Trivago NV (TRVG) 2017 Q4 法說會逐字稿

Good day, and welcome to the trivago Q4 and Full Year Earnings Release 2017 Conference Call.

Today's conference is being recorded.

And at this time, I'd like to turn it over to Mr. Matthias Tillmann, Head of Investor Relations.

Please go ahead, sir.

Thank you.

Good afternoon, everybody.

Welcome to trivago N.V.'s financial results conference call for the fourth quarter ended December 31, 2017.

I'm pleased to be joined on the call today by Rolf Schrömgens, trivago's CEO and Managing Director; and Axel Hefer, our CFO and Managing Director.

The following discussion, including responses to your questions, reflect management's views as of today, February 7, 2018, only.

We do not undertake any obligation to update or revise this information.

As always, some of the statements made on today's call are forward-looking, typically preceded by words such as we expect, we believe, we anticipate or similar statements.

Please refer to today's press release and the company's filings with the SEC for information about factors which could cause our actual result to differ materially from these forward-looking statements.

You will find reconciliations of non-GAAP measures to the most comparable GAAP measures discussed today in our earnings release, which is posted on the company's IR website at ir.trivago.com.

I encourage you to periodically visit our Investor Relations site for important content, including today's earnings release.

Finally, unless otherwise stated, all comparisons on this call will be against our results for the comparable period of 2016.

With that, let me turn the call over to Rolf.

Yes.

Welcome, everybody.

Many thanks for joining our full year 2017 earnings call.

We think -- beginning of the year, we think it's time for a bit of reflection.

So 1 year after our IPO, we are looking back on a very intense year.

We have basically seen the full bandwidth of emotions being a public company.

Still, going public was just one of many steps that we have done and that we will have to do.

I think it's obvious that we are not done yet.

But more than that, we think we just started.

So we are proud to say that the mentality of the organization has not changed during the last 12 months.

We stay focused, we keep on investing and we keep on learning in a very fast pace.

We defined our mission statement about 10 years ago and still it remains unchanged.

We want to use the opportunity today to reaffirm again what we think we are here for.

We want to match every traveler with their ideal hotel.

We are super convinced that at the end, it's all about building a superior user experience.

To build a superior user experience, we need the best search technology.

We need to focus on profiling users and hotels.

And we need to build the capability to match these profiles with each other.

We need to be aware that the user makes us responsible for the whole experience.

So when we speak about user experience, regarding the responsibility that we take for it, it does not end with a lead to another website.

It does not end with booking a hotel.

But it includes the whole user journey, including the stay.

We have to build a broad coverage of the whole accommodation market, including OTAs, including chains and individual hotels across the globe but more and more alternative accommodations.

At the end, we have to connect our offering to a strong, attractive brand that is able to get across the product value that we create.

In the volatile year, the different parts of the puzzle do not always get clear.

So let me sum up where we are standing right now.

We were exceeding the first time EUR 1 billion in revenue.

This might be interesting from a financial perspective.

But what counts for us much more is our scale in terms of data.

We massively increased the number of transaction on our platform, which helps us to learn much faster.

We used the last year to revamp our search infrastructure to cope with the new expectations that we have, including rebuilding of our major data flows and moving into the cloud.

We focused on marketing on long-term value creation to address the right users in the right moment and that are the most valuable for our advertisers.

With a new relevance assessment, we made lots of efforts to progress to create the more optimal user experience, also in the process of our users leaving trivago to book on an advertiser's site.

We started to broaden our offering by including more and more alternative accommodations because we know that our users do not make these differentiations.

They just want the best place to stay.

And last but not least, we drastically improved the way that we organize ourselves as a team, making sure that we have the best foundation to keep on learning in the upcoming years.

To get a good perspective, you have to do both, you have to zoom out and you have to focus back in.

Zooming out on the last years, we see continued strong growth following the exponential pattern of the previous years.

That led to more than EUR 1 billion in revenue.

And actually, the strong euro obviously did not make that very easy.

Looking at it more in detail.

This meant an overincrease of 37% year-on-year, leading to EUR 1,035,000,000 in revenue.

Our most major markets in Developed Europe grew 22% year-on-year; American markets, 37%; and our Rest of the World markets catch up massively in size, growing 84% year-on-year.

Looking at the percentages overall, we got again more global and more balanced in our geographical distribution of revenues.

Focusing back in, we have seen a lot of volatility between the quarters in terms of growth, but also in terms of the structure of our marketplace.

