Digital Marketing Intelligence
RUG 2022
Lecture 1. Introduction DMI
Part 1.
Digital Marketing: “An adaptive, technology-enabled process by which firms collaborate with
customers and partners to jointly create, communicate, deliver, and sustain value for all
stakeholders.” – Kannan and Li, 2017.
Relevance of digital marketing:
Bucklin et al. 2002
Digital Marketing Framework by Kannan and
Li, 2017
New problems of the modern internet age:
,Basic ingredient of DMI: Clickstream data – Clickstream data are defined as the electronic record of
Internet usage collected by Web servers or third-party servers. (Bucklin & Sismeiro 2009)
- Site centric: detailed records of what visitors do when navigating and interacting with a
specific site (offline: compare to loyalty card information of one specific store)
- User centric: detailed records of online behavior tracing across sites (offline: scanner data of
specific consumer products)
Site centric data analysed through website
conversion funnels
Some customers leave in the process: only 2%
make a purchase.
Measurements:
- 1000 Visitors
- 500 Visit shopping area (Shoppers)
- 100 Place item in cart (Pickers)
- 20 Make a purchase (Buyers)
- 10 Return for another purchase (Returners)
Insights
Metric Formula (x 100%) Outcome
Conversion Buyers/Visitors (20/1000) x 100% = 2%
Stickyness landing page Shoppers/Visitors (500/1000) x 100% = 50%
Relevance of the content Pickers/Shoppers (100/500) x 100% = 20%
Checkout effectiveness Buyers/Pickers (20/100) x 100% = 20%
Loyalty Returners/Buyers (10/20) x 100% = 50%
Drawbacks of funnels: it is too simple, it does not include all possible options/events -> “it is just
counting”
To increase traffic:
, - Sponsored search advertising: “A firm pays a certain fee to a search engine operator such as
Google or Yahoo to display its ad(s) in the form of a link along with organic search results.”
(Ghose and Yang 2009, p. 1605)
- Search Engine Optimization (SEO): “Labels efforts of firms that aim to improve the ranking of
the ad/link in the unsponsored search results. The display of the unsponsored (organic)
search results is free of charge, whereas advertisers pay for each click on their ads that
appear among the sponsored (paid) search results.” (Skiera, Eckert, Hinz 2010, p. 488)
Part 2: beyond counting
3 types of analysis by Bucklin and Sismeiro 2009:
1. Browsing behaviour and site usage (/apps)
2. Shopping behaviour on the internet
3. The internet for advertising
Browsing behaviour and site usage (/apps)
- E-commerce sites might want to speed up the process from visiting to purchasing (see also
Ngwe et al. 2019 paper)
o Decrease duration visit
o Learning effects relevant
o Previous findings can be used
- Entertainment and media websites/platforms may want to slow down the process: increase
duration visit
- If the are more frequent site visits, the visits duration decreases (Johnson, Bellman, and
Lohse 2003)
o Perhaps due to learning fewer page views or shorter visits
o (Bucklin and Sismeiro 2003) found that fewer page views, but not less time spent
viewing each page
During a visit, visitors knew where to go learning
So visit duration decreases
Mallapragada, G., Chandukala, S. R., & Liu,
Q. (2016). Exploring the effects of
“What”(product) and “Where”(website)
characteristics on online shopping
behavior. Journal of Marketing, 80(2), 21-
38.
More complicated analysis. Same variables,
but they also imply that website
characteristics etc. influence the
purchasing stage. Also basket value.
Shopping behaviour on the internet
, Understanding shopping behavior – Path to Purpose... (Li et al. 2020)
Customer journey: “A series of actions a customer takes to arrive at the moment of purchase” (Li et
al. 2020, p. 127)
3 research questions:
1. Do consumers use digital information channels differently for H/U purchases?
a.
2. How does this usage vary over the customer journey?
a.
3. Does this usage vary between converted and unconverted sessions?
a.
Traditional Market Basket Analysis:
- Goal: finding pairs of products that are jointly observed in large samples of baskets (also
called associations) one product cause consumer to buy another product
- Assumption(!): purchase of one product leads to the purchase of the remaining ones
- Cross-selling, Recommendation, Bundling
- Three key metrics:
o Support: joint probability of finding pair AB across all baskets
o Confidence: interpreted as the probability that purchase of A will lead to purchase of
B (but is actually the probability of a purchase of B given a purchase of A)
o Interest: discounts the joint probability by the popularity of the individual products
- Traditional MBA assumes that joint occurrence (measured by support and interest) implies
complementarity (= purchase of A leads to the purchase of B) but you could not conclude
that, which is the critism
Sequential MBA (basic idea: extending MBA with new data):
- Online stores can track which product was purchased first
- New metric:
o GAIN[A B]= the percentage gain in purchase probability for product B due to a
previous purchase of product A, relative to the probability of purchasing B
o GAIN is based on purchase sequence
o Compare to Confidence, which is only based on joint occurrence
o Some products have relationships which goes one way
o Key conclusions:
Purchasing sequence contains information
Sequential MBA leads to different conclusions than Traditional MBA on
complementarity
Sequential ‘input’ leads to sequential ‘output’
The internet for advertising
Internet advertising, two functions:
1. Using the Internet to measure advertising effectiveness
2. Using the Internet as an advertising medium (banner advertising)
Using the Internet to measure advertising effectiveness
a. Banner/display advertising may (Van Ewijk et al.2021):
RUG 2022
Lecture 1. Introduction DMI
Part 1.
