15
DETECTING AND COMBATING FRAUD Oliver Weiss, Director Product Coordina5on & Business Development [email protected] AdFraud from a Buyer's Perspec5ve

DETECTING AND COMBATING FRAUD · Graph from White Ops’ Bot Benchmark Report PATTERN: DAY TIME Bots mimic human traffic volume paerns *Case Study by DCN and White Ops PATTERN: IMPRESSIONS

  • Upload
    others

  • View
    5

  • Download
    0

Embed Size (px)

Citation preview

Page 1: DETECTING AND COMBATING FRAUD · Graph from White Ops’ Bot Benchmark Report PATTERN: DAY TIME Bots mimic human traffic volume paerns *Case Study by DCN and White Ops PATTERN: IMPRESSIONS

DETECTINGANDCOMBATINGFRAUD

OliverWeiss,DirectorProductCoordina5on&[email protected]

AdFraudfromaBuyer'sPerspec5ve

Page 2: DETECTING AND COMBATING FRAUD · Graph from White Ops’ Bot Benchmark Report PATTERN: DAY TIME Bots mimic human traffic volume paerns *Case Study by DCN and White Ops PATTERN: IMPRESSIONS

Buyer’sConcernDetec5onSolu5on

Page 3: DETECTING AND COMBATING FRAUD · Graph from White Ops’ Bot Benchmark Report PATTERN: DAY TIME Bots mimic human traffic volume paerns *Case Study by DCN and White Ops PATTERN: IMPRESSIONS

WHYFILTERFRAUD?

Brandsloseconfidenceandmoneyindigitalmedia

NHTcomplicatesgeIngrealinsightsabout

campaignperformance

Failingtopreventfraudulenttrafficfundscriminal

ac5vity

Page 4: DETECTING AND COMBATING FRAUD · Graph from White Ops’ Bot Benchmark Report PATTERN: DAY TIME Bots mimic human traffic volume paerns *Case Study by DCN and White Ops PATTERN: IMPRESSIONS

SOMENUMBERS

Fraudaccountfor(IAB,comScore,WhiteOps):23%ofvideoimpressions

11%ofdisplayimpressions

36%ofadinterac5ons(clicks)involvebots

Mobileisleastaffected

AccordingtocomScore–54%ofdisplayadsarenotviewedbyhumans!

Page 5: DETECTING AND COMBATING FRAUD · Graph from White Ops’ Bot Benchmark Report PATTERN: DAY TIME Bots mimic human traffic volume paerns *Case Study by DCN and White Ops PATTERN: IMPRESSIONS

DETECTION

Page 6: DETECTING AND COMBATING FRAUD · Graph from White Ops’ Bot Benchmark Report PATTERN: DAY TIME Bots mimic human traffic volume paerns *Case Study by DCN and White Ops PATTERN: IMPRESSIONS

FRAUDULENTTRAFFICTYPES

Botnets

Crawler

HijackedDevices

ClickFarms

HidingIden5ty(Proxy)

DomainSpoofing

AdInjec5on

AdStacking

PixelStuffing

CloakedDomains

Fraud-NHT

Page 7: DETECTING AND COMBATING FRAUD · Graph from White Ops’ Bot Benchmark Report PATTERN: DAY TIME Bots mimic human traffic volume paerns *Case Study by DCN and White Ops PATTERN: IMPRESSIONS

PATTERN:FREQUENCY

GraphfromWhiteOps’BotBenchmarkReport

Page 8: DETECTING AND COMBATING FRAUD · Graph from White Ops’ Bot Benchmark Report PATTERN: DAY TIME Bots mimic human traffic volume paerns *Case Study by DCN and White Ops PATTERN: IMPRESSIONS

PATTERN:DAYTIME

Botsmimichumantrafficvolumepaaerns

*CaseStudybyDCNandWhiteOps

Page 9: DETECTING AND COMBATING FRAUD · Graph from White Ops’ Bot Benchmark Report PATTERN: DAY TIME Bots mimic human traffic volume paerns *Case Study by DCN and White Ops PATTERN: IMPRESSIONS

PATTERN:IMPRESSIONS

Comparedimpressions–NosignificantdifferenceVisibility

ExposureTime

Mouse-overTime

1614

65%64% Normal

Suspicious

54sec

sec

secsec

Page 10: DETECTING AND COMBATING FRAUD · Graph from White Ops’ Bot Benchmark Report PATTERN: DAY TIME Bots mimic human traffic volume paerns *Case Study by DCN and White Ops PATTERN: IMPRESSIONS

PATTERN:USERBEHAVIOR

Compareduserdata–SignificantdifferencesSession5me(5meafernextimpressionsisserved)

*Robotsseemoreimpressionsthannormalusers

Cookies

*Mobileandsimilardevicesexcluded

DomainsVisited

30010

0,651,00 Normal

Suspicious

10010

Page 11: DETECTING AND COMBATING FRAUD · Graph from White Ops’ Bot Benchmark Report PATTERN: DAY TIME Bots mimic human traffic volume paerns *Case Study by DCN and White Ops PATTERN: IMPRESSIONS

GETPARTNERS

%Fraudbyadver5singcountry,usingsampleblacklist

Page 12: DETECTING AND COMBATING FRAUD · Graph from White Ops’ Bot Benchmark Report PATTERN: DAY TIME Bots mimic human traffic volume paerns *Case Study by DCN and White Ops PATTERN: IMPRESSIONS

GETPARTNERS

%Fraudintop40domainsbytotalimpressions

Page 13: DETECTING AND COMBATING FRAUD · Graph from White Ops’ Bot Benchmark Report PATTERN: DAY TIME Bots mimic human traffic volume paerns *Case Study by DCN and White Ops PATTERN: IMPRESSIONS

SOLUTION

Page 14: DETECTING AND COMBATING FRAUD · Graph from White Ops’ Bot Benchmark Report PATTERN: DAY TIME Bots mimic human traffic volume paerns *Case Study by DCN and White Ops PATTERN: IMPRESSIONS

ACTIONPLANFORBUYERS

Monitorallyourcampaigns

Scanforadinjec5on

Createallies,notadversaries,inthefightagainstbotfraud

Managetheemo5onsofAdFrauddiscussions

Authorizeandapprovethird-partytrafficvalida5ontechnology

Toeffec5velycombatbotsintheirmediabuys,adver5sersmustbeabletodeploymonitoringtools.Publishersandagenciesmustenablethedeploymentofthesemonitoringtools.

SupporttheTrustworthyAccountabilityGroupTheIAB,4A’s,andtheANAannouncedinearlyNovemberthecrea5onoftheTrustworthyAccountabilityGroup(TAG).

Page 15: DETECTING AND COMBATING FRAUD · Graph from White Ops’ Bot Benchmark Report PATTERN: DAY TIME Bots mimic human traffic volume paerns *Case Study by DCN and White Ops PATTERN: IMPRESSIONS

Thank you!