Criteo S.A. (CRTO) Q3 2017 Earnings Report, Transcript and Summary

Good morning and welcome to the Criteo Third Quarter 2017 Financial Results Conference Call. All participants will be in listen-only mode. [Operator Instructions] And please note this event is being recorded. I would now like to turn the conference over to Edouard Lassalle, Head of IR. Please go ahead.

Thank you, Nicole. Good morning everyone, and welcome to Criteo’s Q3 2017 earnings call. With us today are Eric Eichmann, CEO, and Benoit Fouilland, Chief Financial Officer. During the course of 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. Also 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 published earlier today. Lastly, 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 will now turn the call over to Eric Eichmann, Criteo’s Chief Executive Officer.

Thank you, Edouard, and good morning everyone. I am pleased to report another strong quarter for Criteo. We continue to deliver high growth while increasing profits and cash flow. Before we walk through the quarterly results, let me say a few words about the large emerging opportunity our Commerce Marketing Ecosystem opens up for us to serve retailers and brands in a fast evolving market. The battleground has shifted for commerce players, and today success increasingly depends on the ability to know your customers and serve them intelligently. This comes with multiple challenges for marketers. One, the seamless understanding of the shopper journey requires gathering granular shopper data in real time, both online and offline, and on a mass scale. And two, even more important, the activation of this dataset for commerce requires predictive technology, consumer reach at scale, and real-time advertising optimization to drive sales. Commerce marketing is quickly emerging as the next big marketing category after search and social. It focuses on inspiring people to buy things and is measured by performance, directly driving sales and profits for marketers. And unlike search and social, it is not limited to digital, as close to 90% of commerce still happens offline. To help marketers take advantage of these trends, we are building the highest performing and open commerce marketing ecosystem that delivers performance at scale to retailers and brands. The Criteo Commerce Marketing Ecosystem is powered by over $550 billion of online sales and associated intent data, state of the art machine learning, and unrivaled consumer reach across thousands of publishers. The Criteo Shopper Graph, built through data sharing among ecosystem participants, is made of three trusted data collectives. One, the Identity Graph, allowing us to match users’ identities across devices and environments; two, the Interest Map, collecting and organizing user intent and purchasing data across retailers; and three, the Measurement Network, providing sales attribution for brands across retailers. Every participant gets much more than they contribute. Brands and retailers get measurable sales and profits as they acquire, convert, and reengage customers online and in stores, and retailers get advertising revenue from brands on their site inventory. Our Commerce Marketing Ecosystem resonates well with Chief Marketing Officers worldwide. A Forbes Insights study recently found that 71% of retailers are willing to contribute data to a pool. This provides them with the opportunity to level the playing field with the largest commerce players when it comes to online data, technology, and scale. We believe to be the ideal commerce marketing partner in this fast evolving environment. Turning now to Q3 results, for 16 quarters in a row we have exceeded revenue ex-TAC and adjusted EBITDA guidance. At constant currency, we grew rev ex-TAC 32% to $234 million and adjusted EBITDA 45% to $79 million. We grew same client revenue ex-TAC strongly again, to 14% at constant currency. Better technology and inventory access drove increased client spending. This elastic demand comes from the 78% of our revenue ex-TAC with clients that have uncapped budgets. We performed well across the entire business. We delivered strong innovation across existing and new products, we broadened and improved our access to inventory, and we added new clients across all regions and products. Starting with innovation, I want to highlight three areas; the engine, Criteo Shopper Graph, and increased transparency for clients. One, our various features in the engine that better align our prediction with clients’ gross margin objectives have been rolled out to clients, representing 13% of rev ex-TAC, 2.5 times our Q1 rate. Two, our three data collectives that make up the Criteo Shopper Graph grew further in scale and efficiency. Our world-class Identity Graph provides as good or better CRM onboarding rates than the largest Internet players. In addition, we are making strong progress in building the Interest Map that supplies user intent and purchasing data across retailers in our ecosystem. Many of our clients are now sharing more comprehensive product level information such as brand name, category, and a universal product ID, which helps us build our universal catalog across retailers. The Criteo Shopper Graph is a solid foundation to build compelling new products. Three, we are significantly increasing transparency by providing clients with customized reports showing detailed impression level information. These includes publisher domains where their ads are shown, timestamps of displayed ads, encrypted user IDs, and the price clients are paying for each impression. We believe that more transparency will further increase the confidence clients have in our platform without changing the dynamics of our core business. We have been building standardized reports for the Criteo interface, which will be made available in the coming quarters. Moving to the supply side of the business, we’ve continued to roll out Criteo Direct Bidder, our header bidder technology, now connected to 950 large publishers globally. Recent partner additions include eBay, Baidu, United Internet, Axel Springer, and the LA Times. We also developed a solution specifically for the APAC ecosystem which is gaining good traction. In each market, we are focusing on signing the largest publishers first. The additional publisher yield from Criteo Direct Bidder is 20% to 40%, of which about 50% is incremental to our overall business. We have several more optimization opportunities for this year and next, such as strengthening our bidding strategies for first price auction environments. On the Criteo