4. Augustine Fou- 4 -
“Digital ad fraud is well known and well
documented. However, the magnitude of it
may not yet be well understood.
Furthermore, ‘industry constituencies
[may be] insufficiently motivated’ (Source:
IAB) to act ferociously to deter and reduce
these fraudulent activities.”
-- Dr. Augustine Fou
5. Augustine Fou- 5 -
Why Now?
“As more ad inventory is bought and sold
programmatically on ad exchanges, bad
guys are finding it far easier to commit fraud
because few agencies and advertisers
actually check in detail the hundreds of
thousands of sites on which the ads are run.
It’s easier to hide in a far larger haystack.”
-- Dr. Augustine Fou
6. Augustine Fou- 6 -
The Opportunity
As more digital
ads are placed
entirely
programmatically,
the opportunity for
fraud continues to
increase.
7. Augustine Fou- 7 -
The Motive
“Highly Lucrative, Profitable
The aggregate ad revenue for the
sample of 596 sites was an estimated
$56.7 million for Q3 of 2013,
projecting out to $226.7 million
dollars annually, with average profit
margins of 83%, ranging from 80%
to as high as 94%.”
Source: Digital Citizens Alliance Study,
Feb 2014
8. Augustine Fou- 8 -
Digital Ad Spend (IAB FY 2013)
Impressions
(CPM/CPV)
Clicks
(CPC)
Leads
(CPL)
Sales
(CPA)
Search 43%
$18.5B
Video 7%
$3.0B
Lead Gen 4%
$1.7B
11% Other
$4.8B
Source: IAB, FY 2013 Internet Advertising Report, April 2014
$36.2B
Display 19%
$8.2B
Mobile 15%
$3.7B$2.8B
NOTE: figures from IAB FY 2013
($43B annual ad spend)
CPM Performance
• classifieds
• sponsorship
• rich media
• email
$6.5B
9. Augustine Fou- 9 -
Digital Ad Fraud Estimates
Impressions
(CPM/CPV)
Clicks
(CPC)
Search
$18.5B
$36.2B Ad Spend
Display
$8.2B
Video
$3.0B
Mobile
$3.7B$2.8B
(U.S. digital ad spend only)
DollarsAvg Fraud Rate
$4B50%Display
$2B60%Video
$6B30%Search
$2B40%Mobile
$14BTOTALS 39%
Multiple sources
10. Augustine Fou- 10 -
Ad Fraud Ranges
• Display ad fraud fake ad
impressions created by bots
• Video ad fraud fake video ad
views generated by bots
• Mobile ad fraud fraudulent
or accidental clicks
• Search ad click fraud click
fraud on search ads
30-70%
50-80%
20-40%
$3-7B
$1-2B
$4-8B
(U.S. Only)
30-50% $2-3B
50%
60%
30%
40%
Range DollarsAverage
12. Augustine Fou- 12 -
19 blocked ads on page
Bad Guys Put Up Sites
site = analyzecanceradvice .com
site = missomoms.com
13. Augustine Fou- 13 -
Load Tons of Ads on Pages
http://interiorcom plex.com/
http://modernbab y.com/
14. Augustine Fou- 14 -
Load Ads in Hidden iFrames
Augustine Fou- 14 -
Source: Spider.io May 2, 2013
Ads are hidden in up to
72 layers
Entire web pages are
loaded into ad iframes to
boost impressions
Multiple redirects and
auto page refreshes
15. Augustine Fou- 15 -
Fake their “Viewability”
Augustine Fou- 15 -
Source: Spider.io May 2, 2013
Ads are above the
fold of the page
But their pixel
opacity can be set
to zero (invisible)
Entire web pages
stuffed into ad
iframe; ads counted
as viewable
17. Augustine Fou- 17 -
Sell Impressions on Exchanges
“Modernbaby.com and Interiorcomplex.com
Each of these sites peddles enormous traffic on the
exchanges. For example, on a recent day Modern
Baby was offering 19 million impressions via one
exchange (quite the baby boom) and Interior
Complex 30 million [ad impressions PER DAY] (the
roaring housing market must be back).”
