Release Date: 19 October 2018

Criteo S.A.
Click Here to Buy

Ticker: CRTO.US • We believe Gotham’s short thesis on

Criteo is highly flawed and misleading.
Market Cap: US$3.1 billion GCR seems to have a tenuous grasp of how
internet traffic is distributed and lacks a
Recent OTC Price: US$47.20 practical understanding of how the digital
Target Price: US$64.00 advertising industry operates.
• According to industry insiders and clients
Expected Return: 36% themselves, the key complaints that GCR
levels against Criteo are either baseless or
Opinion: Strong Buy inconsequential. Clients see Criteo as a
valued advertising channel.
• These sources also explained that the GCR
reports have had no impact on clients,
whatsoever.

You should have expected us

aainfo [@] neomailbox.ch

Twitter: @anonanalytics
www.anonanalytics.com
 Disclaimer
Neither Anonymous Analytics nor its principals is a registered investment advisor or otherwise licensed in any jurisdiction,
and the opinions expressed herein should not be construed as investment advice. This report expresses our opinions,
which we have based upon publicly available facts and evidence collected and analyzed including our understanding of
representations made by the managements of the companies we analyze, all of which we set out in our research reports
to support our opinions, all of which we set out herein. We conducted basic research based on public information in a
manner than any person could have done if they had been interested in doing so. You can publicly access any piece of
evidence cited in this report.

All facts, figures, and opinions are as at the last practicable date. This document has been prepared for informational
purposes only. This document is not an offer, or the solicitation of an offer, to buy or sell a security or enter into any other
agreement. We have made every effort to ensure that all information contained herein that support our opinions is
accurate and reliable, and has been obtained from public sources we believe to be accurate and reliable, and who are not
insiders or connected persons of the stock or company covered herein or who may otherwise owe any fiduciary duty to
the issuer. However, we do not represent that it is accurate or complete and should not be relied on as such, in particular,
Criteo S.A. (“Criteo” or “the Company”) and insiders, agents, and legal representatives of Criteo and other entities
mentioned herein may be in possession of material non-public information that may be relevant to the matters discussed
herein. Do not presume that any person or company mentioned herein has reviewed our report prior to its publication.

As evident by the contents of our research and analysis, we expend considerable time and effort to ensure that our
research analysis and written materials are complete and accurate, we strive for accuracy and completeness to support
our opinions, and we have a good-faith belief in everything we write - but such information is presented “as is,” without
warranty of any kind, whether express or implied. All expressions of opinion are subject to change without notice, and we
make no representation, express or implied, as to the accuracy, timeliness, or completeness of any such opinions and
information or with regard to the results to be obtained from its use, and we make no representation that we will update
any information on this. You should assume that all statements contained herein are our opinion and are not statements
of fact – even if certain statements can be perceived as such. That way, we don’t have to sacrifice our (hopefully)
entertaining writing style by starting every sentence with “In our opinion” as advised by our team of neurotic and
overpriced lawyers.

We believe that the publication of our opinions and the underlying facts about the public companies we research is in the
public interest, and that publication is justified due to the fact that public investors and the market are connected in a
common interest in the true value and share price of the public companies we research. We are exercising our right to
express such opinions in a public forum. Any investment involves substantial risks, including complete loss of capital. Any
forecasts or estimates are for illustrative purpose only and should not be taken as limitations of the maximum possible
loss or gain. Any information contained in this report may include forward-looking statements, expectations, and
projections. You should assume that these types of statements, expectations, and projections may turn out to be incorrect.

Anonymous Analytics itself holds no direct or indirect interest or position in any of the securities profiled in this report.
However, you should assume that certain of Anonymous Analytics’ research and due diligence contacts, consultants,
affiliates, and/or clients may have a long position in the stock or debt of AAC and/or options of the stock, and therefore
stand to gain substantially in the event that the price of the stock increases. You should further assume that following the
distribution of this report, the aforementioned individuals and entities may continue transacting in the securities covered
therein, and may be long, short or neutral at any time hereafter regardless of this report’s initial opinions.

Don’t be stupid and invest in the public markets unless you are prepared to do your own homework and due diligence.

1
 Introduction
We believe the Gotham City Research short thesis on Criteo is highly flawed and misleading.

Criteo is a US-listed company that specializes in digital performance marketing and targeted advertising.1
Criteo’s primary product is dynamic retargeting. For example, if you go to an e-commerce website and
browse around but leave without making a purchase, you may see ads on subsequent sites showing you
the products you were just looking at in the hopes of driving you back to complete the sale.