Sponsored Product side, we broadened our inventory supply, signing large retailers like Jet and Saks Off Fifth in the U.S., and mySupermarket in the UK. Finally, we strengthened our brand safety initiatives to ensure that the brand equity of our clients is preserved at all times. We improved our many internal systems and processes to filter out inventory in real time, including the list of suspect IP addresses from the Trustworthy Accountability Group and the lists of invalid traffic from three specialized external vendors. We are also part of the Coalition for Better Ads supported by Google, and are compliant with their recommendations for the most user friendly ad formats. Switching to client expansion, we added 930 net new clients across all regions, products, and sizes, while maintaining client retention at 90% for the core product. We closed the quarter with over 17,000 retailer and brand clients, a 34% increase year-over-year. The midmarket continues its fast trajectory, with revenue ex-TAC growing close to 60% faster than the group and representing 34% of our total business. From a regional performance perspective, we executed in line with plans as our core business continued to grow well across all three regions. The Americas’ revenue ex-TAC grew 35% at constant currency. Within the region, our U.S. business grew 43% at constant currency, driven by the continued high pace of new business and additions in both tier 1 and midmarket clients. EMEA revenue ex-TAC increased 24% at constant currency, largely driven by growth in existing client spends across verticals and client sizes, in particular in our large markets. We expanded Criteo’s Sponsored Products with large European retailers and incurred exceptional publisher costs to seed the future growth of the product in the region. APAC revenue ex-TAC growth accelerated to 40% at constant currency, led by existing tier 1 clients in Japan and Korea and new midmarket clients across the region. I will now update you on Criteo Predictive Search. Given the exciting prospect of the Criteo Commerce Marketing Ecosystem and its associated products, we have decided to discontinue the marketing testing of Criteo Predictive Search and redeploy its associated R&D resources to other projects. Every time we launch a new product, we mean to create a disruptive proposition. With Search, though we created performance for customer and had good market traction, the measurement of the performance uplift over the competition has proven to be very volatile, making it difficult to build a scalable performance-based product. The retirement of Criteo Predictive Search will depend on client and country specific circumstances. Innovation is at the heart of what we do. And as we add new products to our portfolio, we will continue to evaluate all products with strict performance and scalability expectations. Speaking of innovation, we launched two new exciting beta products on October 10. One, Criteo Customer Acquisition helps retailers acquire new customers by using intent information across a pool of retailers. With this product, New Look, a UK-based fashion retailer was able to generate four times more orders at a 74% lower cost per order compared to their other display prospecting acquisition partners. After a successful test in the UK, we are now extending the beta version of Criteo Customer Acquisition to the U.S., France, and Germany. Two, Criteo Audience Match allows retailers to accurately target and reengage customers that are already in their CRM system in order to drive more sales. With this product, La Redoute, a French fashion and homeware retailer, was able to match 70% of their lapsed shoppers online and saw a 50% increase in sales from their lapsed shoppers. The beta version of Criteo Audience Match is currently rolled out across all markets except Japan. These two new products leverage the capabilities of the Criteo Shopper Graph. With these additional products, we will be able to cover the entire consumer journey for retailers and brands. Let me know say a few words about Apple’s Intelligent Tracking Prevention, or ITP. ITP makes it more difficult for third party providers to access data on Safari users and is enabled by default on mobile and desktop. We have always been strong advocates of user privacy, and will continue to actively champion users’ choice over their online experience. For many years, we have provided consumers with privacy friendly controls in Apple environments, and we have done it again for the ITP environment. Our ITP solution is able to collect anonymized commerce data across sites and works well on advertiser websites and with most publishers. We will continue to adapt our solution to potential new changes in the Apple environment. We believe that a data-driven, privacy friendly advertising environment is beneficial to consumers, publishers, and advertisers alike. Our solution for Safari users currently allows us to mitigate about half of the potential impact from ITP. Given ITP was released on September 19, 2017, it had a minimal net negative impact on our revenue ex-TAC in Q3 of less than $1 million. Based on our expectations for the iOS rollout and our coverage of Safari users, we expect ITP to have a net negative impact on our revenue ex-TAC in Q4, between 8% and 10% relative to our base case projections for the quarter. We will continue to improve and deploy our solution for Safari users over the coming quarters. Looking into the fourth quarter, we are focused on four key priorities. First, execute successfully during the U.S. holiday season across the core product and Criteo Sponsored Products, second, improve and roll out our solution for Apple users and mitigate the impact of ITP on the core business, third, build and leverage Criteo’s Shopper Graph based on our three data collectives, Identity Graph, Interest Map, and Measurement Network, to further strengthen our product portfolio and benefit the entire ecosystem, and fourth, further develop and scale our latest products, Criteo Customer Acquisition and Criteo Audience Match, and build new capabilities such as lookalikes, app installs, and omni-channel. In closing, I am pleased with our solid results. We executed well across the business and delivered sustained high growth while increasing profitability and cash. I look forward to updating you as we continue to build the Criteo Commerce Marketing Ecosystem to drive sales and profits for retailers and brands. With that, I will now have Benoit Fouilland, our CFO, walk you through our financial results and guidance in detail.

