Edouard Lassalle - VP & Head, IR Jean-Baptiste Rudelle - Co-Founder, Executive Chairman & CEO Benoit Fouilland - CFO.

Timothy Nollen - Macquarie Research Brian Wieser - Pivotal Research Group Matthew Thornton - SunTrust Robinson Humphrey Sarah Simon - Berenberg Andrew Hargreaves - KeyBanc Capital Markets Nicholas Jones - Citigroup Mark Kelley - Nomura Securities Brian Fitzgerald - Jefferies Daniel Salmon - BMO Capital Markets.

Good morning, and welcome to Criteo 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, Keith. Good morning, everyone, and welcome to Criteo's Third Quarter 2018 Earnings Call. With us today our cofounder and CEO, JB Rudelle; and CFO Ben Fouilland. During the course of this call, management will make forward-looking statements.

These may include predicted 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. Such 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 which and the reconciliations to the most directly comparable GAAP financial measures were provided in the earnings release published earlier today.

Finally, unless otherwise stated, all growth comparisons made in the course of the call are against the same period in the prior year. With that, I now turn the call over to JB Rudelle, CEO of Criteo.

JB?.

Thank you, Edouard, and hello everyone. During Q3, we met the financial objectives we laid out 90 days ago. Despite this achievement, I'm not entirely satisfied with the results, as we are not showing a return to growth yet. As discussed in the last quarter, the road towards growth will not be linear and we may face temporary setbacks on this path.

This said, we feel good about the transformation plan we have for the future. While it's too early to provide guidance for next year, it's still aspiration to return to double-digit growth in the second half of 2019. Our vision is to be the advertising platform for the Open Internet.

To us, the Open Internet is an environment enabling direct interactions with audiences and consumers, that also fair and transparent access to the time measurement. Despite representing 50% of time spent on line, it only captures about 30% of our expense. This means the Open Internet is still highly undermonetized.

This is because today, it does not benefit from the same level of technology than the world guidance enjoy. Using the data and technology to improve the monetization of this ecosystem, we seek to be the advertising platform of choice for the Open Internet.

The opportunity is huge, and we believe we are in a good position to be one of the leaders in this space. To do this, we leverage our unique commerce datasets, our Shopper Graph is among the largest global data sets focused on Shoppers. It has 3 main components.

The first is the Identity Graph, where we match well over 1 billion users across multiple devices. According to clients, our graph ranks in the top global 3 for exact match capabilities in most of the large countries where we operate. The second is the Interest Map.

An aggregated, anonymized view of the Shopper data, who are from a huge pool of participating clients. Thanks to this highly general product level browsing and cross retailer shopping data, we can support large range of scenarios for personalized advertising.

And the third is the Measurement Network, which helps brands connect advertising dollars to actual sales outcomes. We leverage first-party retail point-of-sale data, both digital and in-store to provide brands with sales distribution down to the product level. People often label us as a targeting company.

To be honest, I've never been completely comfortable with this label. Because it's a narrow oversimplified view of what we do. What we actually do is leverage high-quality, first-party data to provide value added advertising to clients. There are many ways to build attractive advertising scenarios on top of high-quality first-party data.

So far, we have mainly focused on conversion scenarios, which is readily to drive immediate sales for brand or retailer. In the past year, we have started building extensive capabilities to expand our offer to our performance scenarios. Leveraging our Shopper Graph and our technology, we offer new advertising scenarios, for what we call consideration.

This is the ability to engage with new audience to drive visits on a site, or installation of an app with not necessarily immediate link to sales. Market research shows that established consideration tactics are particularly effective at creating user engagement at scale. Our first consideration product had been Criteo Customer Acquisition or CCA.

In the coming quarters, we'll keep iterating and improve CSA, while also adding new consideration scenarios like app in-store, that I will discuss in more details later on the call. Beyond our effort to expand consideration, we are also planning to go even further at the funnel by adding [indiscernible] scenarios.

This is the ability to users to discover a new brand or product without necessary focusing on immediate interacting with the brand properties. So lots of exciting solutions to be ahead of us. Now to effectively scale upper funnel advertising scenario, we believe we need to excel in two areas.