Digital Marketing: “An adaptive, technology-enabled process by which firms collaborate with
customers and partners to jointly create, communicate, deliver, and sustain value for all
stakeholders.” – Kannan and Li, 2017.
Relevance of digital marketing:
Bucklin et al. 2002
Digital Marketing Framework by Kannan and
Li, 2017
New problems of the modern internet age:
,Basic ingredient of DMI: Clickstream data – Clickstream data are defined as the electronic record of
Internet usage collected by Web servers or third-party servers. (Bucklin & Sismeiro 2009)
- Site centric: detailed records of what visitors do when navigating and interacting with a
specific site (offline: compare to loyalty card information of one specific store)
- User centric: detailed records of online behavior tracing across sites (offline: scanner data of
specific consumer products)
Site centric data analysed through website
conversion funnels
Some customers leave in the process: only 2%
make a purchase.
Measurements:
- 1000 Visitors
- 500 Visit shopping area (Shoppers)
- 100 Place item in cart (Pickers)
- 20 Make a purchase (Buyers)
- 10 Return for another purchase (Returners)
Insights
Metric Formula (x 100%) Outcome
Conversion Buyers/Visitors (20/1000) x 100% = 2%
Stickyness landing page Shoppers/Visitors (500/1000) x 100% = 50%
Relevance of the content Pickers/Shoppers (100/500) x 100% = 20%
Checkout effectiveness Buyers/Pickers (20/100) x 100% = 20%
Loyalty Returners/Buyers (10/20) x 100% = 50%
Drawbacks of funnels: it is too simple, it does not include all possible options/events -> “it is just
counting”
To increase traffic:
, - Sponsored search advertising: “A firm pays a certain fee to a search engine operator such as
Google or Yahoo to display its ad(s) in the form of a link along with organic search results.”
(Ghose and Yang 2009, p. 1605)
- Search Engine Optimization (SEO): “Labels efforts of firms that aim to improve the ranking of
the ad/link in the unsponsored search results. The display of the unsponsored (organic)
search results is free of charge, whereas advertisers pay for each click on their ads that
appear among the sponsored (paid) search results.” (Skiera, Eckert, Hinz 2010, p. 488)
Part 2: beyond counting
3 types of analysis by Bucklin and Sismeiro 2009:
1. Browsing behaviour and site usage (/apps)
2. Shopping behaviour on the internet
3. The internet for advertising
Browsing behaviour and site usage (/apps)
- E-commerce sites might want to speed up the process from visiting to purchasing (see also
Ngwe et al. 2019 paper)
o Decrease duration visit
o Learning effects relevant
o Previous findings can be used
- Entertainment and media websites/platforms may want to slow down the process: increase
duration visit
- If the are more frequent site visits, the visits duration decreases (Johnson, Bellman, and
Lohse 2003)
o Perhaps due to learning fewer page views or shorter visits
o (Bucklin and Sismeiro 2003) found that fewer page views, but not less time spent
viewing each page
During a visit, visitors knew where to go learning
So visit duration decreases
Mallapragada, G., Chandukala, S. R., & Liu,
Q. (2016). Exploring the effects of
“What”(product) and “Where”(website)
characteristics on online shopping
behavior. Journal of Marketing, 80(2), 21-
38.
More complicated analysis. Same variables,
but they also imply that website
characteristics etc. influence the
purchasing stage. Also basket value.
Shopping behaviour on the internet
, Understanding shopping behavior – Path to Purpose... (Li et al. 2020)
Customer journey: “A series of actions a customer takes to arrive at the moment of purchase” (Li et
al. 2020, p. 127)
3 research questions:
1. Do consumers use digital information channels differently for H/U purchases?
a.
2. How does this usage vary over the customer journey?
a.
3. Does this usage vary between converted and unconverted sessions?
a.
Traditional Market Basket Analysis:
- Goal: finding pairs of products that are jointly observed in large samples of baskets (also
called associations) one product cause consumer to buy another product
- Assumption(!): purchase of one product leads to the purchase of the remaining ones
- Cross-selling, Recommendation, Bundling
- Three key metrics:
o Support: joint probability of finding pair AB across all baskets
o Confidence: interpreted as the probability that purchase of A will lead to purchase of
B (but is actually the probability of a purchase of B given a purchase of A)
o Interest: discounts the joint probability by the popularity of the individual products
- Traditional MBA assumes that joint occurrence (measured by support and interest) implies
complementarity (= purchase of A leads to the purchase of B) but you could not conclude
that, which is the critism
Sequential MBA (basic idea: extending MBA with new data):
- Online stores can track which product was purchased first
- New metric:
o GAIN[A B]= the percentage gain in purchase probability for product B due to a
previous purchase of product A, relative to the probability of purchasing B
o GAIN is based on purchase sequence
o Compare to Confidence, which is only based on joint occurrence
o Some products have relationships which goes one way
o Key conclusions:
Purchasing sequence contains information
Sequential MBA leads to different conclusions than Traditional MBA on
complementarity
Sequential ‘input’ leads to sequential ‘output’
The internet for advertising
Internet advertising, two functions:
1. Using the Internet to measure advertising effectiveness
2. Using the Internet as an advertising medium (banner advertising)
Using the Internet to measure advertising effectiveness
a. Banner/display advertising may (Van Ewijk et al.2021):