Thank you, Eric, and good morning everyone. I am also pleased with our Q3 performance. We continue to deliver high growth coupled with increased profit and free cash flow while investing in the business. This attractive combination remains the key differentiator of our strong model. I will walk you through the quarterly performance and share our guidance for Q4 and full year 2017. Q3 revenue was $564 million, up 32% at constant currency. Revenue ex-TAC, the key metric we use to monitor business performance, grew 32% at constant currency to $234 million, driven by a 14% growth in same client revenue ex-TAC and the addition of 930 net new clients across regions, sizes, and products. Revenue ex-TAC margin was 42%, in line with expectations and the prior year level. In the core business, both tier 1 and midmarket clients continued to show healthy growth in revenue ex-TAC per client at constant currency. Compared with our guidance assumptions, changes in ForEx had an approximately $3.5 million positive impact on the revenue ex-TAC, driven by the stronger euro and Japanese yen. In addition, compared with the prior year, changes in ForEx represented a tailwind of approximately 70 basis points to revenue ex-TAC growth. Moving to expenses, other cost of revenue, comprised of hosting and data costs, grew 34% to $30 million, driven by increased hosting capacity across data centers as well as additional third party data to complement our Identity Graph. Operating expenses increased 31% to $171 million. Headcount related expenses represented 75% of GAAP OpEx. Global headcount grew 23%, closing the quarter with over 2,700 employees. Non-GAAP operating expenses grew 26% to $140 million. On a non-GAAP basis by function, R&D expenses grew 42% to $35 million, in line with our investment plans and largely driven by a 37% increase in headcount to 650 R&D and product employees. Sales and operations’ expenses grew 31% to $80 million, also in line with our operating plans. This was largely driven by an 18% increase in headcount to 1,600 employees. Quota-carrying headcount also grew 18% to 755, with over 70% of the growth coming from midmarket. Our G&A expenses decreased 2% to $25 million despite a 21% increase in headcount to 460 employees. Excluding exceptional legal fees incurred in Q3 2016 in connection with the HookLogic acquisition and a litigation, G&A non-GAAP OpEx grew 11%. Switching to profitability, adjusted EBITDA grew 45% at constant currency to $79 million, driven by the strong growth in revenue ex-TAC and sustained operating leverage across the organization. Adjusted EBITDA margin was 34% of revenue ex-TAC, up 340 basis points, or over 170 basis points after excluding exceptional legal fees incurred in Q3 last year. In line with expected seasonality, Criteo Sponsored Products had a negative contribution to adjusted EBITDA in the quarter. Excluding the impact of Criteo Sponsored Products, adjusted EBITDA margin was over 36% of revenue ex-TAC, an increase of 550 basis points compared to Q3 2016. Our adjusted EBITDA margin for the first nine months of 2017 puts us ahead of our plan for the full year. Equity awards compensation expense increased 58%, largely driven by the $2.9 million one-time equity grant in connection with the HookLogic acquisition in November 2016, which represents 36% of the growth. We continue to expect equity awards composition to represent approximately 6% of revenue ex-TAC in fiscal 2017. Depreciation and amortization expense increased 61% including $3.9 million dollars or 44% of the increased in amortization of intangible assets relating to the HookLogic purchase price accounting. Financial expense was $3 million primarily driven by hedging cost on the euro dollar exchange rate on interest expense related to the funding of the HookLogic acquisition. The effective tax rate was 26% based on our estimated effective tax rate of 30% for fiscal 2017. Net income increased 51% to $22 million driven by 44% increase in income from operations and partially offset by