First, we're implementing our Technology platform to be more flexible and modular. This in order to effectively combine assets into new advertising scenarios. For instance, our new identity architecture, run hash e-mails, allow us to offer a CRM onboarding tool that we call Criteo Audience Match or CAM.

As discussed earlier, our very large Identity Graph provides one of the best match rates in the industry. CAM also leverage all the important assets of our platform, like Shopper Data and Concept Management. Thanks to CAM. We offer a variety of flexible new advertising scenarios that a growing number of clients consume in modular way.

Another example is the big effort we are making to build self-service tools for our clients on top of our platform. Still scheduled for release in the first half of next year, we expect those tools to significantly increase our net client additions in the second half of 2019.

Self-service will also enable clients to entirely manage the advertising campaigns themselves, something very important for large digital native players.

Second, to scale upper funnel, we need also to ensure that our sales team is able to present and sell on multiple advertising scenarios, in a way, that is as intuitive and easy to understand as possible for clients. To achieve this, we must complete the ongoing realignment and training of our sales team.

This is a significant undertaking and has caused variability in our reported results. Given the size of our global sales organization, this is a large-scale project spanning several quarters. We are taking this very seriously and with the deliberate careful approach so we minimize negative impact to the business.

This also required us to restore our net hiring talents. As a matter of fact, exiting Q3, we are nearing the phase we need to meet our original hiring goals. Despite this, however, employee attrition increasing in Q3, leaving us a short of our net employee addition goals.

Recognizing that meeting such goals require combining strong high cadence with lower attrition, we have implemented a range of company-wide programs to remediate those drivers. We believe our ongoing actions will enable us to meet our net employment goals in the coming quarters.

Our anticipation is that in doing so, will be one of the factor helping restore net current additions as well as revenue growth. Overall, achieving success in those 2 areas, the flexible and modular technology platform, coupled with an efficient multi-product sales organization, is very much in our hands.

Now that we have set clear strategic goals and built plans to address them, success lies primarily in our ability to focus and deliver disciplined execution. As our transformations like this can come with a few bumps along the road, we feel good about our strategic plan.

As a matter of fact, we began to see early signs of success with some clients in Q3, and have an optimistic view of what lies ahead in the medium term. We are enforcing those 2 big focus areas with a number of important initiatives. Last June, we announced a major investment in artificial intelligence, specifically in deep learning technologies.

As you know, improving our prediction engine has been always the major driver of growth historically for Criteo, and we believe its next generation of neural metric technology could have a similar impact. During the summer, the Artificial Intelligence Research Team has been busy setting up the computation infrastructure for our new models.

In Q4, we will start our first test with real data. From then we expect to fine-tune our deep learning models and hopefully start to see early commercial impacts in 2019. We also significantly accelerating investment in the mobile-app ecosystem in Q4.

As you know, mobile app usage is growing quickly, very quickly and generating significant return on app-spend for our clients who invest in the right app experience for the customers. Because the lifespan value of an app customer is typically much higher than on the web, our clients are increasingly asking us for app-focused solutions.

As a result, when investing in this space, and this investment brings multiple benefits to us. First, apps increase a moat. Like that, delivering in-app advertising product at scale is hard to do, and requires significant R&D investments that many players are not equipped to make. As a result, our technology moat gets larger in the worlds of apps.

Second, apps offer a better monetization environment. Since our business has been built on delivering superior return on app-spend to our clients, the more time shoppers spend in well-designed apps the easier it becomes for us to deliver strong returns to our clients.

Last but not least, user identification is cleaner in apps than on the web as it does not rely on cookies, or any browser. Demonstrating a commitment to app-based solutions, we're delighted to announce that we have closed the acquisition of Manage on October 29. Manage is a Silicon Valley company with an attractive app in-store advertising solution.

The addition of Manage complements already significant app conversion business and allows us to move into upper funnel consideration solution in apps.

Beyond that, it also expands our client base, as Manage serves advertisers not only in a strong vertical or retail and travel, but also in gaming and other app first areas, such as food delivery and ridesharing. With Manage, we also gain additional technical and commercial talents to accelerate our expansion into this fast growing app space.

