Criteo S.A. (CRTO) Q3 2016 Earnings Report, Transcript and Summary

Good day and welcome to the Criteo Third Quarter 2016 Earnings Conference Call. [Operator Instructions] Please note this event is being recorded. I would now like to turn the conference over to Edouard Lassalle, Head of Investor Relations. Please go ahead.

Thank you. Good morning, everyone, and welcome to Criteo's Q3 2016 earnings call. Eric Eichmann, CEO, and Benoit Fouilland, CFO, are with me today. During this call, management will make forward-looking statements. These may include projected financial results or operating metrics, business strategies, anticipated future products and services, anticipated investment and expansion plans, anticipated market demand or opportunities, and other forward-looking statements. These statements are subject to various risks, uncertainties and assumptions. Actual results and the timing of certain events may differ materially from the results or timing predicted or implied by such forward-looking statements. We do not undertake any obligation to update any forward-looking statements, contained herein, except as required by law. In addition, reported results should not be considered as an indication of future performance. We will discuss non-GAAP measures of our performance. Definitions of these metrics and the reconciliations to the most directly comparable GAAP financial measures were provided in our earnings release issued earlier today. Last, unless otherwise stated, all growth comparisons made in the course of this call are against the same period in the prior year. With that, I am now turning the call over to our Chief Executive Officer, Eric Eichmann.