Source:
Adweek – Suspicious Web Domains Cost Online Ad
Business $400m per Year
By Mike Shields
19. Augustine Fou- 19 -
Display Ad Impressions
• By far, the easiest to commit – fill a web page with many
display ads and repeatedly load the page with scripts or bots
• Clients pay on CPM basis and are usually goaled on the
number of impressions delivered and CPM
30 – 75%
fraudulent, not-in-view, adblocked
IAB: FY 2012 Display Ad Spend = $7.7B
21. Augustine Fou- 21 -
Viewability of Display Ads
Source: Integral Ad Science Sept 2013 via MarketingCharts
Publisher direct is
best and lowest
risk
22. Augustine Fou- 22 -
23% of Users Use Adblockers
Source: PageFair, August 2013
Average Adblocking Rate
22.7%
23. Augustine Fou- 23 -
Traffic Firehose On/Off
Legit human traffic does not change rapidly; but bot traffic
(firehose) can be rapidly turned off and directed to other sites.
Source: Alexa
24. Augustine Fou- 24 -
Real-Time Bidding
Real-Time
Bidding to
Account for
25% of Display-
Ad Spending by
2015: eMarketer
25. Augustine Fou- 25 -
Where Fake Impressions Come From
Source: Integral Ad Science, via BusinessWeek Nov 26, 2013
Ad exchanges are particularly prone to fraud because shady sites sell enormous
quantities “ad impression inventory” into the exchanges for re-sale.
26. Augustine Fou- 26 -
Video Ad Impressions
• Advanced bots are programmed to load video ads and wait
till they are counted as “views” before leaving the page
• Video ads have 10x the CPM compared with display ads, and
are therefore a prime target for ‘bad guys’
50 - 77%
fraudulent, autoplay, wasted, not-in-view
eMarketer: Digital Video Ad Spend est. $4B in 2013
Source: Vindico via Adweek, December 15, 2013
27. Augustine Fou- 27 -
Video Ad Views Exploded
Source: comScore, Jan 2014 via MarketingCharts
Dec 2013 35.2B
video ads per month
28. Augustine Fou- 28 -
Digital Video Ad Spend
Source: eMarketer, May 2013
Digital video ad
spend continues to
rise rapidly as more
TV ad dollars shift to
digital.
Because video CPMs
are so much more
lucrative, it is more
lucrative for bad
guys too.
29. Augustine Fou- 29 -
Video Ads Are Lucrative
Video ad CPMs
are $8 - $12,
which means
they are more
than 10X more
lucrative than
display or
mobile ads.
Source: Turn, Inc.
October 2013
30. Augustine Fou- 30 -
A new study [download page] from
VideoHub indicates that during the first
quarter, average viewability for online
video ads in the US was 83%, although
rates varied widely among properties,
from 43% on the low end to 94% on the
high end. Broadcast TV sites fared best
(89% on average) among property
categories, with networks and
exchanges (73%) bringing up the rear.
Video Ad Viewability
Source: VideoHub August 2013 via MarketingCharts
31. Augustine Fou- 31 -
500 billion
display ad impressions /mo
Impression (CPM) Fraud
29%
confirmed bot traffic
~$1 - $3.50
cost per thousand
$2-6 billion
wasted ad spend (annualized)
Display Ads
35 billion
video ad impressions /mo
40%
estimated fake views
~$8 - $12
cost per thousand
$1-2 billion
wasted ad spend (annualized)
Video Ads
Source: Vindico, 2013Source: Solve Media 2013
37. Augustine Fou- 37 -
Redirect Link(s)
Known blackhat
technique to hide real
referrer and replace
with faked referrer.
SIX (6) redirects
before ending on the
landing page
See how-to:
http://www.blackhatworld.com/
blackhat-seo/cloaking-content-
generators/36830-cloaking-
redirect-referer.html
40. Augustine Fou- 40 -
Search Ad Clicks
• Search ad fraud is a bit more involved to commit and usually
occurs on “search partner” sites (not the main search sites)
• Bad guys set up sites with no content, execute searches with
lucrative keywords, and click the ads with bots
20 - 40%
fraudulent, accidental, wasted
Source: Adometry Click Fraud Report 1H 2013
IAB: FY 2012 Search Ad Spend = $16.8B
42. Augustine Fou- 42 -
Bot Clicks vs Humans
Humans actually click
on buttons and menu
items (mouse moves
and clicks)
Bots don’t bother
disguising their click
locations and don’t
show mouse traces.
