CUSTOMER JOURNEY
Der Weg des Kunden zum Produkt umspannt Touchpoints in Online, Mobile, In-Store, Kundencenter, sozialen Netzen, Werbung im TV, Print, auf Plakaten, im Radio – einfach gesagt: Er sieht alles und das überall. Wie man dem Konsumenten trotzdem ein sinnvolles Bild der Marke gibt und ein unverwechselbares Angebot schafft. Wir begleiten den Konsumenten auf seinem Weg zum Kauf.
Prof. Dr. Jan Hendrik Schumann ist Inhaber des Lehrstuhls für Marketing und Innovation an der Universität Passau. Seine Forschungsschwerpunkte liegen in den Bereichen Onlinemarketing, Technologie und Innovation, Wertorientiertes Kundenbeziehungsmanagement und Internationales Marketing. Seine Forschung im Bereich Online-Marketing beschäftigt sich Prof. Dr. Schuman besonders intensiv mit dem Thema Customer Journey. Ziel der Forschung ist neben der Entwicklung eines besseren Verständnisses für die Such- und Entscheidungsprozesse im Internet auch die Entwicklung praktischer Tools zur Optimierung von Conversions und Budgetallokationen. Kooperationspartner aus der Praxis sind unter anderem die IntelliAd Media GmbH, die ValueClick Deutschland GmbH sowie Plan.Net. Die Forschungsarbeiten von Prof. Dr. Schuman werden vom deutschen Bundesministerium für Bildung und Forschung gefördert und wurden mehrfach international ausgezeichnet – zuletzt erhielt er einen Research Grant on Innovations in Advertising Effectiveness Measurement der Wharton Customer Analytics Initiative.
The Path to Product Excellence: Avoiding Common Pitfalls and Enhancing Commun...
20130711 - Customer Journey - Universität Passau - Jan Schumann
1. Innovation – Insights – Interaction
Prof. Dr. Jan Schumann, Universität Passau
Using Online User Journey Data for Conversion
Prediction and Attribution Modeling
Prof. Dr. Jan Hendrik Schumann
Lehrstuhl für Marketing und Innovation
Universität Passau
5. Werbeplanung.at Summit
Wien, 11. Juli 2013
2. Innovation – Insights – Interaction
Prof. Dr. Jan Schumann, Universität Passau
Advertisers employ various channels to reach consumers
over the Internet
Search engines Social Media Affiliate networks
Price comparisons Display/content ads Newsletter
Typical online channels for consumer communication
3. Innovation – Insights – Interaction
Prof. Dr. Jan Schumann, Universität Passau
Online user journeys are diverse and can comprise
multiple points of contact on different channels
Search
(SEO)Search
(SEO)
Display
Search
(SEA)
Blogs
Social
Networks
Newsletter
Online
shop 1
Online
shop 2
Online
shop 1
Price
comparison
Sites
Online
auctions
Search
(SEO)
Search
(SEA)
Group
Buying
Portals
Blogs
Forums
Review
Video
Portals
Affiliate
Price
comparison
Sites Newsletter
Social
Networks
Micro
media
Micro
media
Display
Customer journey
Conversion
4. Innovation – Insights – Interaction
Prof. Dr. Jan Schumann, Universität Passau
Despite major technological developments advertisers
often still struggle with fundamental questions
Typical questions of advertisers
To what extent does it pay off to
reach consumers on multiple
channels?
How should marketing
budgets be optimally
allocated?
How can I use information about the
prior user journey to predict
conversion probabilities?
5. Innovation – Insights – Interaction
Prof. Dr. Jan Schumann, Universität Passau
We used cookie-level data to analyze drivers of conversion
probabilities in multichannel campaigns
Conversion probabilities
of individual users
User history
Did user purchase
before?
Intensity
Number of clicks
Duration
Channels
Number of involved
channels
Channel switching
Informational
Navigational
Navigational
Informational
1
2
3
4
5
User journey
characteristics
Study 1
6. Innovation – Insights – Interaction
Prof. Dr. Jan Schumann, Universität Passau
We argue that channel switching behavior is a good
proxy for users„ purchase decision processes
Navigate to a
specific website
Classifying online advertising channels by primary user intent
(based on research on user intention in information retrieval scenarios)
Find information
on a specific topic
Search engines
Newsletter
Affiliate networks
Display/content ads
Study 1
7. Innovation – Insights – Interaction
Prof. Dr. Jan Schumann, Universität Passau
User history, number of channels involved and channel
switching are strong predictors of conversion probability
+2200%
User history
Did user purchase
before?
Intensity
Number of Clicks
Duration
Channels
Number of involved
channels
Channel switching
Inf. Nav.
Nav. Inf.