Thank you Edouard. Good morning, everyone. I am pleased to report we achieved another strong quarter of profitable growth and continued execution on our operating plans. Before going through the quarterly earnings, I will say a few words about our recent investments and their strategic importance to Criteo. The acquisition of HookLogic, which we anticipate will close shortly, and the launch of Criteo Predictive Search are two key milestones in the execution of our long-term strategy. With these two strategic initiatives, our performance marketing platform becomes an even more critical part of our advertisers' marketing program and success. HookLogic allows brand manufacturers to advertise their products on a performance basis on the world's largest e-commerce sites. With HookLogic, we will be able to capture a new source of advertising dollars from brand manufacturers while driving valuable site monetization for retailers. These new revenue streams for large retailers makes us more strategic to their business. And the new Criteo Predictive Search product provides retailers with a sophisticated end-to-end automated solution that delivers significant performance improvement for their Google Shopping campaigns. Predictive Search brings Criteo's unique machine learning approach to search to drive systematic uplift in sales of up to 49%. Predictive Search opens up a large new market for Criteo, almost as big as our core business today. Since retailers use search as their largest marketing platform, Predictive Search makes Criteo solution even more relevant for them. In addition to our core offering, the HookLogic search and cross device products provide retailers with a full set of solutions. This portfolio of products leverages the power of our network and large investments in technologies to help retailers compete on equal terms with the largest e-commerce platforms. We bring large-scale through our pooled assets across retailers and state-of-the-art technology that retailers could not develop on their own. While we are not seeing significant changes in the competitive landscape, the addition of HookLogic and Search will further differentiate Criteo in the marketplace. With these latest strategic initiatives, our journey to deliver relevant, accountable, and seamless marketing to our customers continues, strengthening our ecosystem and providing large opportunities for future growth. Turning to Q3 results. For the 12th consecutive quarter, we exceeded revenue ex-TAC and adjusted EBITDA guidance. At current currency, we grew revenue ex-TAC 30% to $177 million and adjusted EBITDA 51% to $54 million. We continue to make solid progress towards our long-term operating targets, demonstrating the leverage in our model. We performed well across the entire business. We delivered on our innovation roadmap, we further expanded our publisher relationships, and we added a record number of new clients globally. Technology innovation drives more client sales every quarter. The same clients that were live in Q3 2015 and were still live in Q3 2016 generated 15% more revenue ex-TAC at constant currency. The 75% of our business coming from uncapped budgets continues to drive the same client growth. I would like to highlight four technology innovations. One, our Adaptive Revenue Optimization feature in the Criteo Engine is now live and we've started deploying it across our client base. This new feature allows advertisers to bid on a cost of sales target as opposed to a cost per click, or CPC. This makes the return on investment calculation and optimization much simpler for advertisers and will be a key enabler in rolling out the additional optimization models of the engine, such as our value optimizer which maximizes gross margins for advertisers. Two. Our Universal Match Technology continues to see strong client adoption, with users matched through these technologies now, representing 52% of revenue ex-TAC. To help our advertisers with improving marketing measurement and consumer understanding, we have started sharing our granular cross-device data back to their analytics and attribution systems. We expect that this service will allow advertisers to better measure the impact of their marketing and ultimately, attribute more value to the cross-device convergence we drive. With this robust pooled assets, we believe we are in a unique position to help our clients gain better cross-device insights, a necessity for competing and winning in e-commerce. Three, we continue to invest in our publisher-facing products, taking advantage of recent shifts in the markets such as header bidding. As we described at our Investor Day in September, we are making modifications to our bidding strategy to maximize yield for advertisers, such as using our machine-learning technology to predict the prices of opponent bids. And we launched a direct header-bidder technology that allows us to be in the auction without going through the traditional RTB platforms, reducing cost and capturing more data on the auction bids. While this is still in progress, our smart approach allows us to embrace the changes in the publisher landscape and take advantage of the opportunities that header bidding opens up for sophisticated technology players like Criteo. Four, mobile in-app commerce continues to enjoy strong momentum, with app advertising revenue ex-TAC growing over 300%, making it a meaningful part of our mobile business. Moving to publisher relationships, we grew our publisher base to over 17,500 direct relationships. Starting with Native, we launched a partnership with Outbrain, a native advertising platform specializing in content discovery and sponsored links and expected to grow in Q4. Additionally, close to 400 publishers are already live with the direct native offering that we launched in Q2. Overall, revenue ex-TAC from Native, excluding Facebook, grew over 80% quarter over quarter. And our flexible integration and dynamic creative capabilities remain strong advantages in Native. As mentioned, we launched an early version of our own header bidder in Q3, which connects directly to publishers ad servers. The broader deployment of our header bidder is scheduled for Q1 2017. Around 7,000 advertisers are now live on dynamic ads on Facebook and the Facebook Audience Network, both on mobile and desktop, and a similar number is live on Instagram. We launched our app-to-app solution for dynamic ads. We are now able to track sales on the same device and are working in partnership with the Facebook team to make this feature available to our clients. And we further broadened our mobile reach by expanding our partnerships with Interactive globally, and by adding MoPub and InMobi as new partners in China. Moving to clients additions. With over 1,000 net new clients, we set a new record and ended Q3 with close to 13,000 clients, while maintaining client retention at 90% for 21 consecutive quarters. We signed up large and mid-market clients across all regions. Mid-market continues to grow fast worldwide, bringing in over 80% of new clients. Mid-market now represents 26% of revenue ex-TAC and remains a significant opportunity for us. To launch new campaigns even faster, we continue to roll out a set of automation tools. Following the Q2 rollout of the product-feed module, we just rolled out the creative module in Q4 which automates the production and online validation of creative templates, in line with advertisers branding guidelines. The complete range of these self-service tools from registration and tagging to product feed, creative and payment will be available by early 2017. Turning to regional performance, we executed well across all geographies, in line with our plans. The Americas grew revenue ex-TAC 31% at constant currency and remain the largest contributor to our growth. Our granular cross-device data service is increasingly a strong differentiator in our discussions with large US prospectors. In EMEA, revenue ex-TAC growth accelerated to 27% at constant currency. Established markets continued to post solid double-digit growth, especially Germany, where client additions were particularly strong. We signed several new large clients, including Idealo, Germany's largest price comparison website. Travel clients further expanded their business with us across several European markets. Finally, APAC revenue ex-TAC grew 34% at constant currency. Momentum remains solid across key large markets. Japan, once again, delivered strong performance, especially with existing large strategic clients. Our partnership with Kakao is quickly ramping up and now represents close to 20% of Korea's revenue ex-TAC. Business in Taiwan was strong and we saw growth accelerate across the Pacific markets. Looking to Q4, we remained focused on a clear set of operating priorities. Number one, execute flawlessly during the upcoming holiday season, particularly in the U.S. Number two; drive successful adoption of Criteo Predictive Search in the U.S. before rolling it out in additional markets next year. And number three, start the integration of HookLogic once the acquisition is closed. In closing, I'm pleased with our strong Q3 performance. We continue to deliver high growth while significantly improving our profitability. We executed in line with our plans, made a strategic acquisition, and launched an exciting new product in search. The acquisition of HookLogic, our new search product and our cross-device graph will make us more strategic to advertisers, increase the overall stickiness of our solution, and provide large growth opportunities. I look forward to updating you on the progress of these initiatives in the coming quarters. With that, I'll now turn the call over to Benoit, our Chief Financial Officer.

