Professional Documents
Culture Documents
Arthur M. Hughes
Database Marketing
And the Web
Introduction
“You can not make money from transaction buyers, the money is in the
relationship buyers”.
Marketers are faced with a dilemma on the approach. Should they base
their message on the price (a left brain argument) or on the benefits of the
product (which appeals to the right brain)?
It is well known in retailing that “buy one, get one free!” out-pulls “50
percent off!”
Price and quality have been the staples of marketing for years. Today,
there is a new dimension entering into consumer decision-making that is
of equal importance. It is time. People have less and less time available
for shopping (or for anything).
What we marketers are selling today, therefore, is more than the product.
It is the product, plus the delivery method. The convenience and service is
part of what we are selling.
The numbers a, b, c, and d are weights which vary with each customer.
“In the 1970’s, across the Texas border in Mexico gasoline cost much
more, but there were no lines and unlimited gas supplies. For a number of
months, Texans drove south of the border to fill up their gas tanks at high
prices, while Mexicans came north across the border to wait in lines for the
cheap gas”.
Database marketing and the Web came along just when they were
needed. They solved the information problem; the provided recognition,
personal service, and a profit to the customer. The web is an extension of
that database, providing recognition and helpful information.
Database marketing and the Web are aimed at the customer’s right brain.
Instead of being bombarded with discounts, the customer is showered with
attention, recognition, friendship and service.
Database marketing and the Web are the only way to start a two-way
dialog in which the customers are able to tell you what is on their minds,
and you are able to react to their thoughts by varying your services and
product mix.
PART
Marketing
Strategy Development
3. Lifetime Value – The Criterion of Strategy
“The only way to grow a business is to get customers to come back for
more and tell their friends” – Enterprise Rent-A-Car’s CEO Andy Taylor
Lifetime value is the net present value (NPV) of the profit that you will
realize on the average new customer during a given number of years.
The lifetime value is simply the Cumulative NPV profit in each year,
divided by the original group of customers: LTV = CUM NPV / acquired
customers.
“If you are selling software the customers are likely to tie up you customer
service lines during the first 60 days until they learn how your software
works. For the next 60 months, you may never hear from them again”.
A customer relationship building strategy can affect five (and only five)
basic things:
1. Retention rate
2. Referrals
3. Increased sales
4. Reduced direct costs
5. Reduced marketing costs
“Research shows that referred people are more loyal, have a higher
retention and spending rate than the average new acquisition”. In your
database you put the ID number of the referred person in the referrer’s
record, and vice versa. Research shows that those who refer other
customers are also better customers. They are advocates. They spend
more and are more loyal.
When a business is sold, goodwill is often what the buyers pay the most
for. Goodwill is nothing other than the value of the customer base that the
company has built up over the years, and currently is holding on to.
Compute lifetime value for each of several years, use the period of time
that makes the most sense to you based on your particular product
situation.
What is the annual retention rate if 50% of the customers buy again after
four years? Here a formula is necessary. The formula is:
RR = (RPR) (1/y) RR is the annual retention rate, RPR the repurchase rate
and Y the number of years between purchases.
4. Designing a Successful Customer Strategy
Sometimes ‘losers’ have temporary reasons why they are not profitable:
- They may actually be doing a lot of business with competitors
- They may be small, but fast growing businesses
- They are buying the wrong products from you
- They are true losers, transaction buyers who only buy when you are
on sale. Give them your competitor’s number!
There are two types of facts that you can learn about customers: who they
are (demographics) and what they do (behavior). Accurate behavior
predictions are important for making profitable marketing decisions. The
best predictor of future behavior is past behavior. RFM is pure behavior.
Recency: to code your customer base for recency, you need to store one
vital piece of information in every customer’s database record: the most
recent purchase data. Then you divide the databse into five exactly equal
parts (quintiles), which you number from 5 (most recent) down to one
(most ancient). More recent customers will respond better than ancient
customers. Recency is a powerful predictor of response. Everyone in the
world feels a rush of enthusiasm when he purchases a new product or
service. The feeling lasts for a while.
There is another way called ‘hard coding’. Using hard coding, the top
quintile is set as being some arbitrary date range, such as 0-3 months. The
next is 3-6 months etc.
If you are a telephone company or a bank, recency cannot be the last time
the customer made a phone call or paid a bill. Recency is the last time
they changed their service or bought an extra service.
The 1’s on a frequency graph respond better than you would suppose,
because they contain of a lot of recent (new) buyers. A customer who just
joined you yesterday is your most recent buyer. If you lowest quintile on
frequency is not higher than the trend, then you probably have done
something wrong.
Monetary: same methods as above, but now divide the same people by
their monetary spending. Divide by the total amount spent on our
products or services in total, by month, year or in some other way.
