Transcript Slide 1

Digital Marketing
Measurement
MKTG 437
Scott W. Flexo, Ph.D.
Why Measure?


Every decision we make as marketers is
based on what we know about the target
market.
Decisions about product/brand positioning,
product or service configurations, media
placement, ad content, promotions, etc.
are all based upon our knowledge of the
consumer target.
Why Measure?


Gathering consumer data – called
"marketing intelligence" – are the most
critical first, middle and end steps in the
development and successful execution
of a marketing strategy.
Otherwise, you have "Death-Wish"
marketing – when marketing strategy is
not connect to markets.
Types of Measurement

Primary Data Measurement



Data that is collected specifically for the
problem at hand.
Typically primary data is collected in a
survey using a questionnaire and survey
research methods.
Can also be collected on the Internet
though "digital footprints" the data left
behind by consumers when they are doing
something on the Internet.
Types of Measurement

Secondary Data


Information collected for another reason,
but stored and available for use.
Typically stored internally in a company
database, published on the Internet by a
data aggregator, or available though other
on and offline syndicated resources.
Database Marketing and Cluster Analysis
http://thedma.org/
Comedy Klout
Retailer Klout
Two Forms of Research: Qualitative
and Quantitative

Qualitative Research: Listening and Counting
What is Happening Online

Observational Research: The process of recording behavior
as it occurs, or the residual (footprint) of this behavior.




Offline examples are watching and recording people's behaviors in
retail stores and restaurants.
Online examples include gathering and tracking comments/ posts
made by consumers online.
One of the most often used forms is to count "brand mentions"
online at selected sites (or from social media listening aggregators)
and compare the number, location and shifts in mentions to digital
marketing activities.
http://twitter.com/search/
What is Your Klout?
Observations: Counting
Visitors and Clickthroughs
Another way to use observational research
is to simply count the number of visitors to
a web site or "likes" on a Facebook page
and any "footprint" information that is left
behind.
 This data can be tracked daily and
compared to digital campaign activity.
Google Analytics
Facebook Counter

Forms of Research: Qualitative

Qualitative Research

Observational Research: The process of recording
behavior as it occurs, or the residual (footprint) of this
behavior.





Problems: High risk of subjective bias in the selection of sites to
monitor and data to collect.
Problems: Most WOM occurs "offline" which is not picked up in
this form of online research.
Problems: Online conversation is different from offline
conversation so if just online monitor is used, will yield a biased
picture of the marketplace.
Problems: It is very labor intensive and time consuming – again
most rely on aggregators for this purpose.
Example: Online Listening Vs Survey Results
Two Forms of Research: Qualitative
and Quantitative


Quantitative Research: Measuring
online behavior though the use of a more
formal research design and measurement
process than just simple observation
(listening and counting what you hear).
Quantitative research also involves a
more systematic analysis of data in a
more formal way.
Online Quantitative Research

Employs a formal research design
before data is collected.


Design specifies the plan to collect and
utilize data.
Designs include decisions on methods,
sampling plans, procedures for data
collection and how the data will be
analyzed.
Online Quantitative Research:
Benefits of a Research Design


A research design or plan helps to improve the
precision of data and allows for hypothesis
testing.
Research designs help improve the RELIABLITY
AND VALIDITY of study data.
Reliability: Same measure result will occur each time.
 Validity: Data actually measures the phenomena
Gives researchers and marketers more confidence that
the information can be generalized to the population and
will lead to a higher probability of making correct
decisions.


Quantitative Research:
Systematic Approaches

Sentiment Analysis/Opinion Mining


How people think and feel about brands and
products online. Can replace attitude
surveys if done properly.
The method involves analyzing online
content to determine the attitude of the
writer – notably, whether the online
comments are positive or negative toward
the brand/product.
Quantitative Research:
Sentiment Analysis

How It's Done


Comments/data from previously identified key sources (sources
the marketer believes is relevant to the brand/product) are
collected using "web crawlers"
 Web Crawlers: Applications that move across designated
web sites to collect and store content that they find.
 Many commercial web crawlers are available.
Text Mining Software is then used to filter and
analyze the "sentiment" or content found.