higher financial expand and marginally higher provisions for income taxes. Excluding non-cash accounting impact from the HookLogic acquisition on equity awards compensation and amortization of intangibles, net income increased 98% to $29 million. Adjusted net income per diluted share increased 37% to $0.65. Cash flow from operations grew 41% to $62 million, driven by strong operating growth and profitability in the quarter, offset by a negative contribution from changes in working capital. This represented a conversion rate of adjusted EBITDA into cash flow from operations of 78% in the quarter and 88% over the first nine months of 2017. CapEx increased 40% to $28 million, driven by continued investments in data center equipment, leasehold improvements, and capitalization of IT projects, in line with our projections. As a result, free cash flow increased 43% to $34 million, driven by the stronger cash flow from operations and slower growth in CapEx over the period. Finally, cash and cash equivalents increased by $88 million in the first nine months to $358 million. I will now provide you with our guidance for Q4 and fiscal year 2017. The following forward-looking statements reflect our expectations as of today, November 1, 2017. Based on our expectations for the iOS11 rollout and our coverage of Safari users, we expect ITP to have a net negative impact in Q4 of between 8% and 10% of our revenue ex-TAC relative to the base case projections for the quarter. In addition, compared to Q3 growth, Q4 comes with a tougher comparable as HookLogic was included for eight weeks in Q4 last year. These two factors provide a more challenging backdrop for the year-over-year growth in Q4. As a result, we expect Q4 2017 revenue ex-TAC to be between $260 million and $263 million. This implies constant currency growth of 13% to 15%. The inclusion of HookLogic in the fourth quarter 2016 base drives a seven point sequential decrease in the year-over-year growth compared to prior quarters. We assume that year-over-year changes in ForEx will have a positive impact of approximately 213 basis points on our reported growth in Q4. And we expect Q4 2017 adjusted EBITDA to be between $106 million and $109 million. driving a year-over- year improvement in the adjusted EBITDA margin for Q4 of approximately 390 basis points to 450 basis points. For fiscal 2017, we are adjusting our revenue ex-TAC growth expectation to between 26% and 27% at constant currency to reflect the impact of ITP in Q4. We assume changes in ForEx to have a positive impact of 30 basis points on our reported growth for the full year. In parallel, we are increasing our guidance for adjusted EBITDA margin for fiscal 2017. We now expect our adjusted EBITDA margin to improve by 100 to 120 basis points in fiscal 2017 as measured against the 30.8% adjusted EBITDA margin we reported in fiscal 2016. In addition, we believe we will continue to see a strong transformation of adjusted EBITDA into cash flow from operations in excess of 70% for fiscal 2017. .: Looking at 2018, we currently expect that Apple’s ITP could have a net negative impact on our revenue ex-TAC of 9% to 13% next year relative to our base case projections for the year. This assumes that the roll-out rate of iOS11 increases to 90% of addressable Apple users by the end of the summer 2018. This also assumes that our ability to mitigate the ITP impact continues to improve over time. In parallel, we are confident that ITP should not drive a significant impact on our adjusted EBITDA margin next year. In closing, I am pleased with our strong performance in Q3, which delivered sustainable profitable growth on strong cash flow generation. These results highlight the strengths of our business and our confidence in our position and growth prospects. We remain focused on executing on our plans for the fourth quarter and building the Criteo Commerce Marketing Ecosystem for continued growth in 2018 and beyond. With that, I will turn the call back to the operator to take your questions.