Manage is profitable with healthy EBITDA margins, as -- and this team has demonstrated disciplined execution. For this asset, we are paying just under 5x projected EBITDA for 2018. Speaking of acquisitions. We're pleased with the first steps of our integration of Storetail, the French startup we acquired in August.

Their platform nicely complements Criteo's sponsored products and Criteo's reseller program. We will now offer global monetization solution for retailers to activate their audience and traffic data to generate more revenues from brands and retailers, both on the site and across the Open Internet.

This broader monetization approach is receiving very encouraging feedbacks from retail clients, both in the U.S. and in Europe. Overall, as we've seen, returning to top line growth will be mainly about disciplined execution and focus.

And to illustrate the confidence we have in our plan for the business, we are also announcing today an $80 million buyback program. Given the current market conditions and the outlook for our business, we believe this program could create significant shareholder value in the future.

Before I turn the call over to Benoit, I wanted to point out that in contraction with our earnings release this morning, we posted a supplemental document on our website that answers common investor questions. With this, I turn the call to Benoit to discuss our performance in detail and guidelines..

Thank you, JB, and good morning, everyone. I will walk you through our Q3 performance and share our guidance for Q4 and 2018. Revenue was $529 million. Revenue ex-TAC, our key metric to monitor performance, decreased 2% at constant currency to $223 million.

This performance was driven by growth with large clients in the U.S., more than offset by software performance in the mid-market, primarily in the Americas and EMEA. Revenue ex-TAC includes store retails contribution of approximately $0.6 million.

Excluding the impact of Search e-mail on new project travel, that we have discontinued over the past year, revenue ex-TAC decreased slightly less than 1% at constant currency. Using the Forex assumption provided in our Q3 outlook, revenue ex-TAC was $225 million, slightly more than $2 million above the higher-end of our guidance.

Compared with Q3 2017, Forex changes were a tailwind of about 220 basis points to revenue ex-TAC gross. Overall, currency changes costed us $5 million versus last year or $2 million more than, what we had anticipated when we last provided guidance. Looking at our operating highlights.

We added 218 net new clients, roughly in line with our expectations, bringing total client count to over 19,000. We maintained retention at close to 90% for our Criteo marketing solution -- solutions, or customer acquisition [indiscernible] we are targeting.

In line with our prior view, we expect to see an improvement in net clients addition from the second half of 2019, following the release of our self-service tools earlier in the year. 10% of our live clients used at least two of our products compared with only 1% a year ago.

Revenue ex-TAC from Customer Acquisition, Audience Match, Sponsored Product, and Storetail combined, increased 82% at constant currency to over 7% of our total business. In that revenue ex-TAC grew 67%. Going forward, we expect to include Manage results in this category.

Same clients revenue ex-TAC decreased 5% compared with a decrease of 3% in Q2, primarily driven by user reach limitations in Safari.

And last, we continue to increase our direct access to publishers by further deploying Criteo Direct Bidder to now over 2,600 publishers, including Fox News, CBS, Marketplace, Two Dylan's, M6 and Ed Hunter, Point RU in Q3. Turning to the original performance. In the Americas, revenue ex-TAC was flat at constant currency.

Despite the impact of user reach imitation in Safari, and grew 1% in the U.S., driven by strength with large customers.

EMEA revenue ex-TAC decreased 5% at constant currency in line with expectations, driven by relative strengths with large clients, offset by the impact of user-reach imitation in Safari, softness in mid-market due to the shortfall in such capacity and approximately $4 million impact from GDPR, below our expectations.

We are seeing early positive signs from our self-source reorganization in the region and expect the full impact to take a few quarters to filter through results. In APAC, revenue ex-TAC decreased 2% at constant currency, in line with expectation, driven by strengths in Korea and India, offset by weakness in Japan.

Revenue ex-TAC margin improved 70 basis points to 42%. Similar to first half 2018, the large part of this improvement was driven by the increased share of mobile app supply, a large portion of which is purchased at a lower cost than expected. In line with plans, our margin level has started to normalize and we expect this trend to continue.