GCR has released a series of three reports critical of Criteo over the last few weeks, each increasingly more
unhinged in its rhetoric. Yet, the stock is trading higher now than it was prior to the reports.

The muted response by the market is not particularly surprising. Judging by their analysis, we believe GCR
has a tenuous grasp of how internet traffic is distributed and lacks a practical understanding of how the
digital advertising industry operates.

To believe GCR’s short thesis is to believe that they know more than the sophisticated customers that
actually use – and are directly plugged into – Criteo’s service. Criteo has a decade-long operating history
and a 90% client retention rate. The idea that there is fraud in the adtech industry, that there are issues
with click misattribution, and that somehow Criteo’s clients – most of which are top enterprise retailers –
have no idea exactly what’s going on or have not found workaround solutions is next-level patronizing.
We can imagine GCR’s thinking as they were writing up their research:

“Check out all this stuff we found on the internet! If we write this up in a report we can
finally show all these Fortune 500 companies that use Criteo and constantly monitor their
advertising ROI what’s really going on! Short seller with minimal understanding and
experience in digital advertising to the rescue!”

The contents of the GCR reports may be news to some investors, but it’s a decades old issue to anyone in
the industry. We know this because we called multiple Criteo clients and industry experts to better
understand the industry, and learn if there has been any fallout from the GCR reports. Not a single client
or industry insider we reached out to actually cared about the reports:

“This has been a topic in the industry for a long time, and so there’s not anything new
that’s facing Criteo in that respect. There was that Gotham Research report or whatever
that seem to have caught the eye of the analyst community but haven’t really impacted
the actual customer base, whatsoever. This isn’t a topic that you hear amongst Criteo
customers. A lot of those articles that have come out recently are for people in the industry
like ‘well, welcome to 2003’. These are topics that have been at play for a while.”

- Major Criteo competitor

1
https://www.sec.gov/Archives/edgar/data/1576427/000157642717000021/criteo10-
k2016.htm#s37214477C32FDF7423CA7ADC22B50BB3 pg. 1

2
 “We’ve seen all this hedge fund activity around Criteo and so we basically wondered if we
had missed something. We looked into this and we didn’t find any practices or any fishy
things that would lead us to believe that there is a significant amount of fraud behind this
business… So, in a nutshell, do I have any reason to believe today and after our own
discussions that Criteo is cheating? The answer is no.”
- Major e-commerce client

“Criteo customers are not reading these reports [Gotham City Research] that you
mentioned in a large way, so the competitive dynamics have nothing to do with click fraud
or not syncing with Shopify data.”
- Major Criteo competitor

“No. Not at all. If my costs of sales would increase alarmingly, then we are going to have
a conversation, but through years of 30% year-on-year increase in investment into the
Criteo channel they’ve performed and really haven’t missed a beat.”

- Fortune 500 client

If anything, the GCR reports almost read like paid advertising for retargeting companies like Criteo,
masquerading as a parody of a short thesis. Advertising has always been sort of a pseudo-science. It’s not
always easy to determine advertising ROI metrics with any sort of scientific precision, particularly if a
company is running multiple advertising channels. But for the most part, there is more accountability and
transparency from companies like Criteo which are closer to the end-point of a sale transactions than
companies that primarily deal in softer impression advertising.

GCR is correct that there is fraud in the digital advertising industry. No one disputes that. But most of the
fraud doesn’t come from ad retargeting where Criteo derives nearly all revenue. It comes from banner
advertising and branding. If anything, GCR’s reports show why advertisers are more likely to flock to Criteo
and similar companies focused on performance-based metrics.

“There’s kind of two worlds in display ads – there’s brand ads and there’s performance
ads. So, brand ads would be like ‘I’m Kellogg’s and I just want to run a bunch of display
ads.’ And the metrics that they’re looking at is ‘how many clicks or impressions did I get?’
And for that the fraud actually matters. On the performance side, we’re focused more on
output metric so like on e-commerce sites how many times did someone actually purchase
our products? The fraud becomes less of an issue because I don’t care if there are fake
clicks, because a click is not a metric I look at. I just care about what’s my cost per booking?
What’s my ROI? In theory it shouldn’t affect anything because even if I pay for it, it’s still
baked into my bidding model.”
- Major online booking company

3
 “Fraud in general and more broadly is an issue for digital advertising overall – that’s
absolutely true. But once you look at that from a relative standpoint, performance-focused
players like Criteo are a little bit more insulated from a lot of that fraud issue than some
of the other players that are focused on brand dollars and softer metrics. Guys that are
paying $10, $15, $20 CPM for branded video ads, that’s where fraud is the real problem
and actually that’s where most of the fraud actually happens.”