Thank you. [Operator Instructions] Our first question comes from Dan Salmon of BMO Capital Markets. Please go ahead.

Good morning everyone. Eric, could you perhaps expand a little bit on the decision to retire Predictive Search? I think in your prepared remarks you mentioned a little bit around that variability of traction and performance. I’d just love to hear a little bit more about that. And then second, you spent a good amount of time on retail data co-op. Could you just expand a little bit on how you’re working with retailer partners to get offline sales data in particular? Thanks.

Okay, great. Thank you, Dan. So let me first talk about Criteo Predictive Search; a couple of things to mention on it. First is, as you know, this for 2017 really doesn’t have a significant impact on our top line. We have provided sort of guidance for 1% to 2%, so it’s minimal from that perspective. We had been in the market now for about eight, nine months. And in that period of time, we have about 200 clients live with Criteo Predictive Search. And when I mentioned we have some traction in the market, so we’ve been very sort of proud of the effort that the team has put in place to get clients signed up. One of the challenges that we had since the beginning, and we probably sort of got a better appreciation as we were in-market, is the ability to measure the performance of the solution over other solutions. And that has proven to be a very difficult thing to do for a couple reasons. One, there’s no established head to head protocols out there like you have in display. And so we had to work to establish one of our own, which we think is stable enough, but it’s not standard in the industry and it’s hard to sort of implement with partners. And the second this is that, unlike sort of the elasticity that exists in display, you don’t have one metric of sort of uplift in sales. You often have two metrics that come into play. One is the cost that you’re able to sort of achieve for the partner and the sales that you’re able to achieve. And so, those two metrics don’t always calculating to an easy sort of one metric comparison. And so though in the clients, we obviously have a number of clients and those clients are happy with sort of what we bring to them, we didn’t believe that we could make this a scalable product that we could sort of take into many markets and make it a big product for us. And given the opportunity that we have with – and I’ll talk in a second about the data co-ops, but the opportunity that we have in building a Commerce Marketing Ecosystem that has a number of products, of which we just talked about too today, we thought it was wiser to sort of redeploy those resources to work on those other products. So that’s regarding search. In terms of the data co-op, obviously we’re excited about not just the fact that we’re in a great position to establish these data co-ops, but also the reception from CMOs about this. More and more people are realizing that you need scale and you need the technology to activate that data. And a lot of the retailers we talk to don’t have the capability to do those two things, and we become a great partner for them to do that. When it comes to offline sales, which is one of sort of – we perceive offline assets as assets as opposed to what’s perceived as sometimes being liabilities because of their cost. Our ability to bring those into the data co-op relies heavily on our ID Graph. Our ID Graph allows us to take CRM data and convert that into or associate it with online identities to the tune of 60% to 70%, depending on the regions that we’re in. And so one of the things we talked about in the last earnings call was about the test that we’re running for store to Web retargeting. And that basically takes data from offline and then targets those customers or those visitors to offline stores online with sort of taking into account their purchase history. So I would say we’re at the early stages of sort of using that data. In a lot of cases, sometimes to activate that data you just have to go through old systems and sort of bring that data in. But we see a lot of potential over time for that data to be part of the co-op.

Great. Thanks, Eric.

Our next question comes from Sarah Simon of Berenberg. Please go ahead.

Yes, can you hear me?

Yes, yes.

I’ve got couple of questions. First one, just staying on the e-commerce marketing thing, obviously the fact you’re retiring search is something that some people will say is disappointing. But can you give us a feel for how reception of e-commerce is – the e-commerce marketing proposition has been relative to the sort same point in development of search, and also whether there are any issues we’re going to encounter in terms of measurability here that could also sort of impact the take up of e-commerce? And also, I know you have quantified customer acquisition previously in terms of the market opportunity. But do you have a sense for the overall market opportunity here versus what you were talking about with search? And the second one was on the margin. Your Q4 margins increased – or let’s say the margins implied by your Q4 guidance increase quite a lot year-on-year, and that’s despite the impact of ITP. And sorry, can you just go through in terms of your Q3 – or, sorry, the Q4 guidance, what the components were in terms of sort of like-for-like HookLogic drag, FX, and whatever’s going on in terms of ITP? Thanks.

Great. And thank you, Sarah. So let me just maybe take the first couple questions, and then I’ll pass it on to Benoit. So one, I would say the commerce marketing narrative has been very well received. And we’re helped by the fact, obviously, that large e-commerce platforms are creating a lot of pressure on existing businesses. So the urgency around sort of doing something and the understanding that data and technology are driving a lot of the value today, sort of what we call understanding your customer and serving them appropriately, has helped. I think, look, if you look at the products that we have, one of the tenets of search is that we would use our technology in optimization, which I think we’ve done well, and to a certain extent what we’re calling the data collective, the Interest Map or sort of the ability to see what shoppers have done on sites. And what we’ve found also in looking at that is the technology we’re able sort of to – as I said, the clients are with us and we’re able to add value. But the value coming from the Interest Map, on top of what is already an environment where you have a stated intent with the search keywords that happen, was not as significant. In terms of the Commerce Marketing Ecosystem and the products we’re talking about, a lot of them rely on an open environment that is not sort of driven by one or two suppliers, and so the ability to measure success is very much similar to what we do in dynamic retargeting. And so we don’t see a problem there because there’s a very well established way of looking at comparables. I think search, since the beginning, was a particular market where doing that was more difficult, because the ability to sort of split up users and see the impact of one technology over another was always quite limited. One of the exciting things about the Commerce Marketing Ecosystem is all the products that I’ve talked about in the call all rely heavily on the data collectives, and I don’t think that was as much the case with search. So we feel quite strongly that not only can we bring to bear the technology expertise that we have and the supply network that we have, but we can leverage what is turning out to be a very unique asset. You also asked about customer acquisition. Customer acquisition is probably the biggest sort of ask from our customers when we ask them what else can we do for you. They’d like us to move up in the funnel and sort of bring customers that haven’t come to their site. In the tests that we’ve done – we launched actually a test in the UK for this before launching it or sort of rolling it out more broadly – we’ve seen customers spending about 15%, one-five, of what they spend on retargeting. Having said that, we believe that the potential is much bigger as we sort of improve the technology over time. And so, if customer acquisition is something not unlike retargeting, the more you can do at the right ROI the more people will spend. So it will be sort of reliant, I think the potential, on how well we’re able to build the technology and sort of like how well we’re able to leverage the data pool. So with that, let me just turn it to Benoit to answer the questions about the margins for Q3.

