case studies of reducing bots fraud by augustine fou
Post on 21-Apr-2017
807 Views
Preview:
TRANSCRIPT
Case Studies of Reducing Bots/Fraud
April 2017Augustine Fou, PhD.acfou [at] mktsci.com 212. 203 .7239
For Advertisers
April 2017 / Page 3marketing.scienceconsulting group, inc.
linkedin.com/in/augustinefou
Advertiser myths about ad fraud1. Fraud is 2% and ‘on the wane’
No. It’s because you’re catching less of it, because bots are better at hiding (think Methbot)
2. Fraud is ‘priced in’No. You may be paying 1/10th the CPM, but you’re buying 10x more impressions; an ad shown to a bot is useless.
3. We’ve got fraud-free guaranteesNice. But this assumes the detection tech can detect it; what if it can’t detect it, or it gets tricked? Is it fraud free?
April 2017 / Page 4marketing.scienceconsulting group, inc.
linkedin.com/in/augustinefou
Reduce bots/NHT in display campaignsPeriod 1 Period 3Period 2
Initial baseline measurement
Measurement after first optimization
After eliminating several “problematic” networks
April 2017 / Page 5marketing.scienceconsulting group, inc.
linkedin.com/in/augustinefou
“Humanness” of different sources
Organic sources have more humans (dark blue)
Conversion actions (calls) are well correlated to humans; bots don’t call
April 2017 / Page 6marketing.scienceconsulting group, inc.
linkedin.com/in/augustinefou
Higher quality means lower cost per human
Lower quality paid sources mean higher cost per human acquired – like 11X the cost.
Sources of different quality send widely different amounts of humans to landing pages.
April 2017 / Page 7marketing.scienceconsulting group, inc.
linkedin.com/in/augustinefou
Way better outcomes from better mediaMeasure
AdsMeasure Arrivals
Measure Conversions
good publishers, clean media, low bots
low-cost media, ad exchanges
346
1743
5
156
30X better outcomes
• More arrivals• Better quality
A
B
April 2017 / Page 8marketing.scienceconsulting group, inc.
linkedin.com/in/augustinefou
Make analytics more accurate and clean
7% conversion rate 13% conversion rateartificially low actually correct
For Publishers
April 2017 / Page 10marketing.scienceconsulting group, inc.
linkedin.com/in/augustinefou
Publisher myths about ad fraud1. Fraud doesn’t affect us, there’s low bots on our site
Bots don’t come in large quantities to your sites; they just collect a cookie and go elsewhere to create ad impressions
2. We have bot protection on our siteNice. But what if bad guys pretend to be your site by passing fake data, and put your brand reputation at risk?
3. We have high quality trafficGreat. We believe you. But what if bot detection tech accuses you of high bots (falsely)?
April 2017 / Page 11marketing.scienceconsulting group, inc.
linkedin.com/in/augustinefou
Good publishers take action to reduce bots
Publisher 1 – stopped buying traffic
Publisher 2 – filtered data center traffic
April 2017 / Page 12marketing.scienceconsulting group, inc.
linkedin.com/in/augustinefou
Filter bots to protect advertisers
On-Site measurement, bots are still coming
In-Ad measurement, bots and data centers filtered
10% red
-7% (filter data center traffic)
3% red
April 2017 / Page 13marketing.scienceconsulting group, inc.
linkedin.com/in/augustinefou
Defend against incorrect measurementsNHT counts higher
than ad impressions
1. In-ad measurement can’t look outside iframe
ad tag / pixel(in-ad measurement)
Challenges to measurement accuracy
2. Extrapolations from small sample or short time-based sample
Sources 1 and 2 corroborate
Source 3 completely off
April 2017 / Page 14marketing.scienceconsulting group, inc.
linkedin.com/in/augustinefou
Reduce javascript trackers to protect users
42 trackers24.3s load time
8 trackers1.3s load time
April 2017 / Page 15marketing.scienceconsulting group, inc.
linkedin.com/in/augustinefou
Reduce cookie matching, re-identification
specialized audience:oncologists
Journal of Clinical Oncology
specialized audience can be targeted elsewhere
“cookie matching”(by placing javascript on your site)
FOR EXAMPLE ONLY
ad revenue diverted away
Fraud comes in large numbers
April 2017 / Page 17marketing.scienceconsulting group, inc.