Source: Spider.io Feb 2013
43. Augustine Fou- 43 -
Before and After
18% of spend shifted from fraudulent websites to “top 5” good guys
BEFORE
Top 2 “good guys” = 76%
AFTER
Top 5 “good guys” = 94%
52. Augustine Fou- 52 -
Algo-generated Pages
Source:
http://www.highdefdigest.com/tags/s
how/Seiki_Digital
Characteristics:
• typically blog platform
• keywords optimized for each
post to attract organic traffic
• content plaigiarized from press
releases and other sites
• stuffed with display ads,
affiliate links or link bait
53. Augustine Fou- 53 -
Not Human-Readable
Characteristics
• Auto-generated by bots,
stuffed with search keywords
• To attract organic search traffic
• Not human readable (low
instance of natural phrases)
• Stuffed with affiliate links and
display ads
55. Augustine Fou- 55 -
Other Sites Using Same
Affiliate ID
Source: http://reverseinternet.com/amazon-
affiliate/panandscathed-20
Source: http://sameid.net/amazon-aff/panandscathed-20/
57. Augustine Fou- 57 -
Prime Targets for Fraud
Source: Ben Edelman
Since 2004, I've been tracking and
reporting all manner of rogue
affiliates -- using spyware and
adware to cover competitors'
sites; using trickier spyware and
adware to claim commission on
merchants' organic
traffic; typosquatting; stuffing
cookies through invisible IFRAME's
and IMG's, banner ads, and
even hacked forum sites; and the list
goes on.
61. Augustine Fou- 61 -
Industry Players
Solve Media
Using CAPTCHAs to detect
humans versus bots.
Global Bot Traffic on Pace to
Waste Up to $9.5 billion in 2013
Ad Budgets. Sep 2013
Spider.io
Nielsen/IAB
Algo detection of bot-like
activity and other malware.
Botnet Costing Display
Advertisers $6 Million per
month. Feb 2013
Industry working group to
define ad viewability.
Viewable rates (of display
ads) ranged from 14% to
79%.
IntegralAds (AdSafe)
Verify ad placement against
blacklist of known fraudulent
websites.
Viewable rates (of display ads)
ranged from 14% to 79%.
DoubleVerify
Ad placement, behavioral
compliance, fraud
detection
PageFair
Ad blocking detection
WhiteOps
Realtime bot detection
algorithms
63. Augustine Fou- 63 -
Blacklisting Sites
Value
Exclude sites from
serving your ads
Caveat
For every site excluded,
bad guys put up more
(because they don’t have
to play by the rules).
64. Augustine Fou- 64 -
Enforcing Viewability
Source: Spider.io May 2, 2013
Value
Only pay for ads which
are viewable (i.e. above
the-fold)
Caveat
Bad guys have already
defeated “viewability” by
stuffing ads in hidden layers,
all above-the-fold.
65. Augustine Fou- 65 -
Detecting Bot Traffic
Source: Spider.io March 2013
Value
Good guys use algorithms
to detect unusual
behaviors indicative of
bots (rather than humans)
Caveat
It’s an arms race between
good and bad; bots are more
sophisticated and can fake
mouse movements and keep
cookies.
66. Augustine Fou- 66 -
Using CAPTCHAs
“Startup called Vicarious
automatically solves
CAPTCHAs.” Oct 2013
http://bit.ly/1bFo9lZ
Value
Captchas deter bots from
filling in forms and stealing
content and cookies.
Caveat
Some bots can now solve some
captchas, most captchas don’t
protect content pages.
Source: Solve Media Dec 31 2013
67. Augustine Fou- 67 -
“The above countermeasures are all good, and
advertisers should continue using them. But they are
not enough. If the good guys fight the fight
individually, there is little chance they can overcome
the entire ecosystem of the bad guys. The good guys
need to band together into their own ecosystem and put
the bad guys on a ‘digital ad fraud equivalent to
the National Sex Offenders Registry’.”