Yes. So the margin – thanks for the question. On the margin for Q4, it’s a combination of the fact that we have been delivering – we are ahead of our plan for the nine months, and we’ve been delivering better margin in Q3 and expect to deliver also a better outlook for Q4. It’s a mix of several factors. First, we had a very disciplined and focused hiring on critical functions. If you look at where we have been hiring, we’ve been prioritizing MMS on one hand, on R&D on the other hand, with a modest increase in terms of resources in other functions. So we had a paced and focused hiring. We have also – with respect to Q4, we have pretty effective variable pay policy internally which allows us to compensate for some of the impact of ITP from a cost standpoint on the variable pay through the company. And the third point I would highlight is the fact that our model continues to deliver operating leverage as a result of the pull- through levers that have been pushing our margin for now multiple quarters. So if we look at the guidance for Q4, I’d like to just remind that we have never provided guidance specifically for Q4. We had communicated some caution – at the time of the Q2 earnings release some caution around Q4, as we were – it was still early at that point to estimate the impact of ITP. So if you look at compared to the prior outlook that we had given for the year, the variance is approximately around – just above $20 million down compared to the prior outlook for the year, and that is essentially the impact of ITP. There is also a small impact from search, because we are going to see obviously less revenue ex- TAC from search in Q4. But essentially it is the impact from ITP that explains this $20 million decrease in outlook for the full year. With respect to margin for next year, what I had indicated is, given the focus that we have in allocating our resource, we believe that the ITP impact that we have highlighted for next year should not drive a significant impact on our adjusted EBITDA margin next year when it’s compared to the one that we believe we could achieve this year.

Okay, thanks.

Our next question comes from Mark Kelley of Citigroup. Please go ahead.

Hi, good morning. Thanks for taken my questions. first one is just back on ITP. You talked a little bit about improving and deploying some solutions to offset some of the impact. Can you go into a little bit more detail there? Are these technology changes you still need to make and then implement, or are you talking more about procedural changes that you’ve already developed, you just need the clients and publishers to adopt them? And then second, I know it’s just been over a month since ITP was released. But are there any differences or nuances that you’ve seen in the public version versus the beta version that you now have to react to? Thanks.

Okay, great. Thank you, Mark. So on ITP, just a couple of things. What we’ve developed is a privacy friendly solution. We already have a privacy friendly solution for Safari users. And we continue to do that, and so we feel quite good about that. The solution relies on sort of the ability to have an identifier that is not a cookie but is allowed within sort of the transfer information between websites and our servers. I think two things in terms of how we expect this to evolve. One, it’s a new technology for us, and so we sort of have to work out the kinks, if you will. And so those are sort of a box here and there that we have to make sure work. And then the other thing is it works best when we’re connected directly to publishers. So we’re rolling these out with publishers directly, and the more we have a presence on a direct bidder perspective the better it is for us. The other change is sort of working with the exchanges to make sure that that information can be transferred through the exchange. We haven’t done this across all exchanges, and some are more difficult than others in terms of the implementation. And so that’s part of sort of what sort of creates some of the, if you will, uncertainty around sort of how much of it we can mitigate. In terms of the difference between what was public or what is public now in terms of the technology and what we expect, there were a couple of changes back in the days before they – it wasn’t clear that it was going on by default. It wasn’t clear it was going to be on mobile or not. But by and large, sort of the way it’s working is what we expected it to be. And that’s why, at the moment that sort of iOS11 was released, we had a solution that we were putting in place very quickly. So hopefully that answers the question.

That does. Thanks Eric.

Our next question comes from Douglas Anmuth of JP Morgan. Please go ahead.

Thanks for taking the question. I guess maybe first, just to follow up a little bit on ITP, just wondering, Eric, if you can help us understand in a little bit more detail how the solutions are working on both publisher inventory and then also exchange inventory and what some of the key differences might be there? And then second, can you just give us some of your early thoughts here on how you’re thinking about GDPR in 2018? And then if we think about Benoit’s comments about 9% to 13% the lower revenue from ITP in 2018, is there any impact from GDPR in there or is that just solely ITP-based? Thanks.