Moving to expenses, all the cost of revenue increased to 10%, driven by hosted cost and significant increase in third-party data to complement our Identity Graph, in line with expectations.

Operating expenses decreased 4%, reflecting a roughly flat headcount over the period and lower equity over compensation expense driven by foreign exchange, attrition and a lower share price over the period, and partially offset by higher bad debt expense.

Headcounts-related expenses represented 72% of GAAP OpEx, but slightly lower levels than in prior quarters. We ended Q3 with over 2,700 employees, up 1% year-over-year on 2% sequentially, including about 60 employees from store retail.

As JB discussed, we are taking active steps to improve our hiring cadence and reduce attrition with a specific focus on increasing our average sales rep tenure, across the company, over the coming quarters. On non-GAAP basis, operating expenses decreased 2% to $138 million.

And on a non-GAAP basis by function, R&D expenses decreased 2%, despite the 3% increase in head count to over 670 employees. We expect to grow R&D expenses in Q4 2018. Sales and operation OpEx decreased 1%, driven by a 1% decrease in headcount to close to 1,600, including over 700 quota carrying employees as well as higher bad debt provisions.

We expect our sales and operation headcount to stabilize in Q4. And G&A expenses decreased 2% despite a 6% increase in headcount to 490 employees, largely driven by adding new hiring capacity in the HR team. We did incur about $1 million of temporary savings compared to expectations, largely related to delayed hiring.

In line with what we said last quarter, we do not expect to entirely catch up to initial hiring plans by the end of 2018, but instead hope to be back on track in 2019. On the other hand, we incurred approximately $3 million of higher bad debt provisions than expected, mostly on sales. Moving to profitability.

Adjusted EBITDA declined 11% at constant currency to $70 million with adjusted EBITDA margin declining 260 basis points to 31% of Revenue ex-TAC. Normalizing for temporary savings, the expected negative contribution of store retail on the higher bad debt provision, adjusted EBITDA margin was over 32%, down only 130 basis points.

The reported effective tax rate was 34%, based on our updated projected tax rate of 32% for 2018. The decrease from 37% last quarter results from updated assumption to mitigate the impact of the dip tax in the U.S. As you may recall, we indicated last quarter that our projected tax rates for the year could go down by 3 to 5 percentage points.

Net income decreased 19% to $18 million, driven by a 22% decrease in income from operations, lower financial expenses by 65%, and 13% increase from provision from income taxes. As a result, adjusted net income per diluted share decreased 18% to $0.53.

Cash flow from operations decreased 19% to $50 million due to a significant increase in income tax paid driven by the mid-tax in the U.S. and the cash payment of 2017 tax in Germany and France. This momentarily drove transformation of adjusted EBITDA into cash flow from operations to 72% compared to a normalized 80%.

CapEx increased 7% at 6% of revenue in the quarter, driven by data center equipment on capitalization of development cost for internal projects. We continue to expect 2018 CapEx to represent approximately 5% of revenue. As a result, free cash flow decreased 39% to $21 million.

In the first 9 months, free cash flow increased 14% to $95 million, representing 44% of adjusted EBITDA, a few points above historical average.

Finally, cash and cash equivalents increased $45 million from the end of 2017 to $459 million, despite a $44 million net cash outflow for the store retail acquisition and an $18 million negative Forex impact on the cash position over the period. I will now provide our guidance for the fourth quarter and fiscal year 2018.

The following forward-looking statements include contributions from Manage and reflect our expectation as of today, October 31, 2018. In Q4, 2018, we expect revenue ex-TAC to be between $256 million and $262 million on a reported basis.

This implies constant-currency growth of minus 6% to minus 4% and assumes a contribution of approximately $3 million from Manage. We expect year-over-year Forex changes to be a headwind to reported gross of about 130 basis points.

Using FX assumptions underlining Q3 the 2018 guidance, our Q4 guidance would be $261 million to $267 million, implying a Forex headwind of approximately $5 million. This year, we are providing a slightly broader guidance range for Q4 to account for the greater variability of our business, in particular around the U.S. holiday season.