http://text-processing.com/demo/sentiment
Uses a word-phrase dictionary to code the data (sometimes called a
sentiment dictionary)
Counts are then aggregated and summarized
Aggregated Online Sentiment
Counts: HP Computers
Quantitative Research:
SPSS Text Analyzer
Content Analysis




Content Analysis : Similar to sentiment
analysis in that text is coded and counted.
But in this case, a schematic of how people
will responded is created prior to data
collected.
Content is then coded to fit into this a prior
developed "response" schematic.
Most often used in coding open-ended
responses to questions, but has been applied
to analysis of online response texts.
Quantitative Research:
Survey Research



Survey Research: The use of questionnaires
and sampling designs to collect primary data
that can be projected to a population of
interest.
Survey Monkey Online Surveys
Internet offers great opportunities to conduct
online consumer research through the
development of online research panels.
Groups of consumers that complete surveys for
an incentive or consumers solicited to complete
a survey online or after an offline purchase.
www.e-rewards.com
Knowledge Networks
Providers of Social Media Monitor

Online companies have developed
proprietary web crawlers and analysis
algorithms to analyze web traffic and
report this aggregated data
continuously.
Top 10 Social Community Measurement
Tools
Social Media Posts
Twitter Counter
Netbase
Most "Loved" Brands
Learning Activity





Visit www.socialmention.com.
Run an analysis on a brand that is of interest
to you.
Do you agree with the analysis?
Should you be concerned with coverage or
sampling error in the analysis reveled?
What do you think this analysis says and how
will you use it?
HOW DO WE DO MARKETING
RESEARCH ONLINE?
Problem ID/Process
The Key Steps To Successful Measurement
Management

Define: What do you want to happen and
how will it be measured?


Must be connected to a monetized outcome
for after all, the goal is to drive revenue.
Examples of what you may want to happen



Motivating people to want to behave
Influencing people to have positive attitudes
Motivating people to purchase a product
How to Measure:
The Key Steps To Successful Measurement

Define: Requirements


Specific, Measureable, Appropriate, Realistic and
Time-Oriented
Example: We will promote our new Facebook
page in print advertisements in issues of Rolling
Stone and Maxim. On July 15 of this next year
we will count the number of Facebook users
who "like" our brand and compare sales to the
same period last year.
How to Measure:
The Key Steps To Successful Measurement

Define: How Do You Define Success?


ROI: Return on Investment is a measure of
profitability and captures how effective you are in
using marketing capital to generate profits (or
revenue).
 Cost of Marketing – Revenue Gains/Loss
Often costs are connected to intermediate outcomes such as:
 Impressions Model – How many media impressions or
"opportunities to see" were generated
 Social Impact Model – Correlation of a social media event
with sales
 Target Model – Surveys of target consumers to asses the
effectiveness of the campaign among the target market
How to Measure:
The Key Steps To Successful Measurement

#2 Assess: Build a Cost/Benefit Model

Identifying the costs and returns of the marketing
investment in ways that are easily measureable
 This starts with a cost-benefits table:
* Estimated costs of the campaign itemized
* Estimated benefits of the campaign itemized
* Subtract costs and benefits to yield net value of the campaign
Page 212 of SMM provides an excellent example of a typical
spreadsheet used in building a digital marketing campaign cost
and benefit analysis.
How to Measure:
The Key Steps To Successful Measurement




#3 Track and Adjust: Measure and adjust
based on your measurements.
Continually measure and track changes
based on your definitions and metrics.
Produce tracking maps and compare to
marketing activities.
Adjust activities based on tracking results.
Web Measurement: Cautions

Measurement is NOT optional it is required. But
there are cautions:




Coverage Error: Failure to cover all elements of a population in
your measurement. Which websites, consumers, etc., should be
included in your data collection?
Sampling Error: Bias in the selection of cases for inclusion in the
study.
Online Echo Effect: Bias from duplication in conversation
volume that occurs when content is shared automatically.
Spam: Individuals paid to discuss brands and create "buzz" on
the Internet
Measurement Samples
That's It Folks!

Thank You for a Wonderful Semester!
100%
I hope your
interest and
knowledge of
Digital Marketing
increased
exponentially
this semester.
80%
60%
40%
20%
0%
Start
End