While the unit picture was pretty stable, a partly expected drop of commercialization in the second half of the year was followed by a rebalancing of the marketplace.

While, of course, hurting short-term commercialization and consequently growth, the significant testing activity of our advertisers also helped us to gather valuable insights and calibrate our relevance assessment.

Setting this apart, we see that the underlying growth trajectory of transactions through our marketplace remains intact.

So now, we'll hand over to Axel to go even more into detail.

So comparing to a very strong fourth quarter of 2016, where we grew 70% year-on-year, we grew in the fourth quarter 2017 our revenues by 7% from EUR 169.2 million to EUR 181.5 million.

For the total year, the growth was 37%, growing from EUR 754.2 million to EUR 1,035.4 million.

Adjusted EBITDA declined in the fourth quarter from EUR 11.9 million to a loss of EUR 8.7 million whereas for the full year, it dropped from EUR 28.2 million to EUR 6.7 million.

Looking at the net income, there was a drop from EUR 100,000 profit in the fourth quarter 2016 to a EUR 9.6 million loss in 2017 whereas for the full year, the loss was reduced from EUR 51.4 million in '16 to EUR 13 million in '17.

If we look at the return on advertisement spend, our key profitability metric, the full year 2017 came out at 115% comparing to 120% in the previous year.

And the fourth quarter, slightly above average, is 118% comparing to 134% in the previous year.

Our adjusted EBITDA margin developed from a 7% positive margin in the fourth quarter 2016 to a minus 4.8% in 2017.

And for the full year, it came down from a 3.7% in '16 to 0.6% in '17.

If we now look at the KPIs on a global level, qualified referrals grew 14% in the fourth quarter, again compared to a very strong Q4 2016, from 122 million to 139 million.

And the full year saw a 36% growth from 535 million to 727 million.

RPQR for the overall business was slightly down by 7% in the fourth quarter from EUR 1.36 to EUR 1.27, and for the full year, pretty much flat, going slightly up from EUR 1.39 to EUR 1.40.

The return on advertisement spend, as I mentioned on the slide before, 134% in the Q4 2016 to 118% in 2017, and for the full year, 120% going down to 115%.

So if we now look at the Q4 performance, there are basically 4 drivers that impact all the 3 regions.

And that's why I would like to comment on them just once for the overall business, not to be too repetitive.

So when you look at the qualified referrals, obviously it is driven by our marketing activities.

But the additional effect that needs to be considered is both the rollout of the attribution model and the continuous optimization of our platform both leading to a higher quality of the traffic, and as a consequence, slowing down the qualified referral growth.

The commercialization dropped compared to Q4 2016 across all the regions.

And we saw significant testing activity, in particular in the -- I mean, for Q4 but also overall in the second half of the year.

That obviously had an impact on the revenue that we can derive per qualified referral.

In Americas and in Rest of the World, there were currency effects.

As Rolf mentioned, a very strong euro had a negative effect on those 2 regions.

And those 2 negative effects were partially compensated through the rollout of the attribution model and the platform optimizations that lead to a higher booking conversion, and as a consequence, are positive drivers for the RPQR.

Looking at the return on advertisement spend, the biggest driver across all the regions was the drop in commercialization compared to Q4 2016 and the testing activities on our platform by our advertisers.

So if we now look at Developed Europe, in Q4, the qualified referrals dropped slightly from 51 million to 49.7 million.

And in the full year, they were up 16% from 255.4 million to 295.5 million.

RPQR was flat in the fourth quarter, EUR 1.42 to EUR 1.41, and for the full year, was up by approximately 5%, EUR 1.37 to EUR 1.44.

On the return on advertisement spend, the Q4 showed a decline from 158% down to 136% whereas the full year showed a decline from 136% and 131%.

So in addition to the trends that I mentioned before, there are basically 2 things worth mentioning.

One is that the testing activities on our marketplace were more pronounced in Developed Europe than in the other regions, so had a greater impact on the business.

And in Europe, there tend to be more pre-commitments in the marketing expenses.

And as a consequence, we could not adjust our spending to the level that we would have liked.

In Americas, the qualified referrals in the Q4 grew 13% from 36.7 million to 41.6 million and in the full year by 36%, 149.1 million to 203.4 million.

Revenue per qualified referrals dropped in the fourth quarter from EUR 1.72 to EUR 1.58, and for the full year, were flat coming from EUR 1.92 to EUR 1.93.