Our guidance is within the full year range, when factoring in year-to-date result, that represents a decline of approximately $4 million at the midpoint compared to our prior implied guidance for Q4 at constant scope and Forex. This mild reduction can be attributed to sales execution.

As part of our go-to-market transformation, we are reallocating sale resources across client accounts, resulting in short-term decreases in such productivity. We believe this setback is temporary and we feel good about the direction of our transformation plan.

With regards to the full year 2018, we are reiterating our revenue ex-TAC growth guidance of minus 1% to plus 1% at constant currency. Compared to 2017, we see Forex changes bringing about 110 basis points of reported growth, or approximately $10 million. However, using our assumption for the Q3 guidance, Forex would be a $5 million headwind.

On the profitability side, we expect Q4 2018 adjusted EBITDA of between $86 million and $92 million, including approximately $1 million for Manage. Using the Forex assumption underlying the Q3 2018 guidance, our Q4 guidance would be $88 million to $94 million.

For the full year 2018, we are maintaining our full year 2018 adjusted EBITDA margin guidance of between 30% and 32% of Revenue ex-TAC. As usual, our FX assumption reporting guidance for the first quarter on fiscal 2018 are including in the earnings release published earlier today.

Before closing, I want to point out that in addition to the FAQs that JB mentioned, we have added to our Investor Deck in our -- we have added an Investor Deck to our website, some supplemental information on the impact of that currency has on our actual result on guidance.

While we are currently working on building our 2019 budget, it is too early to provide guidance for 2019. Our aspiration remains to return to double-digit growth in the second half of 2019. We will provide guidance on 2019 when we report Q4 earnings. With that, we'll now take your questions..

[Operator Instructions]. And the first question comes from Tim Nollen with Macquarie..

Can I ask two questions please? First on the sales force hiring.

If you could please elaborate a bit more on what the issues are? Is it a matter of finding the right people is it difficult to find people to train? Is it about ability to incentivize them, et cetera, a bit more and what the issues are there? And secondly, appreciate some progress on the acquisitions and, I think, several investors will be happy with the buyback? I just wonder you still have $400 million in cash and no debt.

Could you talk a little bit more about future plans for use of cash please?.

So regarding sales force hiring, this is JB, there are different things. First, you need to hire people who are doing the hiring and that's immediate, because we use, we rely mostly on internal staff to hire those people. So we've been repopulated this team, where we had some attrition in the beginning of the year.

So now we have a full-fledged hiring team and we are going full speed into hiring people. You need to find the right people and you're absolutely right we need also to train them. And it takes typically 6 months for someone to be completely operational.

Probably a bit longer now as we're asking them to sell more complex product suites, rather than just the one product thing that we were selling the years before. So it makes the training a bit more sophisticated. But we are confident, we are back on track now..

With respect to the use of cash, as you noted the announcement today, so we are planning to dedicate $80 million of our capital for share buyback. For the rest of the capital allocation, we continue to prioritize allocation of capital beyond this share buyback program to M&A..

And the next question comes from Brian Weiser with Pivotal Research..

Just two please. Could you tell us about the degree which you think that ITP chip now is expected to have any incremental impact or not at all in your current expectations? And relatedly, you mentioned that large advertisers drove the U.S. results, which leads me to ponder.

As agencies are launching people-based marketing platforms increasingly and pushing it to their clients, does that represent an opportunity or a threat to maybe work more closely with those platforms of your business?.

Sure. Thanks very much for the two questions. So regarding ITP, as you know, most of the effect of ITP came in Q4 last year, where we got the full effect of ITP beginning of 2018. So I would say the effect is really reflected in our numbers and all our projections are taking this into account. So I think we are very much in steady mode..

Yes, with respect to the ITP 2.0, which was introduced in September, we don't expect any incremental impact from this new release of ITP, I mean, we saw the full impact in the initial ratings..

One, when we go up a funnel additionally, agency has more budgets there. So it's going to help us in this capacity. And also, as we are expanding the monetization space to our brands, it's also an area where our collaboration with agency will be very important to the success.