- Major Criteo competitor

We think the reason that Criteo gets more than its fair share of bad press is because it is the biggest stand-
alone public company in this space and therefore the most visible. Likewise, short sellers can’t go after a
private company, or adtech companies that are subsidiaries of Google or FB, because there is no
opportunity to short those companies. Instead they are reduced to awkwardly tailoring their short thesis
to Criteo, which actually focuses on a segment of digital advertising where fraud is less of an issue.

In any case, this report will look at the key allegations GCR has made in their reports and provide a
systematic breakdown of why we believe GCR’s thesis is flawed. Specifically, we will focus on the following
points from the first two GCR reports:

Source: GCR 1st report, pg. 5 & 2nd report, pg. 4

We will also discuss Apple’s ITP, which is the focus of GCR’s third report, but presented in an over-the-top
dystopian fashion.

4
For what it’s worth, this is not our first disagreement with GCR. Earlier this summer GCR released a report
suggesting fraud at AAC Tech, a multi-billion dollar Apple supplier listed in Hong Kong. That report was
laced with obvious factual errors and inaccuracies which led us to respond with our own analysis.
Following our report, shares of AAC have nearly doubled and anyone who relied on GCR’s research as a
reason to short AAC probably lost a lot of money. We don’t imagine it will be much different with Criteo.

1-Year Share Price Performance – AAT Tech (2018.HK)

GCR suggests fraud

We respond

Source: Google Finance

5
 Criteo Websites
Key GCR claims:2

• Over 50% of Criteo’s revenues originate from suspect sources (e.g. clickbots, fake/low quality
websites, etc).
• 49.9%-56.2% of Criteo ads seem to be displayed on suspect websites, per our analysis of 513,000+
websites.
• 28.7% of Criteos peers’ ads seem to be displayed on suspect websites, implying that Criteo is twice
as bad [as peers].

*****

GCR’s thesis centers on the allegation that over 54% of websites using Criteo are suspect, as per the
following screen grab:

Source: GCR 2nd report, pg. 6

The way GCR appears to have gotten their results is by using the BuiltWith tool to scan a universe of
websites and filtering them using specific markers, or lack thereof. For example, you can choose to filter
a data set that has an email address listed or twitter/FB buttons. The sites that don’t have either of these
properties one *may* conclude are possibly dubious because they lack contact information or social
media presence.

Using similar methodology, GCR concludes that over 54% of the sites using Criteo fall in the suspect
category because they are missing a certain list of markers. This analysis is novel, but doesn’t actually
provide any economic substance. It’s more akin to the type of analysis a first-year MBA student would
conduct to show their professor how creative they are than an analysis that has any sort of real relevance
to the discussion.

2
GCR 2nd report, pg. 4

6
The problem with GCR’s analysis is that it assumes all websites are equal-weighted. For example, it
assumes that traffic to CNN.com is equal to an obscure site like toastedballs.com, and therefore the ad
revenue from both sources is equal. Here is the actual assumption from the ‘Methodology’ section of
GCR’s report:

Source: GCR 2nd report, pg. 17 – note that this methodology is not ‘conservative’ as GCR claims, it’s just wrong.

But in fact, total internet traffic is heavily skewed to just a handful of sites. The vast majority of us visit the
same 100 or so sites, and those sites account for nearly all the traffic on the internet. To illustrate, we
used Quantcast data to build this graph that shows the number of visitors to the top 250th ranked US sites:

Monthly Traffic to top 250 US Sites

250,000,000

200,000,000

150,000,000

100,000,000

50,000,000

0
134
1
8
15
22
29
36
43
50
57
64
71
78
85
92
99
106
113
120
127

141
148
155
162
169
176
183
190
197
204
211
218
225
232
239
246
Source: https://www.quantcast.com/top-sites

We can see that by the time we get to the 250th ranked site, the number of visitors is less than 3% of the
1st ranked site. By the time you get to the fringes of the internet that GCR’s analysis would suggest, the
50,000th ranked site gets almost no visitors, and therefore almost no advertising clicks.