Thank you, Eric. Good morning, everyone. Just like Eric, I'm pleased with our strong performance in Q3. Rapid growth on astounding profitability remained key attributes of our model. I will walk you through the quarterly performance and provide you with our guidance for Q4 on 2016. Revenue was $424 million, up 27%, or 25% at constant currency. Revenue ex-TAC, the key metric we used to monitor our business performance, grew 32%, or 30% at constant currency, to $177 million. This was driven by our largest quarterly addition of clients to date as well as a continued growth of our business with existing clients. Compared with guidance assumptions, change in ForEx had a positive impact of approximately €2.5 million -- $2.5 million, sorry, on reported revenue ex-TAC, mostly driven by the stronger Japanese yen and the euro. Compared with Q3 2015, ForEx represented a tailwind of approximately 200 basis points to reported growth in revenue ex-TAC, also largely driven by a stronger yen and partially offset by a weaker British pound. Now a quick word about our traffic acquisition cost, or TAC. At constant currency, our average CPM or cost per thousand impressions decreased 4% year-over-year in the quarter while our average click through rate improved 11% over the same period. Revenue ex-TAC margin was 41.7%, slightly above prior quarters, in line with our expectations. Our dynamic margin management approach allows us to optimize margin across regions. With a single objective of maximizing absolute revenue ex-TAC at both the local and the global levels. As a result of this approach, the margin differentiation between markets has increased, while the margin in EMEA trending higher than historical patterns this quarter. Shifting to expenses. Other cost of revenue, comprised of hosting and data costs, was approximately $22 million, mainly driven by increased hosting capacity across data centers. This was approximately $1 million below expectations, driven, in equal parts, by hosting and data costs. Operating expenses were $131 million. Non-GAAP operating expenses grew 23%, to $111 million. Headcount-related expenses related -- represented slightly below 75% of non-GAAP OpEx. We added 130 net new employees and closed the quarter with over 2,200 employees, a 27% increase compared with September 30 last year. OpEx came in approximately $5.5 million below our Q3 guidance assumptions, excluding Forex impact. The HookLogic acquisition expenses, included in our guidance, account for approximately 30% of these volumes. Since the transaction is now signed, acquisition-related costs have been excluded from non-GAAP OpEx, in line with our definition of adjusted EBITDA. About 30% of the volume relates to employee-related savings, primarily travel expenses and variable compensation, while the remaining 40% relates to conservative estimates on market conditions, including employer contributions on RSUs on other items. On a non-GAAP basis by function, or excluding amortization, equity awards compensation expense, pension service costs, acquisition-related costs, and deferred price consideration, R&D expenses grew 28% to $24 million, largely driven by the 32% growth in headcount to close to 480 employees. Sales and operation expenses grew 15% to $61 million, also largely driven by the 25% increase in headcount to over 1,350 employees. As the percent of revenue, Sales & Operation OpEx decreased by 160 basis points on a non-GAAP basis. Quota-carrying headcount grew 27% to 640, with 75% of the growth still coming from the mid-market. G&A expenses increased 40% to $26 million while headcount grew 26% to 380 employees. Excluding exceptional items on timing differences, primarily related to litigation expenses and employer contribution on stock-based compensation, G&A expenses grew only 28% to $23 million. We grew adjusted EBITDA 55%, or 51% at constant currency, to $54 million. Adjusted EBITDA margin for the quarter was 12.6% of revenue, or 30.3% of revenue ex-TAC. This represents an improvement of 230 basis points of revenue or 450 basis points of revenue ex-TAC, driven by our gross margin improvement on the continued operating leverage in Sales and Operations. In the first nine months of 2016, we increased profitability by 180 basis points of revenue, or 400 basis points of revenue ex-TAC. This highlights the strong leverage of our model and our ability to drive adjusted EBITDA expansion over the long term. Speaking of our operating model, we do not expect the addition of HookLogic to materially impact our ability to reach our adjusted EBITDA margin target of 40% of revenue ex-TAC in the long term. Financial income improved by $6 million. This increase was primarily driven by the much lower foreign exchange loss compared with last year, mainly as a result of the conversion of our intercompany position in Brazil from debt into equity in Q2 this year. Net income increased 154%, to $15 million, driven by the growth of our income from operations on the significant improvement in financial income over the period. The effective tax rate was 34% for the quarter, and 30% for the first nine months of 2016. We now expect our effective tax rate for fiscal year 2016 to be approximately 30%, slightly above our initial expectations. Adjusted EPS, on a diluted basis for the quarter, increased 173% to $0.48. Cash flow from operations grew 149% to $44 million as a result of our increased profitability and a positive change in working capital, in line with our expectations. This represents the conversion of adjusted EBITDA into cash flow from operations of 82% in the quarter. CapEx decreased 17% to $20 million, or approximately 5% of revenue. This compares to a peak quarter last year when we had spent a third of our annual CapEx Program on a new Hadoop cluster, additional servers and fit out for new facilities. Free cash flow increased by $30 million to $24 million and by 81% in the nine months of 2016. This translates into a 44% conversion of adjusted EBITDA into free cash flow for the quarter. Finally, total cash and cash equivalents were $407 million at the end of September, up $54 million from December 31, 2015. I will now discuss our guidance for Q4 and 2016. The following forward-looking statements reflect our expectation as of today, November 2, 2016. We expect the HookLogic transaction to close in the coming weeks. The contribution of HookLogic is not included in our Q4 and full-year guidance communicated today. We expect Q4 2016 revenue ex-TAC to be between $207 million and $210 million, excluding HookLogic. This would imply a growth of constant currency of 28% to 30% without HookLogic. We expect changes in ForEx to have a positive impact of approximately 160 basis points on our Q4 reported growth. And we expect Q4 2016 adjusted EBITDA to be between $72 million and $75 million, excluding HookLogic. ForEx assumptions underlying the Q4 2016 guidance are included in the earnings release we published earlier today. For FY16, we expect revenue ex-TAC to grow between 33% and 34% at constant currency, excluding HookLogic. We expect changes in ForEx to have a positive impact of 70 basis points on our reported growth for the full year. And we expect FY16 adjusted EBITDA margin, as a percentage of revenue to improve by 120 to 140 basis points, excluding HookLogic, up from our prior guidance of 60 to 100 basis points improvement, excluding HookLogic. In closing, I'm very pleased with our strong performance in Q3, combining a rapid growth with continued leverage and increasing profitability. I'm excited about the expected positive impact of HookLogic and the new search product on our future growth. And I am confident about the outlook for our business. With that, let me now turn the call back to the operator to take your questions.