Looking at the RPQR in U.S. dollars.

The development was slightly different, which is explained obviously by the significant change in exchange rates.

So if you would look at the RPQR in U.S. dollar terms, it would have gone from $1.86 to $1.86, so basically stayed flat for the Q4, and for the full year, from $1.13 -- $2.13 to $2.18, so slightly going up by 2 percentage points.

Return on advertisement spend in Q4 went up from 136% to 123%, and for the full year, from 118% to 116%.

In Rest of the World, the qualified referrals grew by 40% in the fourth quarter from 34.3% to 48% (sic) [34.3 million to 48 million].

And for the full year by 75%, 130.8 million to 228.3 million.

Revenues per qualified referral in euro terms went from EUR 0.90 to EUR 0.86, so a decline of 4% in Q4 and from EUR 0.85 to EUR 0.89 for the full year, which is a 5% increase.

Looking again at the revenue per qualified referral in U.S. dollars.

The development shows a slightly different trend.

So there, it would go from $0.97 to $1.02, so an increase of 5% in the fourth quarter, and for the full year, $0.94 to $1.01, which is an increase of 7 percentage points.

Return on advertisement spend in the fourth quarter went down from 97% to 93%, and in the full year, increased from 90% to 92%.

So if we now look at the outlook for 2018, I think it is important to say that Q4, as Rolf mentioned before, has obviously seen a lot of volatility.

And so it is -- with this volatility, it is obviously not that easy to have a very clear view.

What we expect, nevertheless, is for the first half 2018 a decline in revenues compared to the first half of 2017; and for the second half of 2018, an increase in revenues; on return on advertisement spend, a decline in the first half and an increase in the second half compared to 2017; the operational expenditure to go up compared to 2017, both in the first half and in the second half; and the revenue per qualified referral to drop in the first half 2018 compared to '17 and to be flat by region in the second half of 2018.

As a consequence, the adjusted EBITDA in the first half 2018 is expected to go down whereas we expect it to increase in the second half of 2018.

Summing all of this up.

We expect the total revenue to grow for the full year between 5% and 10% and the adjusted EBITDA margin to be slightly negative for the full year.

So I'll hand back to Rolf for the closing remarks.

No.

I think we have the questions, first.

(Operator Instructions) And we will take our first question from Douglas Anmuth in JPMorgan.

I was hoping you guys could talk about some of your key initiatives to move beyond Priceline and Expedia.

And in particular, can you talk about your traction with other regional OTAs and hotel direct and if there's anything kind of incremental that you're doing in that area to increase those efforts?

Yes.

So I think you can see on this slide that I've shown for Q4 that I've shown that there is already quite a movement in the numbers for Q4.

And so we have seen that the small advertising brings more share.

And that also means more share than all other advertisers, so also more than the Expedia brand.

So we see there already a quite healthy trend.

And that is also due to the initiatives that we take and we currently like -- do that continuously, so continuously onboarding more advertisers to Automated Bidding, which we think is a very strong argument, a very good tool for our advertisers to be competitive.

We get in more and more hotels direct.

We onboard them to have their rate on trivago.

We see them more and more competitive.

And we also have more and more positive cases where they can get a significant volume of their overall -- of the overall share on trivago, they can get directly to their website.

So this is part of the story.

I think the other part of the story is going more aggressively into alternative accommodation to widen the offering and to create more competition in the long tail.

And our next question is from Naved Khan in SunTrust.

Can you comment on the advertiser behavior in the platform in Q4, i.e., was there -- did you see more volatility at the beginning of the quarter than at the end?

Did things seem to stabilize towards the end of Q4 and into January?

Can you just give us some color?

Yes.

So overall there, I would say that there has been significant testing activity and changes in the marketplace throughout the quarter.

But I think it is fair to say what you're mentioning that the structure, the advertiser structure stabilized more at the end of the quarter.

But there were various tests by various advertisers throughout the quarter.

Okay.

And then in terms of the changes you are making on the marketing attribution, I guess there is a negative impact on the qualified referrals.

But has that led to improved ROI for your advertiser?

Is it positively affecting the RPQR metric?

Or are you still not seeing that?

So yes, from our perspective, when you look at the stable RPQR, for example, in Developed Europe, that is exactly the reason.

I mean, the commercialization that we are seeing has dropped.

So we derive less revenue from the same booking amount that we transfer to our advertisers.

And that has been compensated partially or fully, depending on the region, by an improved booking conversion.