So absolutely, this is part of the key initiatives we are taking to expand our reach..

And the next question comes from Matthew Thorton with SunTrust..

A couple if I could, I guess.

Number one, Benoit, what take rate is assumed in the 4Q revenue guidance? Secondly, Firefox just kind of curious, what you are baking in for incremental impact from some of the privacy settings and changes that they have made there? And finally, thirdly, on Manage, I think you talked about the valuation of that acquisition, can you talk to the size, I'm assuming, it's an all-cash deal, but can you talk to the size of that acquisition that we should expect to hit the balance sheet in 4Q?.

Okay, so with respect to the take rate. We don't expect any significant swing in tax rate in Q4 compared to what we've seen in Q3. So Q3 was at 42%, so we expect to be pretty close to that take rate. With respect to Firefox, I think, we have commented in the past on Firefox. I mean, Firefox represent a small portion of our revenues..

Specifically, it's less than 1% of our revenues in the mobile space, yes,.

Especially in mobile, where it's very small. So Firefox is baked into our plans, I mean, we don't expect any impact in Q4 and obviously, we'll factor that in our '19 plan. With respect to Manage, we have not disclosed the price of the acquisition. It's a multiple. It was a pretty attractive multiple.

We bought the business on an approximately 5x multiple on EBITDA. It's a business running on this EBITDA margin. And we have no plan to disclose more at this stage..

But I guess with the contribution of Manage in Q4, you can have a sense of the size of the business..

And the next question comes from Sarah Simon with Berenberg..

I've got three questions please.

First one was JB, can you just give us an update on where ePrivacy is going versus when you commented at Q2? Second was on Japan, as to whether the difficulties you've been seeing there had improved in Q3? And then the third one was on your mobile in-app business, you reported 68% gross, and I think, I'm right in saying, it was 38% at the half year stage.

So what prompted that massive acceleration or am I just not comparing the right numbers?.

So, let me take the first one on ePrivacy. So ePrivacy is still in discussion. So I think it's too early to drive any conclusion there. There are different scenarios. As you know, next year, it's going to be an election year in Europe. So this might derail the project.

We expect, generally speaking, ePrivacy to be roughly in line and consistent with what's happening -- what's been happening for GDPR. Because at the end of the day, the spirit of ePrivacy is to be the application of GDPR specifically for the digital space.

It's still that the project is not finalized, it's -- we cannot give definitive answer on one side. And we don't even know if it's going to be for next year or the year after. Once more, with the election happening in Europe next year, sometime it create delays in those projects..

So with respect to the in-app figures, so the 67% gross in-app is on the pure in-app scope. Pure in-app scope. You know that we have within our in-app business we have web-to-app as well as in-app.

Now that we have, with the acquisition of Manage and going forward, we want to measure the success of the in-app business purely on the in-app scope without accounting for the web-to-app. And that's the main reason for....

And coming back to your second question on Japan end market. So this is part of our full cruise to reaccelerate our hiring. I was in Japan twice in the last 3 months and the management team is really focused on bringing back the market team on track. And I think the early signs are encouraging.

And we still cautious in Q4, which is not particularly important seasonal quarter for Japan as it is in the U.S. Q1, in a way, is even more important in Japan. But we are working hard to bringing this back on track..

The next question comes from Andy Hargreaves with KeyBanc..

Just want to ask on pricing in the last quarter you guys had mentioned potential....

Can you speak up because it's very difficult to hear you..

Can you hear me now?.

Yes, much better..

I was asking around pricing, you mentioned potential changes to pricing towards the CPM model last quarter and wondering, if you've defined that strategy more clearly and if there's any initial learnings from it?.

Absolutely. So that's one of our major initiative as we go into, what we call consideration, which is the idea that we are not -- we are focusing on driving visits to websites and not necessarily immediately sales.

Because when you are interacting with a user, who's never been in contact with your brand, it's sometimes a bit and really seek to expect the user is going to converge directly into sales. And usually, the first step is to bring this user to the website and start to a first visit, which is a whole consideration model.