linkedin.com/in/augustinefou
$83B digital spend (2017)CPM and CPC fraud in context of digital spend
Source: eMarketer March 2017
Search Spend$40 $40
Display Spend Other
$16$30
$3
Google Search FB Display
$4E $11E
What’s left for good publishers
CPC Fraud
$5 Google Display
CPM Fraud
(75% of search) (40% of display)
$8$6
60% fraud
$29(outside Google/Facebook)
(display opportunity)
Source: eMarketer March 2017
• $7.2B out of $12B (2016)• $11B out of $19B (2017)40%
April 2017 / Page 18marketing.scienceconsulting group, inc.
linkedin.com/in/augustinefou
“sites that carry ads”
Top good domains vs “sites that carry ads”
Source: Verisign, Q4 2016329M
domains
$80B search + display
Google Search
FB+GOOG Display
$29billion
“mainstream sites you’ve heard of”
WSJESPN
NYTimes
EconomistReuters
Elle
top 1 million + next 10 million
159 million unknown sites
100% botpageviews on
“fraud sites”
99% human pageviews are on
“sites you’ve heard of”
3%
carry adsno ads
April 2017 / Page 19marketing.scienceconsulting group, inc.
linkedin.com/in/augustinefou
Little ad spend left for good publishersAd
verti
sers
Publishers are left with 30%
Bad Guyssiphon dollars OUT of the ecosystem
30% ($6B)
60% ($11B)
Ad Blockingusers use ad blocking to
protect themselves
10% ($2B)
Ad Tech“plumbing” and verification
Source: The Guardian, Oct 2016
$5B to Google Display
$16B to Facebook Display
Display Spend $
40B
Disp
lay
Spen
d
April 2017 / Page 20marketing.scienceconsulting group, inc.
linkedin.com/in/augustinefou
Example – 92% impressions cleaned up
Increased CPM prices by 800%
Decreased impression volume by 92%
Source: http://adexchanger.com/ad-exchange-news/6-months-after-fraud-cleanup-appnexus-shares-effect-on-its-exchange/
260 billion
20 billion
> $1.60
< 20 cents
April 2017 / Page 21marketing.scienceconsulting group, inc.
linkedin.com/in/augustinefou
Single botnet steals 15% of video ad spend
Source: Dec 2016 WhiteOps Discloses Methbot Research
“Methbot, steals $2 billion annualized; and it avoided detection for years.”
1. Targeted video ad inventory$13 average CPM, 10X higher than display ads
2. Disguised as good publishersPretending to be good publishers to cover tracks
3. Simulated human actionsActively faked clicks, mouse movements, page scrolling
4. Obfuscated data center originsData center bots pretended to be from residential IP addresses
April 2017 / Page 22marketing.scienceconsulting group, inc.
linkedin.com/in/augustinefou
About the AuthorApril 2017Augustine Fou, PhD.acfou [at] mktsci.com 212. 203 .7239
April 2017 / Page 23marketing.scienceconsulting group, inc.
linkedin.com/in/augustinefou
Dr. Augustine Fou – Independent Ad Fraud Researcher2013
2014
Follow me on LinkedIn (click) and on Twitter @acfou (click)
Further reading:http://www.slideshare.net/augustinefou/presentationshttps://www.linkedin.com/today/author/augustinefou
2016
2015
April 2017 / Page 24marketing.scienceconsulting group, inc.
linkedin.com/in/augustinefou
Harvard Business Review – October 2015
Excerpt:
Hunting the Bots
Fou, a prodigy who earned a Ph.D. from MIT at 23, belongs to the generation that witnessed the rise of digital marketers, having crafted his trade at American Express, one of the most successful American consumer brands, and at Omnicom, one of the largest global advertising agencies. Eventually stepping away from corporate life, Fou started his own practice, focusing on digital marketing fraud investigation.
Fou’s experiment proved that fake traffic is unproductive traffic. The fake visitors inflated the traffic statistics but contributed nothing to conversions, which stayed steady even after the traffic plummeted (bottom chart). Fake traffic is generated by “bad-guy bots.” A bot is computer code that runs automated tasks.
top related