Smart Multi-Channel Attribution Modeling

Beyond the Last Click: Smart Multi-Channel Attribution Modeling

Crispin SheridanSenior Director, Global Search, SAP

SAP Today

Annual revenue (IFRS) of € 16,22 billion• More than 232,000 customers

in 188 countries• More than 65,667 employees – and

locations in more than130 countries• A 41-year history of innovation and

growthSearch managed centrally

SEO: 70+ countries

PPC: 50+ Countries funded centrally

SAP.com Site Search

SAP.com Test Lab

500,000 pages, Search Team: 12 + 5 Agencies

>86%

of the world’s athletic

footwear provided

>50%

of the world’s

packages couriered

>70%

of the world’s

chocolate production

>30 million

households use energy

more responsibly with Smart

Grid solutions

>77,000

people work more safely

using People Safety

solutions

>2 million

people work more safely

using People Safety

solutions

SAP Customer Facts…

SAP’s Attribution Issues

• Long sales cycles

• Multiple individuals

• Interacting with all media types

• Multiple platforms

• Significant spend on Pull & Push marketing

So! What is the Customer journey?

Digital Journeys with SAP Prior to Conversion

1

2

3

Touch 1 Touch 2 Touch 3 Touch 4

Multiple Sources – Multiple Paths

Source: Strategic Oxygen – now Forrester

Multiple Individuals – One Deal

$

Marketing Touch Points

Then …..

The customer research team agrees :

“let’s take the next step with SAP”

Which Digital Tactics get credit? The default methods:

Last click

• Unfairly assigns credit to direct or Pull channels (e.g. SEM)

• Emphasizes branded keywords

• Unfairly assigns credit to awareness or Push channels (e.g. banners)

• Assigns credit even within a long look-back window

DCBA

DCBA First click

Challenge: Reports and analysis by silos

Lead Generation Report: Tactic LeadsPPC 451SEO 620Content Syndication 236Social Media 134Display 276Email 98Direct to web 502Mobile site 43Banners 90Campaign Microsite 348

Total 2,798

1,521

Demonstrative data only

Attribution Models to the Rescue?

Linear • Splits credit evenly for simplicity’s sake• Value diminishes with more touches• Potential to be just as misleading as last or first click

25%25%25%25%

?

• Assigns credit on a curve• Requires judgment calls/ guesswork• Few tools do it on the fly

40%10%10%40%

Weighted

• Assigns credit based on length of time passed• Sophisticated but tricky to code• Usually requires manual programming

Today 40%

2 daysago 30%

20 daysago 20%

40 daysago 10%

Recency

So, Right Attribution Model, but what else?

Micro & Macro Elements of Investment Optimization

MACRO Media Mix Modeling

MICRO

Digital Attribution

Test and Learn

So, Lets Focus on Micro & Digital Attribution

Digital Attribution

Multivariate Attribution Analysis Reveals:

• How channel overlap impact conversion rates

• The true reach and frequency across touch-points

• The contribution of each action, channel, or message

• The optimal sequence of channel exposure to drive a conversion

What is required? TAGS

• Tagging Governance and Process

• Master Tag across all Digital Touch points

• Tag Management system – Adobe, Tealium, etc.

• Tag Manager

ChannelContribution Index

to Conversion ABanner

Contribution Index to Conversion A

MessagesContribution Index

to Conversion A

SEO 150 Discovery 160 Banner A 170

Banner 130 NFL 125 Banner D 150

Social Media 125 ESPN 110 Banner C 98

Online Video 115 CNET 95 Banner F 91

SEM 113 NY Times 85 Banner B 70

Microsite 105 Forbes 50 Banner E 65

Content Page 102

SAP Homepage 95

e.g. Touch contribution is indexed & data is pivoted

Demonstrative data only

Volume of Uniques

Con

vers

ion

Rate

ACTION: Optimize Volume

Initial Results: MIX

Demonstrative data only

0

100

200

300

400

500

600

700

0 10 20 30 40 50U

niqu

es

(000)

Frequency of Exposure

Frequency & Uniques

7-170.0%

1.0%

2.0%

3.0%

4.0%

5.0%

0 10 20 30 40 50

Conv

ersi

on R

ate

Frequency of Exposure

Frequency & Conversion Rate

Finding: Optimal Frequency (for Conversion Rate)

frequency to conversion sweet spot

Demonstrative data only

7-17

retarget for more frequency

sweet spot wasted exposures

cap frequency

Actions:

Vehicle Last Touch Attributed DifferenceIntroducer

Volume IndexInfluencer

Volume IndexCloser

Volume Index

Banner 4,634 4,248 -8% 53 104 147

SEO 3,732 3,839 3% 81 121 92

SEM 1,792 2,035 14% 128 99 71

3rd Pty DG Promo 887 924 4% 138 77 90

Social Referral 4 3 -25% 130 83 102

Attribution for Media Vehicles

• Action – re-allocate budgets from Banners to SEM • Surprisingly, Banners act more as closers & SEM as

openers• SEM gained 14% additional conversion credit

because it indexes very high as an introducer and has high overlap with other vehicles. Introducers get majority of the credit from the models.

Low Sample

Together these drive Efficiency & Effectiveness

MACRO Media Mix Modeling

MICRO

Digital Attribution

Test and Learn

Key Takeaways

Distinguish between Macro and Micro optimization

Drive to your most relevant attribution model (s)

Ensure proper tagging structure (master tag)

Apply multivariate modeling & testing to optimize the impact of each tactic

Thank you Crispin Sheridan Senior Director, Global Search, SAP