As the biggest stand-alone player in the industry, it’s not surprising that Criteo’s ads are found deep into
the internet. But almost no one actually visits the 100,000th ranked site. Therefore, almost all of Criteo’s
advertising audience is from the top few hundred or so websites such as CNN, USAToday and Tripadvisor.

7
GCR’s assumption that because over half of the websites that use Criteo are suspect and therefore half of
Criteo’s revenue is suspect is highly flawed and shows a basic lack of understanding of how internet traffic
is distributed.

No adtech company would survive for as long as Criteo has and maintain a 90% client retention rate if half
of the clicks they got were from suspect sites. GCR’s analysis is based on an assumption that doesn’t even
pass the common-sense test.

Criteo vs Peers
Based on their flawed assumption, GCR also concludes that because over 54% of Criteo ads are found on
suspect websites, Criteo is twice as bad as peers which average around 28%, as per the following
screengrab:

Source: GCR 2nd report, pg. 7

Again, this analysis is flawed because it’s not about how many ‘suspect’ sites Criteo has ads on vs peers
– it’s about where the views are coming from. The vast majority of Criteo’s ads are viewed on reputable
sites because internet traffic is heavily skewed towards those sites.

In fact, we contacted one major competitor of Criteo to ask about the claim that Criteo is twice as bad as
peers. Here is what they had to say:

“I’d love to say yes, but there’s nothing we are doing specifically that Criteo wouldn’t be…
if anything, we don’t maintain our own direct publisher network and as far as I know Criteo
still executes 60-70% of their campaigns on their own publisher network which are the
most premium sites.”

-Major Criteo competitor

8
 Client Ad Spend
GCR continues to build on their flawed assumption by concluding that Criteo clients can cut spending by
50%+ and not hurt sales:

Seriously?! - You really believe

that Chase Bank, RH, and Uber
reduced their digital ad
spending by 99-100%?

Do you realize that would

nd
Source: GCR 2 report, pg. 8 mean they have NO online ad
presence anymore?

Most of these claims are either comically false or grossly misleading.

Here, GCR continues to demonstrate their tenuous understanding of how internet traffic works. The linear
conclusion that 50%+ of Criteo ads are on suspect websites and therefore clients can cut 50%+ of their
advertising spend completely ignores the fact that most internet traffic, and therefore most advertising
dollars, flow through a few hundred sites. These sites are reputable sites and have real audiences. Ted
Turner is not sitting in his office scheming of ways he can use botnets to rip-off advertisers.

There is something absurd about GCR thinking that Criteo clients are not constantly monitoring and
optimizing their return on ad spend, and were just waiting for a confused short seller to swoop in and
save them.
Wake up,
Sheeple!

9
Furthermore, as some sort of apparent confirmation of their claim, GCR points to five companies that
have purportedly cut their digital ad spending by 50+% without losing sales. Let’s take a look at each:

Chase Bank

We’re going to start with Chase Bank because we think it perfectly captures everything wrong with GCR’s
thesis. We spent a lot of time wracking our brain trying to figure out from where GCR got their claim that
Chase Bank reduced ad spending by 99%. Then we realized GCR keeps making the same mistake.

The only source GCR cites for their claim is this NYT article which actually makes no mention, and provides
no dollar value, of Chase Bank’s digital advertising spending. Instead, here is what the article actually says:

This screengrab states that JPMorgan Chase limited the number of websites that show their ads from
400,000 to 5,000 with little change in the cost of impressions or visibility of its ads on the internet.

We think the way GCR calculated their 99% reduction in ad spending is by calculating the 5,000 sites from
the original 400,000 sites. But this is NOT a reflection of ad spending, this is a reflection of where internet
traffic is going. If anything, this proves our point about the most frequently visited sites accounting for
nearly all internet traffic.

The reason that JPMorgan Chase did not see a decline in online visibility is because the vast majority of
internet traffic goes to the handful of sites that sit within the 5,000 figure that JPMorgan whitelisted.
Almost no one is visiting the majority of those other 400,000 sites.

This does not mean that adspend was reduced by 99% proportionately. This might mean that adspend
was reduced probably more to the tune of 1-2% because most of those original 5,000 sites account for
nearly all traffic and therefore nearly all advertising dollars. If anything, it is the publishers of those
remaining 400,000 sites which rely on CPM that are affected, not Criteo or its advertisers.