We will now begin the question-and-answer session. [Operator Instructions] The first question comes from Mark Kelley of Citigroup. Please go ahead.

Just curious, your margins in EMEA continue to get better. Just wondering what's driving that? Is that 60/40 mix between auction and publisher direct inventory in that region or is that all click-through rate improvements and then second, any early indications of your header bidder in terms of how much inventory you're able to see that maybe you weren't able to see before?

Regarding the margin, yes, we had a strong margin during the quarter. That is primarily the result of our dynamic margin management approach, which allows us to optimize margin across regions with the objective for the second year optimize level to maximize absolute revenue ex-TAC both the local level and at the global levels. Now that approach allows us to take advantage of our large-scale and optimize our pricing power by market. So what we’ve seen is that the result of this margin management approach, we have differentiated the margin between markets based on the dynamics of each market, which has resulted to drive a slightly higher margin in the EMEA.

And Mark, on your question on the header bidding environment. I think from the beginning, we thought that any change in the ecosystem bid on the publisher, either on the advertiser side are changes that generally carry with them opportunities, in particular, if you have a technology or innovation that has been used to adapting to these environments and taking advantage of it. For header bidding, in particular, what you have is an opportunity to access more inventory, to be at the front of the line and one of the things that we've done and we've talked about this in the script is that we actually have now started to deploy a direct bidder where we are in the wrapper with header bidding and that allows us to -- and gain access to more inventory, but also as we use machine-learning technology to predict the bids, being in a better position to win more impressions. It's early in the deployment of that technology. We're still in the beta phase of it so I wouldn't be ready yet to talk about the impact that it could have -- the positive impact. But what we've said is that header bidding was not a big issue for us. It represented an opportunity and at any rate, we also thought because supply and demand haven't really changed and this is just -- header bidding is just another way of connecting the two. That over time, you would get to a new balance that would be similar to the past, except that, of course, players like us would be able to reach more inventory at better prices. So very positive for us and we'll probably see some more results of this in the future.

That’s helpful. Thank you.

The next question comes Douglas Anmuth of JPMorgan. Please go ahead.

A couple things. First, just a follow-up on the last one. It’s clear that you try to optimize the margin in the quarter and did that differently across geographies, but is there anything more specific that you can talk about on the 4% lower CPM that you mentioned year-over-year? And then secondly can you also just review on Search, the economics? How we should think about take rates there and the impact on financials going forward? Thanks.