This is and for this model, we are promoting the CPM model, which happened to be the startup of the industry, whenever as Google or Facebook for this type of consideration scenario. They are also charging on CPM. So we thought that it's much more efficient to be a consistent with the rest of the industry. It's still early.

We need to train our sales team. It's a new type of pitch. And when I was mentioning that we had this big confirmation into our sales organization, this is included because for sales people, who have been selling only clicks their whole life, this is a more sophisticated pitch.

And the early signs, we've been focusing on our champions most advanced sales people that are creating the first case studies for this. And as we going to see more and more success in those case studies, this will create emulation for the rest of the sales organization.

So long story short, we are still early in this process, that it's one of our key initiative and key area to focus to expand this consideration space and scale our Criteo Customer Acquisition product..

And then on the sales force, specifically in Europe, are the changes to compensation and account packages now completed and it's just about getting people up to speed? Or are there still more changes to come?.

It's mostly around train -- hiring and training. I think one more -- once more, it's about the ability to sell a different pitch and some time to a different people in a client organization. People in charge of a funnel are, they tend to be the same people for small clients, but for large clients it can be different teams.

So you also have to create those new relationships. So in terms of the way the composition is structured, there's no major difference.

It's more about sales pitch and making sure you create the right expectation for the clients because if -- if you're doing an upper funnel campaign, and you're promising your clients, you're going to deliver immediate sales, then you are setting the stage for disappointment.

So it's very important that we set the right expectation from the start, when we pitch to our clients of what exactly they should expect from which type of marketing scenario..

And the next question comes from Mark May with Citi..

Hi this is Nick on for Mark. Just a quick question on GDPR.

How is that impacting the business today compared to last quarter? Do you think it's mostly implemented by advertisers and publishers? Or is there any update you can provide there?.

So we provide an update in terms of quantitative impact, which is smaller than what we expected. In a qualitative way how does it materialize in real life? Is that -- there are different ways a client can implement GDPR guidelines.

There is the standard way, where we -- where the client is asking for consent in a way which is not intrusive in terms of the user browsing. And you have the small minority of players, who are implementing content management in a way that we believe it's not the spirit of the law and it's way too rigid.

And where pretty much the whole site is frozen before you give your consent, which can be very annoying for the end users. It's a small minority of sites doing this. But when they do this, they technically limit the traffic on their website and this impact the business relationship we're having with them.

We believe those outliers are going to be smaller and smaller because they're going to quickly realize that one, it's not what is -- what the privacy agencies are asking for, it's not the spirit of the law. And it's detrimental to their overall business.

But we're still in the transition phase, where the market is adjusting to this new state of art of GDPR..

And the next question comes from Mark Kelley with Nomura..

Two quick ones good to hear the self-serve launch is on track for the first half of next year. How do we think about operating margins as that starts to layer into the business in 2020 and beyond? I'm assuming that will be margin accretive but I want to get your thoughts there.

And then second, can you remind us the seasonality of CSP in 4Q? And I know the travel business is small, but backing that out, what should we expect for same-store sales?.

So with respect to the self-serve solution, yes, we share your enthusiasm and the fact that we still see a -- see the release in the first half. What, I mean, clearly for us, the agenda here is to be more effective in the way we gain client addition, especially in the smaller part of the mid-market.

So the first agenda is client addition agenda, and that should contribute to restore more active figures in terms of net addition as well as to contribute to the gross.

There's a benefit of you see in terms of margin because the cost of growing and serving this part of the market, is going to be economically viable, while if we were to do that by hiring sales people, it would be extremely difficult to find the right number of sales people and would come at margins that would be far less attractive.

So in order to maintain our margin profile, self-serve is going to be a critical part of this -- in this market. With respect to the seasonality of CSP, seasonality of CSP has been traditionally pretty high in Q4 around 50% of the business of the year. I mean, we would not expect any significant change in profile of seasonality this year.

Obviously, there is one aspect that you need to be aware that we discuss is that there is a transition going on in terms of market demand around a sponsor product, where you -- we see large clients, a large client working increasingly directly with brands on that's the reason why we are adapting our offering in order to be able to support those large clients with more, much more an offer, which is to sell our technology to them on the subscription basis, in addition to the traditional offering that we have in CSP..

