CHAPTER 2

Planning a Customer Analytics Initiative

Prepared by:
Dr. Rubi Ana A. Saludario
In This Chapter

1. Outlining the scope and goals of an analytics initiative

2. Choosing metrics, methods, and tools

3. Setting budgets and sample sizes

A Customer Analytics Initiative Overview
Before you start on your initiative, keep these four things in mind:

✓ Access to the right data: It’s hard to increase the frequency of customer
purchases, conversions, or attitudes if you don’t know what customers are
purchasing, when they are purchasing it, or what they are thinking. Be sure you
know that the data you need exists, or that you’ll be able to collect and analyze it.

✓ Customer level data: To do the most with customer analytics, you’ll want to
gather data for each customer, not aggregated data at product or company levels.
✓ Analytics that focus on the customer: The “right” analytics depend
on the method. But one thing that all good customer analytics have in common is that
they are meaningful to the customer. Just like airlines should care more about on-time
arrivals than on-time departures, your analytics should be felt by customers at all phases
of their journey. If you want to improve the customer support experience, customer
satisfaction with the call outcome is a better metric than the number of calls answered
in an hour. The latter is an example of company centric and the former is customer
centric.
✓ Getting buy-in: Planning, collecting, and analyzing data is only good if something is
going to be done about the insights. All too often, I see organizations spend a lot on
research and customer measurement projects but the results stop at the executive
presentation meeting. Unfortunately, insights aren’t acted upon because the people who
can change the product, price, or experience aren’t involved with the data collection and
planning.
• Customer analytics should be shared with not only executives, but also with product
development, sales, and support staffers. As part of the planning and getting buy-in,
be sure the analytics will cross customer touchpoints and be accessible across the
organization.

• Not all customer analytics initiatives lend themselves well to the more systematic
methodology I include here. So don’t try to force extra steps or complexity just to
meet this framework.

• A lot of creative thinking goes into making a plan. Don’t feel like you need to fit every
project into this process. That’s especially the case if your goals and methods are
narrow in scope. The rest of this chapter covers the details of putting together your
customer analytics initiative.
Defining the Scope and Outcome
The first stage is goal setting: where you define the scope and outcome of your project.
Don’t overlook or rush through this stage. Collecting customer analytics takes time and money.
You can easily exceed your budget if you compiled the wrong data.

1. State the goals of the initiative.

Think in terms of the intended outcome (for example, an increase of 10% in revenue of a
product line over the next year). The more specific you can be, the more attainable the
endeavor.
2. Write down the questions you want to answer.
Data is meaningless unless it’s collected for a reason. Articulate what business questions
you’re hoping to answer. Avoid being vague and large in scope. Start small and specific
and itemize your questions.

You want to be SMART: Specific, Measurable, Attainable, Realistic, and Timely.

Some examples of questions customer analytics can answer include:

• Which product feature should I add to this product?

• What is preventing customers on the website from completing a purchase?
• What labels should I change in the website navigation?
• Why are customers not recommending a product and how can I improve positive
word of mouth?
• What percent of high-income mothers are aware of the brand and website?
• Who are the most profitable customers?
• How long until a customer makes a repeat purchase?

There’s a good chance you aren’t the first person to collect and use customer analytics in your
organization. Look for past initiatives, past projects, and what worked and what didn’t work.
The documents, results, and people involved in past initiatives will save you a lot of effort and
prevent you from reinventing the wheel.
Identifying the Metrics, Methods, and Tools
During this step, identify the metrics and methods you’ll use to answer your questions
and achieve your goals:

• Look for metrics that are meaningful to customers. Think on-time arrivals instead
of on-time departures. See Chapter 2 for ideas on the right data to collect.

• Identify what tools you’ll need for data collection.

• Consider collecting customer data by surveying existing customers. Even

something as simple as surveying customers requires several inputs.
Table 3-1 shows how you can go from question (from the preceding section) to metric, to method,
and finally to the right tools for the two examples of customer loyalty (see Chapter 12) and
findability (see Chapter 15).

 
You also need to understand your baseline scores. It’s hard to know if you’ve improved anything
if you don’t have a baseline measure.
 Access to customer data: This is often guarded in organizations because it contains both
sensitive company and customer data.
 Transaction data at the right level of detail: Total revenue by product
is often at too high a level to understand what’s driving purchases. In many cases, you want to
obtain customer transaction data at the product level. This way, you can understand who
these customers are (demo- graphics and so forth), when they made the purchase, for how
much, and how often (for repeat purchases).
Some of the most important insights companies gain from their customer analytics comes from
merging survey data with transactional data. One of the biggest challenges is being able to
properly match customer survey results with past and future transactions. You may need the
help of a database administrator or IT person to be sure you can merge survey data with trans-
actional data.
 