1
2
3
4
5
+108%
+600%
-15%
+3%
-0.2%
no yes
+ 1 Click
+ 1 Hour
+ 1 channel
no yes
no yes
User journey
characteristic Change
Impact on
conversion prob.
Study 1
8. Innovation – Insights – Interaction
Prof. Dr. Jan Schumann, Universität Passau
The results provide three key insights
1. Target recent customers
2. Try to reach individual users on multiple channels
3. Use user journey information for your budget allocation
(e.g., RTB)
Study 1
9. Innovation – Insights – Interaction
Prof. Dr. Jan Schumann, Universität Passau
To address the attribution problem we propose a complex
statistical model for budget allocation
Search
(SEO)Search
(SEO)
Display
Search
(SEA)
Blogs
Social
Networks
Newsletter
Online
shop 1
Online
shop 2
Online
shop 1
Price
comparison
Sites
Online
auctions
Search
(SEO)
Search
(SEA)
Group
Buying
Portals
Blogs
Forums
Review
Video
Portals
Affiliate
Price
comparison
Sites Newsletter
Social
Networks
Micro
media
Micro
media
Display
Marketers employ various online
channels such as SEA or Display
in their promotional mix
Little is known on how to attribute
credit to exposures along the user
journey
Today, marketers often rely on
simple heuristics like "last click
wins"
Advertisers’ questions
• Which framework can be applied
to ascertain the correct value
contribution?
• How should marketing budgets
be optimally allocated?
Our contribution
• Comprehensive analysis
framework based on first- and
higher-order Markov graphs
• Implementation and practical
impact in a real life system
Study 2Customer journey
Conversion
10. Innovation – Insights – Interaction
Prof. Dr. Jan Schumann, Universität Passau
Four real-life clickstream datasets are used to test and
validate our graph-based Markov framework
Data characteristics Descriptives
DS 1 DS 2 DS 3 DS 4
• Data collection in
cooperation with
intelliAd, a German
multi-channel tracking
provider
• 4 real-life clickstream
data sets from 3
industries
• Individual-level cookie
data including
converting and non-
converting journeys
Industry Travel Fashion
retail
Fashion
retail
Luggage
retail
Number of
different channels
8 8 8 8
Number of clicks 1,478,359 926,995 1,125,979 615,111
Number of
journeys
600,978 622,593 862,112 405,339
Thereof with
length ≥ 2
206,519 87,578 142,039 105,031
Average
journey length1
2.46
(8.860)
1.49
(3.142)
1.31
(1.238)
1.52
(4.587)
Number of
conversions
9,860 22,040 16,200 8,115
Journey
conversion rate
1.64% 3.54% 1.88% 2.00%
1) Standard deviation in parentheses
Study 2
11. Innovation – Insights – Interaction
Prof. Dr. Jan Schumann, Universität Passau
The new framework provides an improved measurement
of online channel contribution
Simple heuristic “last click” vs. novel Markov framework
SEO
16.8
14.5
SEA
19.3
18.3
Type In
30.8
44.3
Referrer
4.0
1.5
Display
5.0
2.5
Affiliate
11.3
8.9
Newsletter
12.7
9.8
Retargeting
--
--
Price
Comparison 0.2
0.1
High con-
tribution
Low con-
tribution
3.5
2.5
2.2
1.3
--
--
4.4
5.9
--
--
12.6
9.5
55.5
53.5
19.2
24.6
2.7
2.8
Online retail “apparel”1 Online retail “luggage / equipment”2
- 22%
4%
32%
--
--
68%
43%
- 26%
- 3%
- 31%
6%
16%
30%
26%
101%
159%
65%
--
1) Minimum journey length: 2; avg. journey length: 4.48 (7.73); journey conversion rate: 0.186
2) Minimum journey length: 2; avg. journey length: 3.00 (8.85); journey conversion rate: 0.047
Markov modelLast click wins … Relative change, percent
Value contribution by channel1
Percent Study 2
12. Innovation – Insights – Interaction
Prof. Dr. Jan Schumann, Universität Passau
This framework makes relevant contributions to
multichannel online marketing
1. Complex statistical models can lead to much fairer results than
simple heuristics
2. Channel attribution is a moving target and needs to be constantly
monitored
3. Framework is easy to interpret and understand for practicioners
(IntelliAd Attribution-Analyzer)
4. Framework is highly versatile and can be applied for various purposes
(attribution, RTB…)
Study 2
13. Innovation – Insights – Interaction
Prof. Dr. Jan Schumann, Universität Passau
Next frontiers in our research
1. Include more information (impressions, social media, multiple
devices, offline channels, offline behavior...)
2. Include additional financial measures such as costs, revenues and
CLV
3. Set up large-scale field experiments with randomized exposure
14. Innovation – Insights – Interaction
Prof. Dr. Jan Schumann, Universität Passau
Thank you very much for your attention!