-- Dr. Augustine Fou
68. Augustine Fou- 68 -
Assumptions No Longer Valid
1. bots can’t fake mouse movements and webpage scrolling - they
can easily now
2. captchas can only be solved by humans - bots can solve them
too now
3. it requires malware infected computers to commit ad fraud -
bad guys can set up hundreds of thousands of server instances to
simulate users without having to infect any computers with
malware
4. malware can be caught by virus software when installed -
some malware does not need to be installed, they are carried along
with the code of a toolbar, plugin, extension, etc. or can be
asynchronously introduced later via updates (especially when user
has permitted auto-updates)
69. Augustine Fou- 69 -
Assumptions No Longer Valid
5. if a correct bid record is passed it should be a human user
- bots can easily send fake information to simulate being a
user (e.g. cookies, referrer, search query, characteristics of the
computer and browser, etc.)
6. fraudulent traffic is a small portion of legitimate human
traffic -- bot traffic is far larger than actual human traffic.
This comes from both “legitimate” sources like Google
crawlers or “site up status checkers” and “illicit” sources like
server-side scripts, browser side scripts, browser extensions
and plugins, and other malware.
71. Augustine Fou- 71 -
Areas of Optimization
bots /not seen by humans
delivery
viewability
targeting
waste
reduction
improving
optimization
30%
40%
30%
72. Augustine Fou- 72 -
Low Hanging Fruit
The most immediate, direct impact on ROI comes from reducing waste
23% Ad Blocked (wasn’t shown)
(PageFair)
54% Not In View (not seen)
(comScore)
24 – 29% confirmed bot
(Solve Media)
25% On-Target Delivery
(Nielsen)
82% Ignored (not relevant)
(Harris Interactive)
73. Augustine Fou- 73 -
Digital Ad Forensics Process
Sizing of
ad fraud
Forensic AnalysisPreliminary Scan
Implementation
Augustine Fou- 73 -
• Technology Tools
• Statistical analysis
Maintenance
Preliminary analysis of
paid campaigns and
analytics to determine
magnitude of the ad
fraud impacting client.
FREE
Creating recommended
list of changes,
including list of sites to
exclude in each ad
channel.
$$$
Subscribe to triangulated,
cross-industry database of
“ad fraud offenders” to
continuously update
blacklists and whitelists.
$
• Budget shifts
• Further optimization
74. Augustine Fou- 74 -
Related Articles
Fake YouTube Videos
By: Augustine Fou, December 2013
Fake Linkedin Profiles
By: Augustine Fou, December 2013
Fake Facebook Profiles
By: Augustine Fou, Dec 2013
Fake Twitter Accounts
By: Augustine Fou, August 2013
An Ecosystem of Digital Ad Fraud
By: Augustine Fou, October 2013
Augustine Fou- 74 -
Digital Ad Fraud Briefing
By: Augustine Fou December 2013
Ad Fraud Fighting Techniques
By: Augustine Fou October 2013
How Display Fraud Works
By: Augustine Fou, May 2013
How Click Fraud Works
By: Augustine Fou, November 2013
The Magnitude of Digital Ad Fraud
By: Augustine Fou, November 2013
75. Augustine Fou- 75 -
Dr. Augustine Fou – Digital Consigliere
“I advise clients on optimizing
advertising across all channels. One
main area of focus is reducing ad waste
due to fraud – fake impressions, clicks,
leads, and sales.”
FORMER CHIEF DIGITAL OFFICER, HCG (OMNICOM)
MCKINSEY CONSULTANT
CLIENT SIDE / AGENCY SIDE EXPERIENCE
PROFESSOR AND COLUMNIST
ENTREPRENEUR / SMALL BUSINESS OWNER
PHD MATERIALS SCIENCE (MIT '95) AT AGE 23
@acfou
ClickZ Articles: http://bit.ly/augustine-fou-clickz
Slideshares: http://bit.ly/augustine-fou-slideshares
LinkedIn: http://linkd.in/augustinefou
77. Augustine Fou- 77 -
IAB Full Year 2012
Source: IAB 2012 Annual Report
78. Augustine Fou- 78 -
Display Ads per Month
480 billion display ads per month
1,439 billion display ads per quarter
79. Augustine Fou- 79 -
Fraud Fighting Techniques
• Blacklisting fake websites by URL/domain
• Whitelisting legit publisher sites by domain
• Detecting bot traffic and known botnet addresses
• Scanning sites for malware (Google hosts list)
• Bot flag on bid records
• Real time rejection of fraudulent ad impressions
• Revised traffic numbers based on audience overlap