10
P&G and Unilever

GCR is correct that P&G and Unilever cut their digital advertising budgets by 41% and 59% this year due
to brand safety concerns and ineffective ads.3 But the overarching issue here is the vast size of their
budgets. P&G and Unilever are the two biggest advertisers in the world.4 We estimate that these two
companies spent a combined US$1.7 billion just on digital advertising in 20165. To put this sum in context,
it is approximately equal to Criteo’s entire 2016 revenue from its ~14,500 clients.6

Rather, P&G and Unilever’s spending reduction seems to be a symptom of the law of large numbers and
a shotgun approach to banner advertising. At a certain point, each incremental dollar spent will have less
of a return than the previous dollar until there is no discernible return on investment. This is especially
true when advertising dollars are spent on banner ads and impressions which have less transparency than
retargeting (it’s worth reiterating that almost all of Criteo’s revenue comes from ad retargeting).
According to one industry insider:

“so much of [P&G’s] investment was put into more straightforward CPM banner ads and
I think that’s where a lion’s share of this fraud is happening. They’ve got these huge
budgets so these agencies will place their dollars into the far reaches of the web just to
gain the impressions that they wanted. I can see how it’s not effective. Banner ads and
straight forward display advertising – not retargeting – but serving up banners as kind of
a brand awareness type of campaign was incredibly inefficient for me and I pulled almost
all of it back.”

- Fortune 500 client

“We tried at some point for one month to do no retargeting at all, and we tried for a couple
of days and we just saw the business went down. That was another indication for me that
this creates sales. This creates value.”

- Major E-commerce client

3
http://www.businessinsider.com/two-of-the-worlds-biggest-brands-are-cutting-back-on-on-digital-ads-2017-6
4
https://www.adbrands.net/top_global_advertisers.htm
5
Our estimate based on the (US$140M) 41% drop P&G’s adtech spending in the five months from Jan-May 2017
implies an original budget of $340M in that period, prorated for full year. Considering Unilever’s comparable total
marketing spend to P&G, according to 2016 filings, we assume similar amount is spent on adtech by Unilever.
6
https://www.sec.gov/Archives/edgar/data/1576427/000157642717000021/criteo10-
k2016.htm#sC09FFB47712B72B08A5A7ADC173C09DA

11
Restoration Hardware

The GCR report suggests Restoration Hardware reduced its spending by 100% with no effect on sales.
From the GCR report:

Source: GCR 2nd report, pg. 9

What RH seems to be actually referring to here is keyword searches like those offered by Google. This has
nothing to do with retargeting ads offered by Criteo. GCR is conflating two different issues here. This does
not mean RH has cut its digital ad spending by 100%. In fact, here is a recent online display ad from RH
promoting an October sale:

Source: www.moat.com

12
Uber

The GCR report suggests Uber reduced digital ad spending 100% by pointing to an isolated dispute
between Uber and Fetch Media. In fact, Uber actually increased its online ad spending in 2017. 7 This
should come as no surprise given that Uber is currently in pre-IPO hyper-growth mode. In fact, here is an
online ad that was active as of this writing:

Source: www.moat.com

The idea that Chase Bank, RH, and Uber all reduced their online advertising spending by 99-100% makes
absolutely no sense and doesn’t pass the sniff test. A 99%+ reduction in online advertisement would mean
that these companies would effectively cease to have any online ad presence. Such a marketing strategy
would make no sense for any company discussed here, but most of all Uber.

In fact, a recent report prepared independently by PwC for the Interactive Advertising Bureau (“IAB”)
notes:

“The first quarter of 2017 marked the highest ever Q1 earnings for digital advertising in
the US… Climbing 23 percent, or $3.7 billion, from the same period just a year ago, these
earnings represent the second-highest quarter of all time, and follow on the heels the
industry’s strongest quarter on record—Q4 2016.

‘The first quarter of 2017 represents the strongest beginning to any year yet in digital ad
spend,’ said David Silverman, a partner at PwC US. ‘It’s a testament to interactive’s ability
to attract audiences and the marketing dollars that follow.’”8

7
https://adexchanger.com/mobile/uber-upping-ad-spend-doubling-data/
8
https://www.iab.com/news/ad-revenues-hit-19-6b/ - most recent quarter. Q2 2017 data to be released in Nov.

13
 Click Misattribution
Key GCR claims:9

• Criteo takes credit for clients’ sales it did not contribute to, and in some cases, that never actually
occurred.
• Clients will leave or demand reimbursements from Criteo, due to brand safety & revenue
misattribution concerns.