All right, great. Thank you, Doug; great question. So on ITP, just to be a bit more precise, the solution, which we call HSPF on our end, uses the HTTP strict transport security protocol. And that’s just another field in the information that goes back and forth, and so it allows you to sort of store an ID within that field without having to use the cookies. Again, like I’ve said since the beginning, this is a privacy friendly solution where we sort of seek consent from users, just like we had done before. And so the question, when it comes to exchanges, is are they sort of using this or sort of are they okay using this mechanism. And for most of them, the conversations are good. Some of them are already implemented. For some others, sort of we need to work harder with them to get that. It just – yes, it’s one mechanism where you can exchange the data and see the data. There are other sort of mechanisms that you could use over time. And this is why we could find ourselves in ways where we need to continuously adapt our solutions, and that’s one of the things that we think about constantly. So hopefully that gives you a better sense for it. GDPR; as we’ve said in the past, we feel quite good about GDPR, though there’s quite a bit of uncertainty still about what it means exactly. What we believe is it does a couple of things that we’re excited about. One, it sort of takes regulation that is sort of country specific and makes it sort of region wide. And from that perspective, it’s going to be easier for us to track that. Two, I think one of the things that GDPR is looking for and is not – hasn’t been as specific on yet is to make sure that there is clear consent, something we do today. And I think one of the concerns of the commission as we understand it is that it’s a bit clunky to have each website sort of ask for it, so they’re looking for ways to do this. And then there’s a third element about sort of a legitimate reason to be sort of having the information, which we again feel quite good about. So I – and then I pass on the question to Benoit about the guidance.

So with respect to the guidance, the impact we indicated next year for – the expected impact at this stage for ITP, we don’t expect any additional impact on the top for GDPR. We don’t see a negative impact coming from GDPR on the top of what has been indicated already for ITP.

Great. Thank you, both.

Our next question comes from Charles Bedouelle of Exane. Please go ahead.

Good morning everyone. So, two questions really. The first one is on the – really your new initiatives around getting the CMOs to work with you in a more – a deeper way and a more diverse way. Can you give us any sense of how do you think this can impact next year and maybe the next two years, both in terms of investments but also revenues? And also, just to come back and just to make sure I understand well on the ITP impact, if we take the Q4 impact you mention and the next year’s impact you mention, how much do you think it will really take off in total of your revenues maybe in U.S. dollars or as a percentage of revenues? Because it seems to be a little bit higher than maybe what was indicated before? Thanks.

Great. Thank you, Charles. So on the new initiatives and the CMOs, we mentioned that a Forbes Insights survey showed that 71% of CMOs are willing to share data as part of a pool. And we have seen this actually in practice when we ran the customer acquisition beta in the UK. Out of 10 customers, eight were okay sharing the data in a pool. So in terms of investments, I think this is very much in line with what we’re planning to do with the data collective. Remember for the Interest Map we already have all the data. What we need to get is permissions from the companies to be able to do it. And then the second thing we need to do is really invest in making sure that the data is consistent across retailers so that we can find a red hat at one retailer and sort of also see that that red hat exists at another retailer so we can link sort of the information across retailers. We haven’t provided any guidance about next year in terms of revenues for these products. We see the potential of not just this product, but the portfolio of products that we’re going to bring as part of the Commerce Marketing Ecosystem being a significant opportunity for us. Again, we haven’t provided sort of any guidance on this. But certainly in talking to CMOs and in talking to the companies we work with, the excitement is real. And it also allows us to have a conversation at a higher level at the client, because obviously now we’re talking about activating their data offline, about filling sort of needs across the funnel in a way that sort of makes them more defensible or puts them on an equal footing against a large e-commerce platform. In an ITP I’ll let Benoit…

So before I add on ITP, I wanted to mention with respect to new products for the investment next year, we compare what we’ve seen this year where we had significant investment with respect to – especially on the R&D integration of Sponsored Products into our core technology where we are now 12 months into an 18 months roadmap. So we should see this effort going down next year as we complete the full integration. And with respect to the secondary investment that we had around Sponsored Products, we have now integrated the sales force and we are coming to market with an integrated sales force for next year, which should also lower the requirement in terms of specific sales force specialists with respect to Sponsored Products, just to put this in perspective for next year. With respect to ITP, we indicated that dollar amount for Q4, so it’s around $20 million of an impact from ITP negative on the revenue ex-TAC. Obviously, for next year, as you know, we will guide only next year when we will release our results in February. So what we’ve chosen is to give you a range with respect to the negative impact in percentage, so the 9% to 13%. And what are the factors behind it? First is that we anticipate that the rollout of iOS11 is going to continue to increase to reach around 90% by the mid of the year, which would be we believe, the plateau based on experience. And what we are anticipating as well is that our current ability to offset, which we are close to 50% or about at 50% with also increase over time. So these are the two components. So the revenue at risk; if you look at the revenue at risk today, we are – around 17% of the revenue ex-TAC is, I would say, at risk in Q4, of which we offset approximately 50%. And if we look at next year – and we are at 17%, not 70%, 17%. And with respect to next year, we believe that this revenue at risk is going to be increase slightly above 20% to reach around 23%; 22%, 23%, and as a result of the increased penetration of iOS. But we believe as well that we are going to increase our ability to offset. So these are the parameters that we can share with you today.