Monetary coding is far less predictive of behavior than either recency or
frequency.
The million-dollar customer would be less likely to ‘open the envelope’
because he is on everybody’s list. But the thousand-dollar customer who
opens up will have to think it over to spend (or get permission). The
monetary graph is a combination of two opposite human emotions:
willingness to open the envelope and ability to pay.
Selecting a Nth:
Select a test group using an Nth. An Nth is a test group that is an exact
statistical replica of the full database. For example, if a database had
300,000 customers and we wanted to select a test group of 30,000 using
an Nth, we need to select every tenth record.
Good graph of a test using the RFM cells in calculating break even
in Figure 5-8 on page 101.
Break-even
Break-even in direct marketing means that the net profits from sales to a
test group exactly equals the cost of promoting that test group.
There is software that can help that is called RFM for Windows. You can
download it for free from www.dbmarketing.com
Loyal customers:
1. Have higher retention rates
2. Have higher spending rates
3. Have higher referral rates
4. Have a higher lifetime value
5. Be less expensive to serve
6. Buy higher priced options
Loyalty-building Communications:
- Customer preferences: today its increasingly done on the Web
- Caller ID and cookies: Caller ID permits you to know who is calling
before you answer the phone, the same goes for cookies on your
website.
- Event-driven relationship messages: these are not sales pitches.
They are expressions of friendship. Use trawling to trigger events
like birthdays, making big purchases or an anniversary as a
customer with your company. Send kind notes on the events!
- Event-driven marketing messages: look for what they are
searching for on your website and communicate it as ‘product of the
month’. Record what they are asking for at customer service and
send personalized messages.
Loyalty programs:
The card does two things: it provides the customer with some benefit he
could not get without the card and it provides you with valuable
information that you can use to understand the customer and build
loyalty. The registration is the entrance to the world of loyalty marketing.
“We must water what we want to grow. But first we must decide what we
want to grow…All customers are not equal. Behavior usually follows
rewards”.
Wealthy customers are more loyal. Even more influential than age,
gender, or geography, household income proved to be the most indicative
of the strength and impact of customer loyalty.
Loyalty-based companies should remember three rules of thumb:
1. Some customers are inherently predictable and loyal, no
matter what company they are doing business with. The simply
prefer stable, long-term relationships.
2. Some customers are more profitable than others. They spend
more money, pay their bills more promptly, and require less service.
3. Some customers will find your products and services more
valuable than those of the competition. No company can be all
things to all people. Your particular strengths will simply fit better
with certain customer’s needs and opportunities.
Strategy:
- Communications to the segment
- Rewards designed to modify the behavior
- Controls to measure the success of the strategy
- A budget for implementation of the strategy
- Specific goals and metrics for the engagement
- An organization that accepts responsibility for the segment
Action plan:
- Roadmap
- Budget
- Standard application of the segmentation
- Responsibilities for the segment
- Specific goals with milestones
If your program is not working, it may be because your are not offering the
correct rewards.
If you mail only to the 50,000 who the model identifies as the most likely
responders, you should get a response rate of more than 2 percent –
perhaps 3 percent or more. This will be much more profitable for you, and
avoid bothering people who are not interested in your product.
“You can purchase compiled names from list brokers based on the model,
in that way you only pay for records that have a high ‘buying’ score”.
“Banks have a lot of data about their customers. They can use their
database to examine customers who have only checking accounts. From
the data they can determine whether our next product should be a home
equity loan, an auto loan, a savings account, or mutual funds. How do the
do this? By using a model to see what thousands of their depositors have
purchased mutual funds look like in terms of behavior and demographics,
and how the differs form those depositors who have been offered mutual
funds but have not bought them. The model is used to score all their
depositors. Those depositors whose scores resemble the mutual fund
buyers are likely targets for a mutual fund promotion. Mutual funds are
their next best product. The next best product is put into the customer
record and called up on the tellers’ screen when you come in to cash a
check or make a deposit. It is also used in mailings to depositors”.
“In the US recently, the churn rate in telecom was 2.9 percent per month,
or 35 percent per year”.
The segment in the tree diagram can be shown in a “gains chart”. The
gains chart is a handy tool for seeing what levels of expected profitability
would result from going increasingly deeper into a prospect file. CHAID
helps to create market segments. The tree diagram predicts the
performance of each segment.
10. Customer acquisition
Types of lists:
- Response lists consists of customers who have made a purchase
ore responded to an offer.
- Compiled lists consist of nearly all the names in a country
(consumers)
Choice kills: You should not give recipients a choice. You will find that
your response rate to a specific offer beats your response rate to an offer
about which the recipients has to make a decision.
