Module 05: Discussion Analytics, Techniques, and Risks

Digital marketing analytics
Second Edition

Chapter 5
Digital Analysis:
Audience

Copyright © 2022, 2018, 2016 Pearson Education, Inc. All Rights Reserved

Figure 5.1 Google Ads Personalization matrix example.
From Digital Marketing Analytics: Making Sense of Consumer Data in a Digital World, 2/e by Chuck Hemann and Ken Burbary (0789759608) Copyright
© 2018 Pearson Education, Inc. All rights reserved

Visit Google site to explore

What Is Audience Analysis?
● Analysis—Who is the audience?

● Understanding—What is the audience’s knowledge and attitude toward the brand?
● Demographics—What is the audience’s age, gender, education, location, and so on?
● Interest—Why is the audience reading, sharing, and interacting with your brand
content?
● Environment—Where does the audience spend time online?
● Needs—What are the audience needs associated with your brand, product, or
service?

● Customization—What personalization and/or attributes of the experience should
the brand address in order to add value for the audience?
● Expectations—What are both the stated and unstated expectations that the
audience has for their interactions with your brand?

Audience Analysis Use Cases
● Digital strategy development

●

Content strategy development

●

Engagement strategy development

●

Search engine optimization

●

Content optimization

●

User experience design and optimization

●

Audience segmentation

Figure 5.2 Content marketing growth.
From Digital Marketing Analytics: Making Sense of Consumer Data in a Digital World, 2/e by Chuck Hemann and Ken Burbary (0789759608) Copyright
© 2018 Pearson Education, Inc. All rights reserved

Engagement Strategy
● Brands develop engagement strategies to maximize the number of
desired outcomes produced on social platforms.
● Best-laid plans can go awry and require course correction.
● Nearly every major social platform has a native engagement analytics
tool, with Snapchat being the only current exception.
● Facebook, Twitter, LinkedIn, Instagram, and YouTube all provide native
analytics to anyone with a brand presence on their platforms.
● A diverse third-party engagement analytics ecosystem also exists.
● The problem is not lacking analytics tools to measure and optimize your
engagement, but having too many tool choices.

Search Engine & Content Optimization
● Getting your content discovered through organic searches.
○ inclusion of social data signals into search engine algorithms.
● Infuse the content distributed in your social status updates, tweets,
blog posts, comments, and so on.
● The output of the SEO analysis has multiple uses, because content or
links that are shared on social platforms are ranking signals for search
engines such as Google.
● Exactly how much weighting gets put into these signals when
calculating search rankings isn’t clear, but Google has confirmed they
are indeed factors.
● This means an irrevocable relationship exists between content and
social planning, and publishing and linking efforts.

User Experience Design
User experience design is important for simplifying things enough so that
users can complete desired tasks and leave a digital experience satisfied.
● audience analysis via digital analytics plays a big role in informing user
experience designers about what the relevant steps are along a consumer
journey, by providing what users need and expect, and alternatively,
which steps are broken, causing dissatisfaction and abandonment.
● Web analytics, site surveys, and social analytics can reveal a combination
of what people are doing and saying about their experience. Designers
can incorporate these feedback mechanisms as input and optimize user
flows accordingly.

Audience Analysis Tool Types
Search insights—Search insights tools can be broken down into two distinct
subcategories.
● The first provides insights into search intent based on actual
consumer searches with Google or other search engines.
● The second subcategory of Search insights is very much keyword
based, with a focus on PPC/Paid Search activity and data.
SEO—You can still use SEO tools to monitor, track, and manage both your own
User surveys—Surveys are a valuable source of qualitative feedback that you
can tie to web analytics data to connect what an audience says with what it
actually does.

Website profiling— enter a website URL and get back a snapshot in time
profile of a website based on panel data from that service that reveals website
traffic statistics, search volume, referring traffic sources, demographic data on
visitors, related sites, and more.

Figure 5.3 Digital intelligence technologies marketers use most.
From Digital Marketing Analytics: Making Sense of Consumer Data in a Digital World, 2/e by Chuck Hemann and Ken Burbary (0789759608) Copyright
© 2018 Pearson Education, Inc. All rights reserved

Audience Analysis Tool Types
Social listening—The social listening marketplace has arguably undergone the
most change and disruption in recent memory.
● Despite all of the disruptions and changes in the social listening industry,
more changes were ahead.
Influence analysis—Identifying individuals responsible for driving
action among relevant members of your audience is important.
Sharing analysis—Tools are available that are specifically dedicated to tracking
share activity of content across digital platforms.
Social profile and activity analysis—Social brand profile and activity analysis is
one of the features most commonly found in third-party social analytics tools
currently on the market.

Additional Audience Analysis Techniques
● What conversations are taking place about my brand?

●

What topics and/or themes are most talked about?

●

Who is my audience?

●

Where do they spend their digital time?

●

What do we know about their content consumption habits, preferences
and engagement behavior(s)?

●

What are the affinities and related interests of my audience?