So that's exactly what you are mentioning.

So we see the positive effect of that in the RPQR in Q4.

And for those advertisers where we can see it, we see also improved commercialization on their site, so we see improved ROI on their site.

And yes, and we have good estimates for the others.

So it looks like there is an improved commercialization on their side.

Okay.

That's very helpful.

And so do you still expect to complete the rollout by end of the current quarter?

So on -- it varies by channel.

So on our DEA channel, we fully rolled it out in the third quarter 2017.

In our SEM channel, we are in the middle of it.

And we expect the large part of the rollout to happen in the first quarter with some remaining rollout potentially in the second.

Our next question is from Lloyd Walmsley in Deutsche Bank.

Two, if I can.

First, you showed the chart showing Priceline losing share in the marketplace in 4Q versus kind of flattish early in 4Q.

Curious, can you give us a sense for whether that share loss was a function of them reducing bids or more just aggressive bidding from other bidders?

And do you have any sense, I guess, related to that to what is going on in markets like Germany and Italy, where Booking.com for much of December was essentially off the site but came back in Germany?

And then I guess, the second question, your guidance for '18 does obviously imply a nice pickup in growth in the second half.

So even if you're down, call it, low single digits in the first half, it implies about 20% growth in the second half.

So wondering, what gives you the confidence in that outlook, given the elevated testing and general lack of visibility?

And maybe you can give us a sense for what kind of assumptions you embed in, in terms of auction pricing, growth in ad spend and impact from changes to your marketing attribution models that you embed in that second half acceleration?

Yes.

So maybe let me just comment on the first part.

So what we have seen is a multitude of different things that happened on the marketplace.

And we're seeing testing in every direction from several advertisers also as a reaction of our relevance assessment.

And I think it's totally fair for advertisers to do that, to play around with that and to see where they can make a gain.

And then the marketplace, of course, has seen a little bit of volatility.

So in total, I think when we have spoken about commercialization before then, it's clear that at least on a relative basis, partners like Booking.com seem to have reduced their bids.

And this is affecting us and affecting our commercialization.

And that is also, in that moment, you have the share taken over by other market participants.

And in this situation, it was even taken over more by the small advertisers that it was taken over by the Expedia group.

So I think second part is you?

Yes.

So the second question was why are we comfortable that the second half will be positive whereas the first half will show a decline.

I guess, you really need to look at the last couple of years and think about it from a normalized perspective.

So we had some quarters, and there I would point out Q4 '16, Q1 and Q2 '17, where we had a very, very strong commercialization that we partially reinvested into the very responsive marketing channels to fuel growth even further.

And so when we look now at the plan, we can obviously normalize that in our numbers and say, "Okay, without that reinvestment with a more normal commercialization, how would that year would have looked like?" And looking at that, you see actually a much more steady development of the business.

And also in Q4 and in the first half of 2018, and that gives us comfort that this normalized trend will actually continue.

And then to the outside, we'll show the numbers that you're indicating for the second half of 2018.

And I think what should also give us confidence is when you look at what happened in Q4, where the marketplace kicked in, right?

So I think there was a level, and you can see that also with the testing activity, of some of the large advertisers.

There was a level when they just lost too much share compared to their profitability, so where they did just not want to like even reduce a bit more.

So there was this kind of safety net.

And I think that safety net is there.

And we tested it out now.

Like in the fourth quarter, we tested it out to the bottom.

But that doesn't mean that it stays there.

And even if we would assume it's staying there, we would still have enough reason to believe in the growth in the second half of the year.

Our next question is from Heath Terry in Goldman Sachs.

Your comments around ROAS in 2018 continuing to decline, can you help disaggregate that for us?

How much of that is a function of advertising cost going up on a like-for-like basis versus the declines in revenue per qualified referral or any other component within that calculation that you feel like is important to call out?

And then you started the conference call with comments around diversification, both within alternative accommodations and direct bookings.

Can you give us a sense of what you're seeing from both of those channels, both alternative accommodations but particularly direct bookings in terms of their interest in being on the platform, how their strategy, particularly around pricing both their rooms and their bids might differ from your OTA partners?

Sure.

So on the negative outlook for the RPQRs in the first half, that obviously, as you point out, has a direct impact on the return on advertisement spend.

And from our perspective and what we see today, that is the main driver of the negative outlook that we are giving on the return on advertisement spend in the first half.

So it is not driven by spend, but it's driven by the RPQR development.