Setting a Budget
Every project requires a budget whether it’s large or small. Consider the following as you prepare
your project:

 Software: If you have millions of customers and as many monthly trans- actions, you’ll benefit
from sophisticated software that integrates into sales and accounting systems. Products offered
from SAS, IBM, and Oracle can cost upwards of six figures to implement and service. Some of
the best insights still come from simple calculations in Excel or a calculator. Don’t think you
need to wait to get approval to purchase expensive software to begin making decisions from
customer analytics.
 Time: You can spend a lot on the hard costs of software and services as well as the softer costs
of employee time. Software packages that can fit the budget for single-person companies to
the largest enterprises are available. Throughout this book, most methods and analysis can be
con- ducted with Excel, free web software, or options that don’t require very expensive
software.
 
While it’s easy to calculate the cost of purchasing software and services or hiring
additional employees to handle an initiative, be sure to consider the cost of business
as usual:

1. How many customers are defecting to a competitor?

2. What percent of customers are not returning?

3. How many customers are discouraging others from using your product or
service?
Determining the Correct Sample Size

• If you aren’t measuring every transaction or surveying every customer, you’ll

have to deal with the uncertainties that come with sampling a portion of your
customer population. Even if you sample everyone, you’ll likely want to make
estimates about future customers or future transactions, and to do so, you’ll still
have to deal with the uncertainty.

• In general, it costs more money and takes more time to either sample or analyze
data from a large database, so you must put some thought into how large of a
sample size you need.
I include ways of coming up with the right sample size for each method in each chapter. In
general, you should consider two important concepts when planning your sample sizes:

 You need larger sample sizes to detect smaller differences.

A new design, promotion, or feature may improve customers’ attitudes or sales, but if the
increase is small (something like a 5% increase), unless your sample size is large enough, that
difference won’t be distinguishable from random fluctuations in the data.

 For very large sample sizes, almost all differences will be statistically significant.
Statistically significant essentially means that the differences are not likely due to sampling
error. However, statistically significant does not mean the findings are noteworthy or
important. Will customers notice a one-second reduction in the time it takes to rent a car
online? Probably not. Although it’s a good idea to drive increases in positive attitudes and
sales, watch out for spending a lot of effort for little return. See the appendix for more of a
discussion of statistical significance.
Analyzing and Improving
In the analyze phase, you want to be able to describe the current state of the customer, often
by segment (see Chapter 4), and identify the root causes of problems or insights to make
improvements.

For example, if you’re measuring customer loyalty, you’ll have customers’ likelihood to
recommend scores, satisfaction ratings on other parts of an experience, and open-ended
comments. From this you can

 Compute the baseline Net Promoter Score

 Examine the open-ended comments for the reasons for ratings
 Conduct a key driver analysis to understand which features or attributes of the experience
are having the biggest impact on customers’ likelihood to recommend. See Chapter 12 for
more on this approach.
During the improving stage, you use the data collected in the measure phase to
show quantifiable improvements (or reductions) in your metric. This can include

1. Conversion rates (converting more browsers to buyers online)

2. Improvements in customer attitudes

3. Increases in revenue

4. An increase in the number of repeat purchases

One of the worst things that happens is collecting analytics that clearly show a
problem with the customer experience but then doing nothing about it.
It can frustrate customers and lead to analytics initiatives that were just an
exercise in measuring.

 For some methods, the improvement process is built into the measure step.
For example, when conducting an A/B test on a website (see Chapter 10), the
improvement has to be built so implementation is easy.

  In other cases, making changes is much harder. If you find that one of the
primary reasons customers aren’t repurchasing or recommending your
product or service is because of price, making changes to the price usually
involves corporate politics, entrenched ideas, and shareholders.
Controlling the Results
A lot of effort can go into planning and collecting data. You can prevent waste and
rework by putting in place systems to reduce the time between data collection and
action. Some things to consider are

1. Automatic reporting: Look for ways to output key dependent variables to

executive and team dashboards or scorecards that provide real-time insights into
customers’ experiences.

2. Access to data: It can take a long time to jump through corporate hoops to get
access to purchase history or customer details. Have a documented process that
others can follow and a way for your employees and data scientists to get
creative with the data.
3. Putting the right people and procedures in place: You don’t want to have an
initiative that’s entirely dependent on the knowledge of one person. Be sure to
document

• Procedures for how to get access to the data

• Methods that go into analyzing it

• People who are responsible for the analysis and decision-making

• Who is impacted by the results