Yes, on -- thank you, Doug. Great question. So on the CPM, I think it was important for us to mention a couple of things, like in the header-bidding environment, where I think the market has assumed that it's going to get tighter. What we've seen is the CPM that has gone down. That is driven normally by a number of factors. One is the mix in inventory that we access but also importantly, in any changes that we do in technology and the improvements that we make on that front. I wouldn’t be able to point to one particular issue related to the CPM but it’s a general trend that I think is important to keep in mind in an environment where the belief is that CPMs are going up because of a change in technology that matches buyers and sellers, in this case, header bidding. So on Search, it’s very early stages on search, we are of course, very excited. We have a great product. We’ve tested it with 30 clients that are all very satisfied with the product and have high retention rates with it. So we feel very confident. The way that we’re going to market with it is we are, as tradition calls for Criteo, pricing on a performance basis so revenue share of the sales we generate for those partners and obviously, the levels of that depend on what kind of ROI those partners have in mind but generally, to make it easy, translates generally in the 10% to 20% of spend or search spend. So that’s the way that we think about it when we think about it financially, but the way we approach the market is different and we definitely want to make sure that it’s a performance product and that obviously has a very good reception by the advertisers that are considering the solution.

So it’s one additional point Doug on this how we are going to account for it. It’s a revenue share that will translate into a percentage of the spend but the way we will account for it is as the term set up -- in the current setup, we act as an agent, we are going to record the net revenue on a net basis. So this would be, in fact, beneficial to the margin mix in the short run.

Okay. And then just real quick follow-up. Did anything shift materially in the quarter just on your mix between direct publisher inventory and what you get through exchanges?

No. It’s very much stable in the 35%, 65%, 35% on the direct side, 65% on the RTB side, pretty much.

Okay, great. Thank you.

The next question comes from Heath Terry of Goldman Sachs. Please go ahead.

Hi, this is Daniel Powell. Sorry. I was just wondering if you could give us a sense of your -- at your Analyst Day, you outlined four stages of adoption around header bidding, the last one being an actual decline in traffic acquisition costs. What stage does this quarter reflect and how far do you feel like that last stage is away from becoming reality?

Thank you. That’s a great question. I think we’re still early in the process. What we had said is that we would think that a couple quarters before so the factors fully reflect the new realities and so this was end of Q3 so I wouldn’t expect that to be in place, if you will, by the end of Q1, Q2 next year. And then we talked a bit about the technologies that we would put in place to help us to take advantage of the new reality. And what I mentioned which is this beta product that allows us to be in the wrapper and bid directly is our first stage into it. We will be testing this, this quarter with the hopes of being able to release a broader version that we would deploy across publishers in Q1. So that gives you a sense for where we are -- still early.

Got you. And then as you look at your e-commerce partners going into Q4, is there a sense that you get of what growth in holiday spending they are planning for? What kind of a consumer environment their budgets would apply?

Always hard. I mean we're November 2nd, right? As you know the real, big part of it starts in three weeks. But generally, I would say we've seen different patterns of spend over the last two years that were quite different, like two years ago, we saw early spend that continued on and last year, we saw a much more spike-y curve. I think we seem to be in between the two now, where there's a good spend at the beginning but not as much. So it depends what the pattern will look like. It's hard to predict but no indication one way or the other at this point.

The next question comes from Charles Bedouelle of Exane BNP Paribas. Please go ahead.

A couple of questions. The first one is, can you just give us a quick impact of HookLogic in Q4 on top of, obviously, the guidance you provided? And the second question I had was on the equity awards, which was a bit high in Q3 and probably outside the user trend so can you comment on that and help us to think about it going forward?

Yes, for HookLogic, we have not closed the transaction. I mean we expect the transaction to close during -- in the next coming weeks during the quarter so it's too early to speak about the contribution of HookLogic. I mean, obviously, HookLogic is going to contribute for less than two months. We know it's already in our Q4 so that is going to be probably not very material to the results at this point. With respect to equity awards, we are the -- a specific situation in Q3, because we have granted an equity award which, in the prior year, had been granted in Q4. So there has been an acceleration of the timing so it's just the timing difference of the equity awards. We have a yearly process for refresh, for retention. We've done this refresh slightly earlier this year than we did last year so that's the reason why, from a timing standpoint, it seems to be a more significant impact on this quarter but on the full-year basis, it would be pretty similar.

And maybe just a follow-up question. Can you talk a little bit about what -- how things are going in China and India and in Southeast Asia in general?