Yes. But Benoit, if I just may on this one, since you’ve had a few weeks of ITP and actually a full quarter in Q4, shouldn’t the impact that you mentioned be on three quarters of the year rather than the full year? Because next year when you lap into Q4, you should probably be on a more comparable basis. And Q4 is actually quite big last – next year.

No, approximately – yes, absolutely; absolutely. While we will be in Q4 next year, I would say the initial impact will be behind us. And then we will carry then a pace of growth that should not be impacted any more by ITP at this point.

Right. Another way of saying that is that obviously ITP to a certain extent sort of resets a base from which, once we’re in it, sort of we go back to what – our core growth of the business. And obviously, there’s also a possibility we mitigate a larger part of ITP, right? So that’s what we’re working on.

Okay. That’s very clear. Thank you very much.

Thank you.

And our next question comes from Richard Kramer of Arete Research. Please go ahead.

Hi, couple of questions. And just to follow-on the last one, if people in the market are having difficulties buying Safari inventory and that should lead to lower CPMs there, can you see higher take rates on other inventory sources and increase bids and sort of mitigate the impact with other inventory sources? We’re kind of talking about one inventory source in isolation, so that’s one question. Second, maybe Benoit, you can just help us understand exactly the implied loss in CSP, I think it was suggested it was about $5 million, and what the revenue ex-TAC was in 3Q. And won’t that move into a situation next year where that should be a meaningful contributor to profits? And the third one, I know it’s one we’ve talked about in the past, but the cash is now back at the level that it was before you bought HookLogic, or almost back to the same level. Given how you’ve had a slide in the equity, why wouldn’t you look at putting some of that cash back into work in a buyback or at least to mitigate dilution or use that cash somehow? Because you keep generating more cash but it’s not reflected in the share price. Thanks.

Great. So thank you, Richard. Let me just take the ITP one and I’ll pass it on to Benoit for the other two. So ITP, if users are sort of accessing most of their activity to Apple devices, it is hard to find those users in other places. We can find them in the app environment where Apple sort of has no restrictions about IDs. But in general, it’s not as easy. If they go to work and they’re using a PC or they’re using Google Chrome, we can certainly find them, but there’s part of their history that we lose. Now the positive side of ITP is that through our solution we’re obviously accessing inventory that many people cannot access. So the competition of that inventory is much lower. So not unlike sort of some of the effects on the inventories, we might find ourselves being some of the only ones with the profile that we have accessing and buying that inventory. And so we can get better prices and higher take rates on that inventory. The net effect of that is still to be seen. We frankly don’t have a good idea of what that is. But again, it’s not just an inventory source sometimes. If users are sort of exclusively using Safari browsers, it’s hard to sort of find them in other places. So with that, let me just pass it on to Benoit about the two other questions.

Yes. So with respect to the loss resulting from Sponsored Products, you are right. It’s approximately $5 million loss on EBITDA for the quarter. So we’re still looking for the year at a slightly dilutive impact in terms of absolute impact on adjusted EBITDA. So I would say middle digits in terms loss. So for next year, we expect to see growth in – significant growth there in Sponsored Products. And as a result of lowering the effort on the integration – R&D integration, as we will complete the R&D integration by that time and also the integration of the sales force, we should see Sponsored Products turning to a contributor to the EBITDA, not yet to the margin but to the EBITDA next year. With respect to the cash, you’re absolutely right. We are not yet at the point where we have restored the cash balance, the pre…

Almost.

But almost. And with the comments made on cash flow for Q4, you could expect that we would get there pretty soon. So we will look at the best alternative for the use of cash, which is always a tradeoff between investment needs that the business might have in terms of selective acquisition versus potentially looking at buying back shares.

And I would say obviously in that equation we’re embarking on a significant endeavor to bring a Commerce Marketing Ecosystem for retailers, a brand that has been well received, and there’s a number of products that we’re going to build there. And as we see companies or propositions outside of what we can do organically, we’ll certainly look at acquisitions continuously. And we see a big opportunity there.

Okay. Thanks.

Thank you, Eric. This concludes our Q&A session. We thank everyone for attending the Criteo call today. The team will be fully available if you have any follow-up questions. Goodbye and have a great rest of the day. Thank you.

Thank you.

The conference has now concluded. Thank you for attending today’s presentation. You may now disconnect.