Okay.

And regarding the second question, I think looking at the alternative accommodation section, we ramped up the number of hotels over the last 2 quarters continuously.

And I think it's always a function of like personalization on the platform.

And I think we did a good step forward there.

But it's still a lot of room to grow for us in terms of the personalization, and then accordingly also in terms of showing alternative accommodation.

Because I think that is always very closely related.

Because alternative accommodations are rather for like a subgroup of users, they are not like for all of our users.

And that's why it's so important if you want to commercialize well to really find the right users to show that inventory, too.

So that will be always a function and will remain a function and will be like also growing over time.

And we see it growing over the next 12 months.

On the other side, looking at hotels direct, yes, we're also ramping that massively -- ramping it up massively.

I think the total number is still small, so growing quite fast of individual hotels.

But we see very positive cases.

We see very positive cases where individual hotels are able to sometimes deliver better pricing.

I think they took this now as part of their strategy to undercut the prices and also to bid more aggressively because their technology works or their technology especially works in combination with the technology that we offer to them with our Express Booking technology.

So we see a very positive cadence.

I think what you have to keep in mind, even if the absolute number is small, we see that we can do this like rather on high-volume hotels.

And we will add more and more of them.

So we heavily -- we invest into this channel.

We also invest into this channel in the upcoming year.

We invest into building up relationships to individual hotels.

It's always a function of our relationships to hotels direct.

We have now more than 400,000 hotels that are using our Hotel Manager tool.

We have more than 40,000 hotels that we have a direct commercial relationship with.

And then we use this commercial relationship to basically onboard more and more of these individual hotels and to do -- to use trivago as a marketing channel.

And so just to clarify, you're actually seeing smaller independent hotels pricing their rooms below the prices that the OTAs are showing as a way to drive volume to their sites through trivago.

And so it's happening with the independent hotels, not just with the hotel chains that have launched loyalty direct and direct booking programs.

We see that this is like in their repertoire of strategies, I'd say, which was not the case so much before.

So I think before, you have seen that way less.

I think now it becomes part of the whole game.

If it will be dominating, not sure.

That is definitely part of the game now.

Your next question is from Brian Nowak in Morgan Stanley.

I have two, the first one on advertising.

The third quarter last year, you made some comments and talked about diminishing marginal returns on advertising in the Americas and in Europe.

I guess, could you talk through where you are now at that point and kind of the branded spend in the Americas and Europe?

Are you at a point where you need to start to spend more on performance in paid search?

Or have you fixed your branded advertising diminishing return challenges?

And the second one, going back to Lloyd's question a little bit on the 2018 guidance, could you just help us and investors kind of better understand a little bit: a, the average visibility you have in the business at this point, given the volatility; and maybe walk through some of the forecasting process that went into the second half and what you're assuming for your 2 largest advertisers in the second half of the year?

So let me just comment on what we said in Q3.

I think it was also quite -- so when we spoke about the diminishing returns, it was also quite early in this whole process of like how the marketplace reacted, what we spend on advertising, how our overall commercialization developed.

And I think when we look -- looking back today, I think the effects of diminishing returns, we were probably in the beginning overestimating that.

That's one point.

I think that always, in every market, there is diminishing returns in the moment where you keep your advertising stable.

There is a wear-out effect of your commercials and so on.

I think that is something that we will definitely address in the next year, where we want to have like more commercials, invest more in creative development and so on and so on.

So I think it's too early to really say in general the diminishing returns.

I think that it's very clear that if you do more of the same thing, probably the outcome will in the future be less and less.

But you don't have to do more of the same thing.

So you can do very different things.

And you can do also like address people very differently.

I think we can do a lot of things in personalization and personalizing, individualizing our advertising.

So no, I don't think that there is like a proof of diminishing returns.

But I think we have to, of course, like we have done that basically for the last 10 years.

So we always have to be smarter and smarter and smarter every year basically in addressing our target group.

So coming to your second question, what is our average visibility on the business looking at Q4?

I think that's a fair question.

I mean, the way -- I mean, obviously, the visibility of dramatic changes happening tomorrow is something you can't have visibility on that because it's an independent decision by a separate company.

I guess -- so it is not so much visibility, but it is more, "Okay, what did we actually factor into our plan?

And how well can we react to extreme scenarios?" And I think there, definitely, the fourth quarter was, despite the negative effect that it had on our financials, was for us positive because there were a couple of also extreme tests that we've seen from advertisers on us that didn't only give data and insight to them but also, and more importantly to us, obviously to us.