Yes, I think in -- we're pretty happy with the performance of APAC. Southeast Asia, I think, has seen a number of players come in, spend quite a bit and then sort of rationalize a bit more but we're still seeing continued growth. We feel that Southeast Asia will do quite well. India is at the beginning of the trajectory with us. We have a number of large clients that have signed with us and we feel quite bullish about it. And I think China, as we sort of discussed, we have put the foundations in place. We have data centers that are fully operational. We have a team that's working. We also had to connect to the supply very differently and customize the, if you will, the connections and so we're really at a point where we feel that, China, the domestic market could be attractive. And obviously, we're also making quite a bit of progress on the export market and that is Chinese companies that want to expand their business outside of China and obviously, for that we have a terrific infrastructure for them to do so.

Okay, that’s great. Thanks.

The next question comes from Ross Sandler of Deutsche Bank. Please go ahead. Mr. Sandler your line is open. Please go ahead with your question.

Hi, sorry, this is Aki on for Ross. Sorry, I was on mute. Just one quick question. So on the Analyst Day in September, you been talking about given the long-term opportunity for the Search business. I was just wondering if you could talk about what contribution we could expect from that in 2017 and going forward on a net revenue basis? Thank you.

I think we are just at the beginning. We just launched the product so obviously, for this year, it’s expected to be just very immaterial to the figure. For 2017, we will provide guidance. At the time we will release our earnings early February, so it’s too early today to speak about the contribution of search in 2017. What we would expect, obviously, we launched in the US. We just launched in the U.S. We expect to launch in Europe and France, in particular, through the beginning of 2017. This is -- should be a progressive ramp up of the product during the year, so we’ll update you at the time of the earnings for Q4.

The next question comes from Brian Nowak of Morgan Staley. Please go ahead.

Thanks for taking my questions. I have two. Just to go back to the search point, can you just talk a little bit about your go-to-market on Search and whether advertisers can start buying it now in the U.S.? And how much Search revenue is embedded into the fourth quarter guidance and could that be a source of upside to the holiday season? And secondly, on the Americas, the America's gross revenue kind of continues to slow down. Just to talk about why that’s slowing down and if you look at the quarterly targeting business, what could cause that to stabilize or maybe accelerate longer term?

So great questions, Brian. Thank you. On Search, obviously, it’s early and as we talked at the Investor Day, we felt good, some level of confirmation to launch but we felt good about where the product was. We certainly feel good about the value-add that we bring to retailers on Search and have many proof points to say that it is a great product. We are starting to deploy it now. You have to be conscious it’s the holiday season so we don’t expect a lot of people to -- they might sign and be ready to do it but they might not be ready to implement it in the middle of the biggest time of the year for them. But my expectation is that we will start seeing a lot of customers signing up for it, as the new year comes. We, as Benoit said, it's too early to tell what they will contribute, certainly not for 2017, even less for Q4 2016, but as we progress with the business we will give you updates on that. On the Americas, I would – start to think, we’re still the Americas is the largest contributor to growth for us and so it's been decreasing over time but it's still a large contributor for us. And a couple of things to say about the different segments in the America's: mid-market growth remains solid even though we have talked about in the last quarter a slowdown in hiring. And we’re seeing a little bit of the fact that we don’t have as many sellers because of that, but we’re back on track on that. And on the Tier 1 side, we have had a couple of, what we believe, are temporary execution issues that we are correcting. One is we’re changing our leadership for the region and our CRO, Mollie Spilman, is directly managing the region. The region still has plenty of potential so we feel very good about the growth there. And the other thing that we feel is going to be a big contributor for us in the U.S. is we now have a HookLogic team that’s joining us, that's very US-centric, that has a lot of relationships with retailers and very strong ones, and we've seen the combination of HookLogic and Criteo solution can actually accelerate our ability to deploy in these regions. So we remain very bullish. I would say probably that the Americas is going through an unusual time of slowdown but we remain very bullish for the future.

Okay, thanks.

The next question comes from Tim O'Shea of Jefferies. Please go ahead.

Hi, yes. Thank you for taking my question. So just looking at search, is there an ideal or maybe a target customer for the search product and should we expect a high overlap with the existing customer set or should we expect that this might attract a new or different type of advertiser? And then just given Google’s already strong ROI, just broadly speaking, I’m curious how much additional performance you are able to bring with the Predictive Search product? And maybe in early testing, how does the ROI for Predictive Search compare to the core re-targeting product ROI? Thanks.