●

What is my audience saying about their needs, both met and unmet?

●

How does my audience feel about specific brand, product, or service
attributes?

Audience Preferences
● Conversation Typing
● Event Triggers

Segmentation overview
Chapter 13
Segmentation: in marketing strategy, a method of sub−dividing the
population into similar sub−markets for better targeting, etc.

Strategic uses of segmentation
● Marketing research: learning WHY. Segmentation provides a rationale
for behaviour.
● Marketing strategy: targeting by product, price, promotion and place.
Strategy uses the marketing mix by exploiting segment differences.
● Marketing communications: messaging and positioning. Some
segments need a transactional style of communication; other segments
need a relationship−style of communication. One size does not fit all.
● Marketing economics: imperfect competition leads to price makers.
With the firm communicating just the right product at just the right price
in just the right channel at just the right time to the most needy target,
such compelling offers give the firm nearly monopolistic power.

The four Ps of strategic marketing
Partition
The first step is to partition the market by applying a (behavioural) segmentation
algorithm to divide the market into sub−markets.
Probe
This second step is usually about additional data. Often this may come from marketing
research, probing for attitudes about the brand, its competitors, shopping and purchasing
behaviour, etc.
Prioritize
This step is a financial analysis of the resulting segments. Which are most profitable,
which are growing fastest, which require more effort to keep or cost to serve, etc?
Position
Positioning is about using all of the above insights and applying an appropriate message,
or the correct look, feel and style. This is the tool that allows the creation of compelling
messages based on a segment’s specific sensitivities.

Criteria for actionable segmentation
Identifiability. In order to be actionable each segment has to be identifiable. Often this is
the process of scoring the database with each customer having a probability of belonging
to each segment.
Substantiality. Each segment needs to be substantial enough (large enough) to make
marketing to it worthwhile. Thus there is a balance between distinctiveness and size.
Accessibility. Not only do the members of the segment have to be identifiable, they
have to be accessible. That is, there has to be a way to get to them in terms of marketing
efforts. This typically requires having contact info, email, direct mail, SMS, etc.
Stability. Segment membership should not change drastically. The things that define the
segments should be stable so that marketing strategy is predictable over time.
Segmentation assumes there will be no drastic shocks in demand, or radical changes in
technology, etc, in the foreseeable future.
Responsiveness. To be actionable, the segmentation must drive responses. If marcom
data is one of the segmentation dimensions, this is usually achievable.

Levels of consumer behaviour

Why would segmentation improve
predictive modelling accuracy?

Examples

Why would segmentation improve
predictive modelling accuracy?

Tools of segmentation
Chapter 14
Metrics of successful segmentation

General analytic techniques
● Business rules
● CHAID

Hierarchical clustering
Hierarchical clustering
calculates a ‘nearness
metric’, a type of similarity
via some interrelationship
variables. There are many
options for how to do this
but conceptually the idea
is that some observations
(say customers) are ‘close
to each other’ based on
some similar variables.
Then a dendogram (a
horizontal tree structure)
is produced and the
analyst chooses how to
divide the resultant
graphics.

K-means clustering
The general algorithm (and as with all other techniques, there are various versions) is as
follows:
●
●
●

●

●

●

Set up: choose number of clusters, choose some kind of ‘maximum distance’ to
define cluster membership and choose which clustering variables to use.
Find the first observation that has all the clustering variables populated and call this
cluster 1.
Find the next observation that has all the clustering variables populated and test
how far away this observation is from the first observation. If it’s far enough away
then call this cluster 2.
Find the next observation that has all the clustering variables populated and test
how far away this observation is from the first and second observations (clusters). If
it’s far enough away then call this cluster 3. Continue with steps 2–4 until the
number of clusters chosen is defined.
Go to the next observation and test which cluster it is closest to and assign that
observation to that cluster.
Continue with step five until all observations that have the clustering variables
populated have been assigned.

Latent class analysis
Latent class analysis (LCA) is a massive improvement on all the above. It is now the
state of the art in segmentation.
LCA takes a completely different view of segmentation. Rather than, as in the case of Kmeans, where the variables define the segments, LCA assumes the scores on the
variables are caused by the (hidden) segment. That is, LCA posits a latent (categorical)
variable (segment membership) that maximizes the likelihood of observing the scores
seen on the variables.
It then runs this taxonomy and creates a probability of each observation belonging to
each segment. The segment that has the highest probability is the segment into which
the observation is placed. This means LCA is a statistical technique and not a
mathematical (like hierarchical or K-meansclustering) technique.
There are some disadvantages of LCA. SAS does not do it, at least not as a proc. SPSS
does not do it either: you have to buy special software.
The advantages have been alluded to but, just to be clear, LCA has a LOT of advantages
(see Table 14.1). Ultimately segmentation’s usefulness is about strategy. The better the
distinctiveness the more obviously a strategy becomes levelled on each segment.

Latent class analysis
The advantages, LCA has a LOT of advantages

Review the Business

Purchase answer to see full
attachment