thank you.

[Peter Kern:]

Eric, you want to start to answer the second one, I'm not sure we can.

[Eric Hart:]

I don't - we've not shared or disclosed on the Vrbo. The year, I would say we are healthily up for the year. We clearly had the spike in June, just given the summer compression where people felt comfortable traveling. We continue to be up year-on-year in Q3, let's call it at more reasonable levels than June. But both revenue and gross bookings were up year-on-year, and I think you can extrapolate back to the start of the year as well.

[Justin Post:]

Got it. And then on hotel supply?

[Peter Kern:]

Yeah. On the hotel supply side, we have obviously among the most competitive hotel supply in the world. There's not a lot of discounting that is unique to them. The hotels are discounting, but I don't think there discounting is unique to us. And I think what we provide, what Expedia provides obviously beyond just very good value, is a bunch of ways to find what you want, a bunch of ways to book multi-product, a way to do everything in one place. So there are a lot of offerings with hcom, hotels.com, you get very robust rewards. So there's other offerings that are differentiating for us on the brand side, including where we're trying to go with a product to get a much better personalization, helping make good value choices, et cetera. So, I would say that is where we have to push our business, the differentiated hotel supply, even when we can get it is usually not sustainable. It's not like something that's consistent, but we do want our hotels to be in a better position to optimize our platform. So we are doing an inverse, which is trying to give them better data so that they can be much more effective at pricing and moving the inventory they want to move on our platform. So if that's suites or certain kinds of rooms or whatever, we want to be heating them, the inbound information, so they can understand the market, understand where there's opportunity to price up or down to gain both rooms, et cetera. So that's where we're pushing, and we think that's going to make us a much more valuable partner for the suppliers. But I don't think it's going to be about, if you get a $5 better deal than the next person. I think it's really going to be about how you can help them build their business.

[Justin Post:]

Great. Thank you.

[Peter Kern:]

Yeah.

[Operator:]

The next question comes from the line of Kevin Kopelman from Cowen. Your line is open.

[Kevin Kopelman:]

Thanks a lot. I just have a quick follow-up on Vrbo. If you could comment on the seasonality patterns there that you're seeing into Q4 and the current environment? Are you seeing any change there, any kind of continued shift into Vrbo even as we get outside of what is typically the peak season? Thanks.

[Peter Kern:]

Yes.

[Eric Hart:]

Sorry, go ahead.

[Peter Kern:]

No, no, no.

[Eric Hart:]

Thanks for the question. Vrbo is I mentioned earlier, it is much more seasonal than the rest of our business. So it's concentrated in the summer months, even more so than North America given summer. So we would naturally expect to see it come down into Q3 and into Q4, and that is what we're seeing. It continues to be at healthy rates and continues to be a product that I think consumers, in the environment that we're in particular, find compelling to be able to continue to travel. So you can think about our – the normal travel seasonality of looking in the early part of the year to stay in the middle part of the year and then less activity in the latter part of the year, just magnified in Vrbo relative to the rest of travel.

[Kevin Kopelman:]

Right. And then – well go ahead.

[Peter Kern:]

No, I was just going to add that, that is an area we believe there's opportunity and a lot of the advertising and marketing we've been doing with Vrbo have to do with sort of Staycations, if you will, in North America where kids may not be going to school, people may not be going to work. Sort of take your life somewhere else, and we think there is opportunity to break some patterns, while this going on. But that is just a COVID-centric issue.

[Kevin Kopelman:]

Got it. And then longer term I think on a like-for-like basis, Vrbo's take rate has been a little bit lower than traditionally you would get with hotel. How do you see that playing out longer term in term of revenue take rate?

[Eric Hart:]

Yes, there is a – I'm happy to take that one. Thanks for the question again. There a is gap between us and Airbnb in particular, and it's something that we do feel that it's – it gives us a bit of an advantage with consumers, but it is something that we are looking at. This is something that we test. We do think there are opportunities to increase monetization over time. But I would say there continues to be a gap, we are actively looking at that. We are actively testing it. And I do think that there is upside in our monetization over time.

[Kevin Kopelman:]

Thank you.

[Operator:]

Your next question comes from the line of Jed Kelly from Oppenheimer. Your line is open.

[Jed Kelly:]

Just on Vrbo, just circling back, I mean can you talk about the supply into the winter travel season? It seems like that's going to be premium supply, and then how you think you're there. And then just a follow-up on virtual agent, I mean, what's your confidence level in virtual agent as travel volumes start to normalize? Like is that just around handling normalized volumes?