And so those data points we processed, digested and now have available to react to these kind of situations much better than we could have before the fourth quarter.

So that, I think, is a positive to consider there.

How did we come up with a plan for 2018?

There, we assumed no significant change.

That's obviously the assumption that you need to take, so neither positive nor negative.

So if there is a significant change in the competitive dynamics, and then it would obviously have a negative or positive impact on the numbers.

But you need to have one assumption.

And our assumption is basically stable throughout the year or at the level that we are seeing throughout the year.

And when you're looking at the fourth quarter, there was rather more stability towards the end of the quarter than in the beginning of the quarter.

So I think it's a fair assumption to make that on this level now that we are from a commercialization perspective, there is also like the marketplace now kicking in and the competition is working.

And so I think that gives us quite some confidence that this is basically a very base level assumption of -- it should be stable.

Next question is from Mark May in Citi.

Realizing that it's a difficult comp, as you've already guided for, but was hoping you'd maybe -- that you'd share what the year-on-year decline in revenue was last month in January, kind of how you started the quarter.

And then in terms of the volatility in the marketplace that you've talked about, it seems like one of the main variables to stabilizing this would be QR pricing, so -- which probably still needs to come down.

Just wondering, how much further do you think RPQRs need to, on average, kind of be adjusted further down before advertisers as a whole start to view them more in line with kind of their expectations?

So on the first question, we won't comment on the current development in January that we are so far seeing.

I think that what -- if there is something that you could have seen in the fourth quarter and learn from it, that 1 month can differ quite a bit from the other.

And so I don't think that, that would be good to use 1 month as a proxy for the full month.

So that's the first half of your question.

On the second half, I mean, let me just make sure to clarify that I understand your question correctly.

You are saying that the RPQR has to come down to recalibrate?

I'm not sure I fully understand what you mean.

It seems like one of the things that's going on here is the market is readjusting their pricing in order to reach ROASs that are in line with what they want to achieve, so -- and maybe I had that wrong.

But if that's right, how much further down do you think that we have here in terms of your average pricing before we start to see some stability?

So I think that the -- it is -- okay.

Now I understand it.

Thanks for clarifying.

I think it is a bit the complication of comparing always to the previous year's quarter.

So when the fourth quarter is down versus the previous year and when we are saying the first half of '18 is down versus the first half of '17, that doesn't mean that the commercialization in the first half of 2018 will be down compared to the fourth quarter and 2017.

So what I said before is that we actually assume that the current level will stay, which means that adjusted for seasonality, which is obviously there with different booking conversions, et cetera, that the commercialization, so our share of the overall value, will relatively speaking be stable.

So we don't think that there will be a readjustment.

But if you -- I mean, going forward.

But if you compare obviously to last year, we saw a very, very positive development in Q1 and Q2 '17 and also in Q4 '16.

So if you compare against that very positive time frame, then obviously it is a negative change year-on-year.

Your next question from Robert Coolbrith in Wells Fargo Securities.

This is Rob on the call for Peter.

Two questions, if I may.

Just wondering, it looks like the net effect by region of commercialization versus conversion roughly flat excluding FX, wondering if you could give us a sense of the scale of the inputs there in terms of commercialization versus conversion.

Also on the attribution model, is that more of a step-change improvement?

Or is [this certainly] can continue to drive RPQR higher over time, adjusting for the effect of commercialization?

And then second, as you've noted, your efforts around personalization, wondering if you can talk a bit about your preparation for GDPR coming up in May.

Okay.

So on the scale of the commercialization drop and the increase in booking conversion -- just before I get to that and just to remind everybody, I mean, we've got a certain coverage of full visibility on what is going on throughout the funnel, which varies obviously by quarter but as over the whole year being at approximately 50%.

So obviously, there is some -- we need to estimate the other 50%.

And so there, that's why we don't disclose that level of granularity.

Having said that, we think that it is -- or from what we estimate, it is a significant effect in both directions, so high single digits and potentially more.

And that's why we are calling it out separately.

And that's why we are very comfortable that, despite the fact that we estimate approximately half of the overall volume, that this effect holds for sure.

So looking at the effect, I think there are again like different effects that we will see.

But of course, like when we try to target our visitors better, to target our target group better, we will probably rather invest more into less people, right, so -- and that is, of course, coming also with like the people who are coming to trivago, the people who are maybe then also going towards our advertisers, who produce the lead, that they will also want to have higher value.