Great questions, Tim. So on the search advertisers, this is a Google Shopping product so it does relate directly to the retailers we serve that have an e-commerce business and so it’s very much a complete overlap and the intent initially is to go to large to large mid-advertisers so not the top top-tier but the ones just after that. We believe that those will be the ones where we'll have less friction because there will be less established ways of doing this but the overall intent is that this is a product that’s applicable to everybody in the market. As we’ve said in terms of performance, we’ve seen uplifts of up to 49%. I believe that when we talked at the Investor Day, we gave a range of some of the people that have seen it. You have to realize, 49% or there was another example, I think we had, which was 23%, Revolve, I believe. That’s a significant -- if you think about Search and how long it’s been around, and how important it is for all the e-commerce players getting a 5%, 10% improvement in performance on your Search dollars is huge. So when you think about getting a 20%, 40% improvement in performance, that’s a real disrupting product in that segment.

Thank you.

The next question comes from Rocco Strauss, Arete Research. Please go ahead.

Yeah, good morning. Two questions for me, please. A quick one on HookLogic. Aside from tapping into brand budgets here, I guess my question is more about its strategic importance for Criteo and what I've been learning you would gain from now owning actually an exchange, which would provide you with deep -- with much deeper insights of the header bidding of the bidding structure and is that knowledge and advantage with respect to header bidding changing customers -- changing auctions from second price towards first price? I guess that's the first one. The second question, I would appreciate some more color on what actually is behind the take rate increase? I know you touched briefly on this but I guess we try more to understand if that is sustainable. I guess we're trying to understand if the increases more driven by mobile or is that more mix shift towards mid-market or is that a result of any kind of particular tech improvements or was that simply the result of header bidding allowing you to shift towards an SSP and exchanges with lower take rates?

So let's start with HookLogic. HookLogic is very strategic for us and that's why I think this is -- we're at a key milestone in the Company. We're introducing HookLogic as an acquisition. I'll talk in a second about this but also Criteo Predicted Search, which are completely different and really strengthen our offerings with retailers in particular. The immediate priorities with HookLogic would be to scale their business. We work with 10,000 retailers. They have about -- don't quote me on this number, but 100 or so retailers. So we have a significant opportunity to scale them in the U.S. but also to rapidly expand geographically with them and we're going to work very hard to do that. But there's a number of other things that we can do with them that is in the area of synergies and combinations beyond the scaling on the technology side. We can bring a lot of the magic of Criteo Optimization to drive more performance. Remember, these guys perform on a -- or sell on a CPC basis, on a click basis, of course, but they also, just like us, the intent is to drive conversions and so that is what we do and we can take a lot of that technology and apply it to them. We also have a cross-device graph that we've been building that's becoming more and more important. The ability to see sales across devices is one thing that we will be able to do for the HookLogic offering. And finally, if you think about a brand manufacturer that's spending on the CPC basis for, let's say, how they are spending on a particular toy for the holidays, they want to see when they spend something, how many sales they are getting. They don't have a real good way of seeing attribution happen. And we're now able -- or we will be able to provide attribution of sales, not just in the retailer network that they currently have but a full retail network that we have and so that extends significantly the sales that brand manufacturers can see. And then there's a couple of new opportunities and one of them, interestingly enough, is a HookLogic Search opportunity. As we start doing the search for retailers, we can incorporate brand manufacturer dollars into that and you can have joint bidding that happens from retailers and brand manufacturers to drive traffic back and sales to the retailer of those products. And same idea, for potentially bringing on the retailers, re-targeting dollars with the brand manufacturers' dollars to drive traffic back to the store and sales of those products. Think of this at the equivalent but better, turbo-charged version of freestanding inserts or any trade marketing dollars that are done off-site. In terms of the take rates…

Additional color of the take rate, in fact, it is the illustration of what we've been saying for quite a long time is we have a high level of control on our margin. We monitor that margin by market and geography, so it's also the illustration of the power of our scale. So I would describe the margin of Q3 rather on the high end of what we would expect. But we will continue to monitor market by market what is the optimal level of margin that can drive the best outcome for us on our -- for our customers from a revenue ex-TAC standpoint. So -- and that leads to more differentiation but between markets, as you could see across the various regions.

I would say one last thing Rocco, on your first question, on about header bidding and HookLogic. We run a Criteo marketplace that sort of matches suppliers and – I'm sorry, publishers and e-commerce companies. And HookLogic does the same, and I think there was a lot of similarities between that environment and when we think about an environment on header bidder that's a marketplace. And that's why we've said for a while now that we believe in environment like the header-bidding environment is an environment that's very favorable to us and as we develop technology, we will be able to take advantage of that more as an opportunity than a challenge to us.