[Peter Kern:]

Yes. So just on the Vrbo supply, I think we are well equipped and well quoted I think you know, we think they are watching good-value, high-value inventory show up in lots of places people want to go in the winter, do that for Thanksgiving, Christmas, et cetera. And we think there's an opportunity that those trips will be longer in duration than they typically are, given school holidays, et cetera, around the U.S. in particular. But we don't feel like we're deficient on the inventory side. That does not mean we don't have opportunity, and it doesn't mean we have parity everywhere on everything. But we think we have plenty of inventory to power a very robust winter season if the demand is there. So we're not worried about that. And sorry, the second part was...

[Jed Kelly:]

Confidence around virtual agent.

[Peter Kern:]

Yes, the confidence virtual agent. So I would say to you a couple of things. One, despite the fact that our volumes are down, given COVID, given cancellation rates, given all the issues that travelers have, I wouldn't necessarily assume that our servicing volume is proportionally down. Because we've had – there's been a lot of traveler issues going on in the world that have created huge relative spikes in traffic. So we've been testing the conversation platform into all kinds of use cases. And we will ultimately have it in virtually everything we do, including our dealings with supply partners and other things. So it is being rolled out broadly. As Eric said, the NPS scores have been strong on it. People generally like to deal with the machine if they can and just get it over with and not have to talk to a person. It's not true universally, but that seems to be true for most people. And we feel very good about it. So it's scaling up. It's a platform tool. It's the kind of tool we want to have in our company, across the company, that we have a platform technology that needs to be taught and taught different use cases and applied consistently. So we have high confidence in its ability to continue to roll out and handle as much traffic as comes in. So no issues there.

[Jed Kelly:]

Thank you.

[Peter Kern:]

Yeah.

[Operator:]

We will now take our final question, Lee Horowitz from Evercore ISI. Your line is open.

[Lee Horowitz:]

Great. Thanks a lot for the question. Just one if I could. Peter, you had mentioned some of your North American seeing greater direct bookings, perhaps as people shop less. I'm wondering if you could comment at all on your mix of direct bookings through the quarter. How that may be trending, what you're seeing through this crisis and perhaps your view on how sustainable those trends may be beyond COVID? Thanks so much.

[Peter Kern:]

Yeah, surely. I would say that what we've seen is mix-wise, we've mixed towards a lot more direct traffic. Bookings are up considerably. Direct is up considerably. But I would caution anyone from taking too much from that because we have been less aggressive in the performance market. So it's a little bit of self-blending, if you will, down to a mix that is more free, which of course has helped all our margins and helped us post the numbers we posted last quarter. But that is inherently in part because we have taken our foot off the gas in certain areas in terms of performance marketing and uprofitable performance marketing. So I think again as we blend back into normalcy, assuming we will be more proficient and have these tools we've discussed about on the performance marketing side, et cetera. My sincere hope is we can grow more share through those performance channels at attractive levels and attractive profitability. And I believe that will happen and if that's true, of course, we'll mix our way back to a more historical balance. But the trends have been very good, credit to our brand loyalty, or brands loyalty I guess. And app has been very strong. And of course, we are putting lots of energy into improving the products for all those things and make them stickier, because it is gratefully not been a topic today, we do want to have more direct traffic, avoid the performance marketing pools as much as we can and create recurring business. So that remains a core goal for us.

[Eric Hart:]

Hey, Peter, before we end, I just wanted to clean up one quick comment I made earlier, just on Vrbo on the year-to-date. Year-to-date, we have seen declines due to the impact earlier this year on COVID. Q3 was up year-on-year, and we were growing early in the year before COVID. So just wanted to clean up the fact that in the early days, if you will of COVID that if you aggregate them for the year, it is down relatively, but was up before and up during. I just wanted to clean that up.

[Peter Kern:]

Okay. Thanks, Eric.

[Operator:]

Thanks. I will now turn the call back over to Peter Kern.

[Peter Kern:]

Great. Well, thank you, everybody. Appreciate the questions. Hopefully, we helped to give you some clarity. Again, we're keenly focused internally still on everything we talked about. And I suspect our next call will sound a lot like that. And hopefully the world's health will get better, and that will help our business along the way. But appreciate the time. Thanks again to our people, and appreciate you – listening into questions today. So take care. Bye.

[Operator:]

This concludes today's conference call. You may now disconnect.