So CPC price is going up, they have higher value and also the revenue per qualified referral going up in the long run when we look at our marketing spend, so -- but the question is always like how you roll that out because you have all these different channels.

There's always also balancing out.

So if you're going -- for example, if you're going more into content marketing, then you have naturally in content marketing maybe visitors which are more early in the funnel.

So you still want to make your learnings, you still want to go in, you want to scale up the volume as long as it's profitable, and then to make enough learnings that you can then again individualize and personalize and pick the right ones again.

So I think you have all these different kind of like different effects there.

And sometimes they mix up.

And I think that is -- might be true for also for the next year.

But in general, I think the tendency, that is clear, right?

On your second question...

Anything you could have on GDPR?

Yes.

On your second question, it is obviously additional work.

I think that's clear with additional requirements to documentation and affecting other processes.

And we are obviously preparing for it.

And we will have to spend to comply, and we'll incur additional costs.

Having said that, overall, we don't see a significant impact on our business as we operate.

But it is definitely something that is important and that we are taking serious.

Shyam Patil from Susquehanna.

I have two.

First one on Google Hotel Finder, can you just talk about what you're seeing there, how that's impacting the meta space and trivago specifically?

And then second, just following up on a previous question and maybe a little bit more specifically, can you just talk about how you're thinking about Priceline's spend in your guidance in 2018 and just kind of how you're -- what you're basing your assumptions on?

Yes.

So I mean, in general, what we're trying to do is we're trying to diversify our channels more and more.

And we -- I think that is also the reason why we went very early into brand marketing.

That's the reason why we are heavily learning in video, in Facebook and so on and so on.

So we want to diversify.

And I think Hotel Finder is one of the channels that you can diversify in.

So we see it as an interesting opportunity also.

But I think it's -- for us at least, it's still quite early, so I think too early to comment about it.

But it might be an interesting opportunity.

On your second question on Priceline's spend, I mean, we don't comment on individual advertisers.

But overall, we assume a stable level of commercialization, which basically means that we will capture the stable share of the overall value.

Whether that will be through an increasing, flat or decreasing share of Priceline or not is something that we won't comment on.

But the basic assumptions of the tension in the marketplace will remain stable.

(Operator Instructions) And our next question is from Kevin Kopelman in Cowen and Company.

So just given it sounds like there's a -- you have a lot of volatility in Q4 but then the increased stability in market dynamics, it sounds like, in December and January.

Can you talk about what you've seen in December and January in terms of advertisers' share on the platform, how that compared to what you've shown us, which is the entire Q4?

And then I have another question.

Okay.

So we will not comment on January.

But in the fourth quarter, when we did our last earnings release, we talked about the quarter-to-date.

So there, we at least gave some data point on how the dynamic changed in October.

And looking at the full quarter, you can see that the remaining month returned to levels that were more in line with the levels that we've seen before.

So yes, I think that's what we can say on the development within the fourth quarter.

And as I said before, going forward, we expect stable tension in the marketplace and a stable commercialization level.

Okay.

Great.

And then a separate question, in the fourth quarter 2017, you talked about some negative impact from TV pre-commitments.

Can you quantify the EBITDA impact approximate in the fourth quarter from that development?

And then how much of an issue, if at all, are pre-commitments for Q1 or 2018?

So if you look at the drop in Developed Europe compared to the other regions, the majority -- vast majority of that delta in drop or a stronger drop in Europe than the other regions is attributable to that specific effect.

Most of the commitments are annual commitments or the vast majority of commitments are annual commitments if they exist.

And so for 2018, that is not something that we see as a challenge going forward.

Okay.

And in Europe, is that the ROAS drop that we should focus on?

Correct.

Yes.

Okay.

Got it.

And then one last one on 2018 revenue guidance, can you give us a sense of the -- what kind of impact currency has on that 5% to 10% growth rate?

So we assume -- I'll compare it to 2017.

I mean, we assume for 2018 in our guidance a constant currency as of end of January.

So if that were to change, that would basically have a positive or a negative impact on the overall euro guidance.

Okay.

And presumably already in that 5% to 10%, you've got some negative year-over-year currency impacts then?

Absolutely.

Yes.

The next question from James Lee in Mizuho.

You guys talk about needing to build a better recommendation engine.

Maybe can you talk about the roadmap, how you want to get there in terms of product and maybe timing of the launch?