And the next question comes from Brian Fitzgerald with Jefferies..

Maybe a follow-up to the GDPR question. You highlight the strength of the shopper graph and that's made of 3 components; the Identity Graph, the interest graft and the measurement system or network or graph.

Is any -- in your view, is any one of those 3 components more or less impacted by GDPR? And then the second one I wanted to ask was around Facebook.

With its closing down its partner categories and tightening the standards around custom audiences, have you seen any impact either positive or negative with respect to your ability to target inside their network?.

So when it comes to the shopper graph the impact from GDPR has been marginal. As I said, the impact of GDPR is focused on small minority of specific partners who implement GDPR in a way which is not consumer friendly, and which tend to defer traffic to other websites.

So it's more an ability to buy traffic from those specific publishers that were impacted by GDPR. In the grand scheme of things, it's a fairly -- fairly limited impact. Facebook, so just to clarify, Facebook for us is a media we buy from. It's not a source of data. We don't use Facebook data to target users. We use the first-party data from our clients.

In terms of media, the technical setting we are having with Facebook is not as efficient as what we have with most web properties. As a result, our engine, who's agnostic in terms of source of inventory, tend to allocate less to Facebook than to other source of traffic.

And you can see as a result that the shared voice of Facebook has been declining quarter after quarter. And now represents only 3% of our Revenue ex-TAC in Q3..

And the next question comes from Dan Salmond with BMO Capital Markets..

Just a few questions. Maybe first JB, just to return to some of the new upper funnel and consideration level type of products you're looking at. First, you mentioned that the focus on CPM there.

Previously, you had also mentioned potentially looking at a subscription model for some products just be interested to hear about that? And then just across this potential sort of new step down the product roadmap.

When you look at these products, do you see Criteo being able to find a model where your economics are comparable to the core, we won't call it retargeting, but retargeting business, some of the newer products lately obviously have had a little lower economics and I am just curious to see if our expectations should be along those lines as well.

And then just quickly on the buyback, just to return to it. It sounds [indiscernible] I'd love to hear just a little bit more on why the choice now? The stock price has been volatile but mostly in a range for the last year.

Just more really about using a surplus of cash and then, similarly to the comment earlier about uses of free cash flow, it sounds very clear that acquisitions are still to focus in building the business more broadly.

So should we look at this as a one-time type of buyback as opposed to a renewal, or an authorization that may be renewed regularly?.

Okay, so on the upper funnel. So as we mentioned given the nature of the marketing scenario, we are addressing across point-pressure model, it's probably more consistent than cost-per-click. Because cost-per-click suggests that the value is only in the click, which make total sense when you're doing your targeting.

When you are doing upper funnel and you are going after users, who have never interacted with your website, the impressions has value even if there's no click. Especially, as you move into a format that are more to their like, where people tend not to click on videos, which doesn't mean the video is not effective.

Video can be very effective even when there is no click. This is why the CPM model works very fine. Regarding the related margins. It's about how much value and return on app-spend you create for your clients.

And if we do good work, in this upper funnel, there's no reason why we could not have a healthy margin in upper funnel, especially as we are leveraging something, which is very unique in the market, which is this first partition of data.

I think, this is something no one else can really offer at this scale that we offer right now, which give us a very, very compelling value proposition for our clients. Because they don't -- they cannot find the source of data anyone else than with us. So I think, this put us in a good position..

Regarding the buyback, so why now? I mean, we've always said that, in the past, that our answer on the buyback was not necessary final that we would be open. And with, I think it's an opportunistic timing. We feel good about our transformation plan. We see the share price at pretty low at the moment.

That's the right timing, we believe to launch such a program. Does it mean that we are committed to renew such program in the future? No, I don't think you should assume such a commitment at this point. It's an opportunistic program to signal that we are confident in our transformation..

Thank you, Dan, thanks, everyone. This now concludes our call. We'd like to thank everyone for attending today. The IR team is available for any follow-ups you may have. Good bye everyone, and enjoy the rest of your day. Thank you..

Thank you, everyone..

Thank you, everyone..

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