Admetrics

Admetrics - Ad Fraud

Transcript

1. Ad Fraud: Statistics and Techniques
2. 2 Billions of events every day Viewability tracking Engagement tracking Cookie-less device tracking Fraud prevention etc
3. “Viewability is number one. I would say preventing fraudulent traffic on your sites is number two, and that's from an industry standpoint […]” David Morris, Chairman of IAB Board of Directors 2015
4. Concerns About Digital Media Buying/Planning According to US Media Agency vs. Brand Marketing Executives, April 2016 (Brand marketers) 27 % 49 % 49 % 57 % 57 % 60 % 71 % 78 % 78 % Click fraud Bot traffic Viewability Ad blocking Conflicting audience data Control over content env Lack of audience data Consumer privacy High frequency exposure Source: MyersBizNet, “Survey of Advertising and Marketing Executives on Media Effectiveness” as cited by MediaVillage, May 2, 2016
5. “How much does bot traffic negatively affect your web properties or the performance of online campaigns?” 48 % Not Sure Source: iAB, distil networks, “A Digital Publisher’s Guide to Measuring and Mitigating Non-Human Traffic” October, 2015 40 % Not Sure Buyers Suppliers
6. 6 Questions fraudsters are going to ask themselves How big is the pie? How much money can I get? How big is the risk for fraudsters? What's the effort for fraudsters?
7. How big is the pie?
8. Source: “Digital Ad Spending Worldwide” report, eMarketer, March 2016 $ 300,63 $ 272,99 $ 243,82 $ 214,15 $ 186,81 $ 159,33 $ 133,32 Digital Ad Spending Worldwide, 2014-2020 2014 2015 2016 2017 2018 2019 2020
9. How much money can fraudsters get?
10. $7.2 billion of losses globally 3 to 37% bot rates ANA/White Ops, January 2016 The Bot Baseline: Fraud in Digital Advertising
11. “[…] between 88 and 98 percent of all ad-clicks on major advertising platforms such as Google, Yahoo, LinkedIn, and Facebook in a given seven day period were not executed by human beings, but rather by […] ‘bots.’” US Senator Mark R. Warner, US Senator Charles E. Schumer in a letter to FTC, July 2016
12. $ 186,8 $ 56 $ 18,7 $ 5,6 Estimated ad fraud losses 2016 2017 2020 $ 214,2 $ 64,3 $ 21,4 $ 6,4 $ 300,6 $ 90,2 $ 30 $ 9
13. 3 to 7–10% depends on a country Our clients
14. How big is the risk for fraudsters?
15. $14 Million collected from advertisers 4 Million computers infected 87 months in prison United States V. Vladimir Tsastsin, et. al., 11 Cr. 878
16. $750 000 in damages 3 years to clause the suit Microsoft v. Lam, et. al., 09-cv-0815
17. To plead a fraud claim, the complaint must allege sufficient facts to show: (1) a misrepresentation; (2) knowledge of the falsity; (3) intent to induce another's reliance on the misrepresentation; (4) justifiable reliance; and (5) resulting damages. Facebook, Inc. v. Grunin, No. C 14-02323 WHA.
18. What’s the effort for fraudsters?
19. Impression fraud Click fraud Attribution fraud
20. Impression fraud
21. Page refresh and manual labour @me illustration
22. Ad farms generate as many ad impressions as possible @me illustration
23. iFrame stuffing or pixel stuffing hides an ad in a 0x0 pixel @me illustration
24. Hidden ads are placed in hidden containers @me illustration
25. Ad stacking: multiple ads are placed on top of each other @me illustration
26. Ad injection: ads are inserted into website without permission and payment @me illustration
27. Bot traffic is generated by automated computed programs @me illustration, bots are looking at ads, not clicking (clicking is another slide)
28. Click fraud
29. Clickjacking tricks users to click on ads @me illustration
30. Bot traffic again @me illustration, now bots are clicking on ads
31. Attribution fraud Cookie stuffing Mobile attribution or CPI (cost-per-install) fraud Click fraud with bots Click hijacking with DNS rerouting
32. 32 Publishers You should expect 3, 10 or 30% of traffic to be fraudulent Check traffic for malicious bots Know sources of your paid-traffic
33. 33 Advertisers You should expect 3, 10 or 30% of traffic to be fraudulent Check campaigns for bot-unrelated fraud Check traffic for malicious bots Check your viewability vendor Ideally, you shouldn’t pay for fraudulent impressions and clicks Train your staff RTB: build white/blacklists Direct: keep an eye on bot traffic
34. 34 Hall 6, booth E-020 admetrics: Stefan Krätschmer: https://admetrics.io/events/dmexco2016 contact@admetrics.io sk@admetrics.io