Should we think about launching by region, by user demo, hotel type?

And just curious, what is the friction in terms of recommendation on your engine, recommendation engines?

Is it fair to say that user come to trivago, they already knew which hotels to choose?

Or do people come and they don't know what to choose and the recommendation engine right now is not optimized?

So I think when you look at it is -- what you have to do if you want to be able to move your business from like a business where you list hotels to a business where you match user profiles and hotel profiles, I think that is quite a big step.

And I think that we said last year, end of last year, we said, "Okay.

This is something that we have to do." Because in the future, it will not be able -- you will not be able sell hotels if you're not really matching user profiles and hotel profiles.

So that's our bet for the future.

So we have to act according to that.

And we really had to revamp our back-end infrastructure.

You can imagine that this is something that where you have to invest a lot of time to do that to change the data flow.

At the same time, we moved a lot of what we do into cloud to also be able to do the calculations there and so on and so on.

So I think that was something that we did during the last year.

The way how we roll things out is usually not that we do a big bang or even not do a big bang in one market.

But the way we roll things out is usually in a way that you will probably not even realize or recognize that there's something happening.

And we will -- and that is a continuous process.

So there will be probably a time when you will see it more obviously.

But it's always a process.

And sorry, could you repeat maybe the second part of your question because I'm not sure if I have understood that right?

Yes.

I guess, the reason you're seeing friction within your recommendation engine, is it because when people come to trivago, people already knew what hotels they want to choose?

I already come in, I choose Sheridan.

I just want to compare prices.

Or do most people come in, they say, "I type in Amsterdam, for example.

I don't know which hotels I want to go," and therefore, you feel your recommendation is not optimized based on matching the profile of the users and hotel type?

I'm just wondering which is which.

Well, I think a lot of users come in -- I mean, we see that on trivago, but I think it's in the industry.

A lot of users come in and they're not like know already which hotel they want to go to.

And they don't use trivago like only to compare prices.

But we get a lot of searches coming in top of the funnel.

And I think maybe that's also a little bit different from trivago maybe to other players in the marketplace because we get a lot of brand traffic.

So a lot of users come into trivago are not yet sold on one specific hotel, so -- and that's why we think that the most -- the vast majority of the value creation really happens actually there.

The vast majority happens in like showing them the right hotel, not just showing them a hotel, which is like commercializing very well or, I don't know, which has, I don't know, paid listing like on other platforms, but really showing the hotels that the user wants.

And for that, I mean, how do we find out what a user wants?

I think there's a lot of effects or a lot of data points that we get from a user before interaction.

And there is definitely an intent from us to get users even more interaction during the process, to learn more about the user and to be better able to react.

And in all of these areas, we're currently investing.

If I could ask a follow-up question here, I noticed that on our tracking, your SEO channel came up quite a bit.

You used to have maybe very small SEO channel.

Now your SEO channel is much larger.

And just curious how strategic that channel for you specifically.

Maybe you can confirm that being the case, number one.

And number two, how should we think about SEO channel conversion versus your SEM channel, for example?

I mean, we don't usually comment on the channel development.

But I would not actually like confirm that development.

So I don't see that there is like a drastic change or significant change in the different channels or in the percentages of the different channels.

So no, I'm -- let me generally speak about SEO.

So I'm not a strong believer that the pure SEO model will be like something where you can say like for the next 5 years or so, there will be -- we'll assume significant traffic increases from that.

So I think that would go definitely against our philosophy like how content and commercialization will move into each other.

And I think all trends that you see also, they show that.

It doesn't mean that we don't think it's a good channel.

It doesn't mean that we're not investing or that we're not optimizing for it.

But I think it's fair to assume that, I think, for the next 5 years or so, there will not be like -- the channel will not, in general at least, not only for trivago but in general, gain importance.

As there are no more questions at this time, I would now like to hand the call back over to the speaker today, Matthias Tillmann.

Yes.

Thank you very much.

And Rolf, any final remarks?

Yes.

Many thanks again for joining the call.

I think the results were in line with the expectation that we said on previous calls.

We know, and we were very open about it, that we will have to go through some more difficult quarters from a financial perspective.

Still, what counts for us the most is how the team is creating value and how it is motivated and how it's delivering on the larger picture.

And there, we really see very positive trends.

So yes, again thanks for participation.

Looking forward to talk to you soon.

Thank you.

And this will conclude today's conference call.

Thank you for your participation, ladies and gentlemen.

You may now disconnect.