*****

GCR’s claims here are primarily drawn from the Method Media Intelligence (“MMI”) report that they
published ahead of their own analysis. MMI itself seems to be a two-man operation with a website that
was set up less than a year ago, and a very limited online presence.10

The criticism effectively boils down to the idea that Criteo takes credit for clicks that it did not contribute
to, or were made by bots, and then overbills clients. The MMI report goes through a technical process of
showing how the clicks Criteo reports don’t match up with the client side, but ultimately seems to ignore
industry practice. In practice, advertisers don’t simply rely on Criteo’s self-reported numbers when it
comes to determining campaign efficiency or calculating ROI, but rather use independent third-party
verification to determine click attribution. As one industry insider put it:

“Any sophisticated marketer – and most of Criteo’s customers are sophisticated by the
way, they focus on the top enterprise retailers – are using some other source of truth. Very
few customers are logging into Criteo’s reporting and being like ‘Okay, Criteo delivered
this much, that must be what it is – let me write it down in my P&L.’”

“9 times out of 10, the customer is not even bothering with that metric in Criteo’s
reporting, or at least they shouldn’t be.”

Interviews with major advertisers and Criteo clients show a wide consensus that industry practice is to
use third-party marketing attribution services such as Google’s Adometry or Clario to analyze and optimize
advertising spend. The notion that all these sophisticated advertisers are being scammed by over-
reporting at the hands of Criteo, and somehow GCR alone has figured it out because they are so woke is
complete nonsense. That is not how the industry operates and that is not how the industry measures ROI.

“We don’t rely on [Criteo reporting] whatsoever. I don’t even look at them. I only look at
my stats. I never believe anything that somebody wants to sell me something tells because
they’re biased.”
- Major E-commerce client

9
GCR 1st report, pg. 5
10
Whois data, Google search term for “method media intelligence”

14
Furthermore, our research and industry interviews suggest that MMI and GCR’s concern over bot traffic
and click fraud is highly overblown. At the end of the day, Critieo’s clients need clicks to convert to actual
purchases, and bots don’t have credit cards. It’s pretty easy for advertisers to sift through the noise and
determine the true ROI of a performance-based marketing campaign. No one is getting fooled here except
for the investors shorting Criteo on faulty assumptions.

“I know I can give Criteo X amount of dollars each month and expect Y amount of revenue
in return. So, despite the fraud out there, which I still think is a small fraction of what’s
served up, I’m still getting performance on my investment in the form of hard conversions
to hard revenue.”
- Fortune 500 client

We think GCR is repackaging a non-issue to the advertising industry and serving it up as some sort of
revelatory thesis to the analyst and investor community.

“[MMI] is reporting something that isn’t a surprise to anyone, in many respects. It’s really
hard to get multiple systems to sync up their data for a whole variety of reasons…. [MMI]
was comparing what Criteo reporting said versus what Shopify reporting said, and that
doesn’t to the customer matter, because they are going to look at a third-party attribution
system... There’s kind of a false sense of precision that exists in the digital marketing
world, and the reality is because of people clearing cookies, because of lack of
deterministic identifiers across Facebook and Google, because of slow load times and ad
pixels getting deprioritized when a page loads, there’s a thousand reasons why these
things don’t sync up together that have nothing to do with fraud.”

- Major Criteo competitor

15
 Apple Safari Update
GCR’s third report focuses on Apple’s new Intelligent Tracking Prevention (“ITP”) update and what it
means for Criteo.11

ITP is a feature on the new Safari browser that uses machine learning to identify and quarantine third-
party cookies like those used by advertisers to track browsing habits. Apple announced ITP in early June
and the update went into effect mid-September.12 The June announcement led to shares of Criteo falling
15% from their highs, as investors await Q3 results to see the impact on earnings:

1-Year Criteo Share Price Chart

ITP Announced

Source: Google Finance

We believe the market is over-reacting to the rollout of ITP. Despite the share price drop, ITP isn’t really
a new feature from Apple – it’s more of an iteration on how Safari browsers have always behaved. Safari
has never been viewed as a friendly environment for targeted advertising. In fact, it has been blocking
third-party cookies in one form or another from the very beginning.13

Here are just a few of the headlines on the subject over the last few years:

2012 – Apple rolls out a new feature limiting ad tracking:

11
GCR 3rd report
12
https://techcrunch.com/2017/06/05/apple-adds-a-tracker-blocker-to-desktop-safari/
13
https://www.theverge.com/2013/2/23/4023078/firefox-to-start-blocking-cookies-from-third-party-advertisers

16
2013 – in this article, Firefox’s decision to join Apple in blocking third-party cookies by default is described
as a “nuclear first strike”:

2015 – a headline from this article declares the next version of Safari a blow for mobile advertising:

2015 – yet another doom-and-gloom headline from this article:

17
And so here we are in 2017 with yet another over-reaction to a headline.

Unlike Google and Facebook, Apple doesn’t make its money from advertising, which has helped it shape
its ideological path towards privacy. But beyond the public relations boost for Apple, ITP isn’t really
anything new, it’s just a reminder of what Apple has always done.

And ultimately, the industry will do what it’s always done: adapt.

Two years ago, ad blockers were considered an existential threat to the entire online advertising industry.
Here is a look at the search interest for ‘ad blocker’ back in 2015, courtesy of Google Trends:

Ad blockers were more of a threat to the adtech industry than ITP will ever be, because they potentially
affect 100% of all browsers, not just one. But you know what? The industry kept on sailing along, and
Criteo has continued to post 30%+ growth in revenue year after year.

Past as prologue aside, there is also a technical reason to believe ITP will have no real impact on Criteo.
ITP only begins to quarantine third-party cookies after 24 hours (and because ITP relies on machine
learning, it may take even longer to recognize third-party cookies):

Source: https://webkit.org/blog/7675/intelligent-tracking-prevention/

However, in a ‘last-click world’ like the one Criteo operates in, around 60-70% of click conversion happens
in the first 24 hours anyway.14 So right from the get-go, ITP leaves open the most valuable window for
retargeting.

14
Based on interview with industry expert

18
Criteo could also engineer a technical solution to ITP. For example, in 2014 Criteo developed a solution to
deliver personalized ads to devices that had third-party cookies blocked by default.15

One possible workaround that GCR criticizes is using HSTS protocol.

GCR’s third report is largely a philosophical argument on the regulatory environment facing HSTS. GCR
even went as far as to email Apple and complain about the issue. In our experience, once a short seller
starts getting into these type of arguments, and sending out emails in the hopes of triggering their own
investment catalyst, they’ve already lost.

Bill Ackman’s Herbalife short was a reminder that this is the wrong road to build your short thesis on. We
think investors that are shorting Criteo on the basis that ITP is going to be any sort of game changer or
have any sort of material impact come Q3 earnings and beyond are in for a rude awakening.

“Knowing what we know right now anyway, we’re not particularly worried about ITP
impacting at least Q3, Q4 spend and unless something unforeseen happens we don’t view
it as a significant headwind at all”
- Major Criteo competitor

“In my estimation [ITP] will have no significant impact on our ability to offer our services
to our customers”
- AdRoll co-founder, on conference call
(12 October 2017)

15
http://marketingmag.ca/tech/criteos-solution-for-using-cookies-on-iphone-110794/

19
 Conclusion
Let’s consider everything that’s happened since GCR released their reports on Criteo:

• The shares are trading higher now than they were before the reports.
• The GCR reports have had no impact on clients, whatsoever, according to industry interviews.
• The short interest in Criteo has increased and sits near all-time highs16

It seems all that GCR has accomplished is to set Criteo up for a short squeeze. There is nothing in the
reports that adequately argues a fundamental change in the business, and no information that is new or
revelatory to Criteo’s clients.

In a recent interview with The Drum, GCR sums up their short thesis like this: “In one sentence, our thesis
is if clients know what [Criteo’s] doing, it’s game over.”17

Turns out that clients know exactly what Criteo is doing, because they are the ones actually using the
system, and monitoring performance with independent third-party verification tools – and they love it.

Game over.

If Criteo reports no material impact from ITP on the next conference call, as we believe they will, we expect
the shares to spike up. Criteo was trading around 11x forward EBITDA prior to the ITP scare in June.
Applying the same multiple to 2018 EBITDA gets us US$64 per share. With shares currently trading at
US$47.20, this would suggest potential upside of 36%.

16
http://www.nasdaq.com/symbol/crto/short-interest
17
http://www.thedrum.com/news/2017/09/21/criteo-counters-fraud-claims-made-infamous-short-seller-gotham-
city-research

20