Thank you.

The next question comes from Murali Sankar of Boenning. Please go ahead.

Hi, thanks for taking my questions. You started talking a little bit about the environment for mid-market hiring. I was hoping that you could maybe elaborate a little bit on that on what you are seeing in terms of trends maybe going forward, whether things are changing? It would also be helpful to get a sense of the size of any kind of litigation-related charges and how long do you see that kind of continuing, assuming that it does indeed go to trial? Thank you.

So a couple on mid-market, I think what we’ve talked about is we had a bit of a hiring gap that was mostly driven by our presence in Boston and whether that was due to a tightening of labor markets or just harder period for us is not clear. At the same time, we've gone out of it and we feel quite good about where we are now. We’re able to get back to the forecast that we had in terms of hiring so we feel good about hiring again. It's hard to predict whether we will have another tightening of the environment. We don't expect it again. The last two quarters have been good. On legal expenses, I don't know if you were able to notice but we arrived at a settlement yesterday, I don’t know what has been included there but probably not much.

Yes so we announced yesterday that we are dropping the litigation on the -- so you should see the litigation expenses, the peak litigation expenses which was around $1 million of litigation expenses in Q3. You should see that continuing through Q4 partially but then beyond Q4, the SteelHouse matter should not drive any more litigation expenses.

And just to be clear, when we went into this legal process, our original intent was always to establish transparency and accuracy around the metrics to judge campaigns. It’s very important to us since we're, of course, a performance player and the health of the industry depends largely also on the transparency and the clarity of how people measure results and all of this. And through this process, we’ve arrived at a point where that became very clear to advertisers the types of offerings that are out there are very clear. How they are measured is very clear. So we feel quite good about where we ended up and I think, overall, the industry is better served by the outcome of that action.

Just to clarify, of the outcome of what was announced yesterday in terms of dropping the litigation, there is no impact, financial impact, neither positive nor negative in -- to be expected out of this process nor with SteelHouse.

And the last question today will come from Tom Champion, Cowen. Please go ahead.

Hi, good morning. Thank you. It just sounds like 57% of revenue came from mobile; that’s a very strong number. I’m wondering if you could comment on the breakdown between web and in-app within mobile. It sounds like in-app, you’re maybe growing very rapidly but also maybe is small and also, I’m curious whether header bidding has any impact in the in-app environment? And then secondly, just curious. I think last quarter about two-thirds of your advertisers shared CRM data. Just curious if you can update that number? Thank you.

Okay, on the first number, obviously, we’re very pleased with mobile. We have now said for a while we track what consumers do and since we have a solution that works across all environments, that’s what’s happening. In-app has grown very fast and in part, has to do with the fact that some of the retailers that we are working with now are taking in-app a lot more seriously and the supply side is also evolving the right direction is becoming more mature. And so that's driving a lot of the growth. When we look at mobile overall, in-app is growing very fast and it's still in the low double-digits of the total of mobile. So very promising and we expect that to continue to grow well but still not the majority of what we see. In terms of advertisers participating in the CRM program, it’s still advancing. And so that’s good. What -- I think what’s worth noting regarding our cross-device graph is we talked about a little bit during the Investor Day and this is the ability now for people that participate in the program to get information, granular information about the cross-device graph to do attribution but potentially also to do customization. So the important part of that is that as an advertiser, not only do you get the benefit of getting better performance, but two, you get the benefit of getting access to these pooled assets to do things that you couldn’t do on your own. And this is sort of now back to a theme that is becoming clearer and clearer is that we are providing more and more tools to advertisers so that they can compete with, on equal terms, with the largest e-commerce platform. We’re bringing them re-targeting, the best in the world that brings back customers to drive their product sales. We’re bringing advertising revenues through HookLogic’s Exchange and we are also bringing state-of-the-art search capabilities, which are very hard to do on your own. So if you start thinking about this as a set of products, we’re becoming more and more meaningful to the retailers and filling a space that is badly needed for the retailers to be able to compete with these large e-commerce platforms.

Thank you. This concludes our call for today. We want to thank everyone for attending the Criteo call. Friederike and myself will be available if you have any follow-up questions. Thank you and have a great day.

The conference is now concluded. Thank you for attending today's presentation. You may now disconnect.