If you use e-mail addresses from opt-in lists, there are some rules that you
should follow:
- Always identify yourself clearly in the address line
- Provide a physical address, a URL and a phone number
- Explain how you got the persons name and address
- Provide a simple one-click opportunity to have the person’s name
dropped from you mailing list
What kind of response rates will you get?
Direct mail response rates are seldom very high. For most mailings, a 2
percent response rate is normal. Some mailings have achieved response
rates of 10 percent or 20 percent but that is very rare. The Direct
Marketing Association (www.the-dma.org) provides a report on average
response rates in various industries.
“The holy grail of direct marketing is the single-variable test. You want
only one thing to change in each test. If you’re going to test price, then
you test two packages that are the same in all respects except for price.”
The first step in any testing program is to determine what you are tying to
accomplish. The goal of database marketing programs is usually to:
- Increase sales to existing customers
- Reduce attrition
- Gain new customers
The best method is to define your customer lifetime value, and test the
effectiveness of various alternate ways of increasing lifetime value.
To set up a test you need two groups of customers: a test group (who gets
the offer) and a control group (which does not get the offer). Without the
control group, you will learn very little. Control groups do not have to be
the same size as test groups. They could be larger or similar. They have to
be large enough to give valid results, and they have to be exactly the
same type of people in the test group. When the groups are set up, a
promotional offer is made to the test group. The control group is treated
like everyone else. Their purchase behavior is compared in the marketing
database.
Follow your groups over a longer period of time, because the testing group
will move up higher in RFM cells and there will be positive effects long
after the promotion will be over. Good testing programs will follow the test
and control groups for the following 12 months to determine the residual
effects of the test. In some cases the residual effects can be eve more
important than the initial response to the promotion. Lifetime value, rather
than the immediate short-term payoff, should be the real goal of
marketing database strategy.
“To use the Internet to generate leads for agents, you data strategies
must focus on qualifying the leads generated, so that agents can handle
the increased volumes. If you can improve the conversion rates with
demographics, it’s worth any amount of money”.
On the whole, the aggregate web sales numbers are still small. So why is
the Web important as a marketing channel:
- Affluent people: web shopping people have higher incomes
- Purchasing research: many web users do their purchasing
research before they go out and buy
- Reviews: once people start making a habit of looking up local
retailers before they venture forth, the reviews will become a part of
their lives.
“In designing your site, think like a customer. Do not think what you want
to sell”.
Two tips:
- Let customers register without hassle
- Provide live help (e.g. www.liveperson.com)
Jacob Nielsen point out that transactional e-mail has three goals:
1. Avoid being mistaken for spam
2. Be a customer service representative
3. Prevent customers from calling in
“Good e-mail should respect users’ time and quickly tell them what they
need to know. Start with the information that matters most to users”.
Trawling: looks at ever record in the database which has been changed
since the last update and applies business rules tot the data it finds. For
example it will determine if the customer:
- Is about to have a birthday or anniversary
- Had an unusually large transaction
- Reached a milestone with the company in terms of total sales, years
with the company etc.
Each hit during trawling has a programmed marketing response. These are
automated communications.
This chapter was not very interesting. It shows the author has a little bit
lack in knowledge about modern Internet marketing.
PART
Profiting
by Experience
13. Retailing and Packaged Goods
“We are no longer trying to take customers away from our major
competitors. Our focus is to make money who are already shopping at us”.
“We are concentrating on our top customers, not our top merchandise.
It is more profitable that way”.
There are several reasons why packaged goods database projects seldom
work out on the web:
- There is not enough margin to do database marketing
- You cannot find out what the customer is doing
- Coupons have seriously eroded brand loyalty
“If you feel that I am warning you to go slowly and carefully before you
rush into database marketing for packaged goods, you have gotten the
message”.
14. Building Retention and Loyalty in Business Customers
“There are many products and services where database marketing has
limited value. In business-to-business marketing, however, relationship
building activities always pay off”.
By classifying your customers into these four segments, you can focus
your marketing efforts on the one segment that is really profitable, the
relationship buyers.
You already know the bargain hunters. Program buyers can almost be
ignored. They make small purchases on occasional basis. Your big problem
lies in distinguishing transaction buyers from relationship buyers. You can
identify the transaction buyers by a small survey.
Ask one question: How important are the following in making your decision
about where to buy this product? Rank from 1 – most important to 5 –
least important. Those who code price as most important are probably the
transaction buyers:
a. Price
b. Service
c. Reputation
d. Recommendation of friend
e. Company policy
f. Previous experience
g. Customer service
The first step in an RFP is a definition of the work to be done. There are
several parts to a scope of work:
- Background: company, products and its customers
- Problem: you would not be writing this RFP if there was not some
problem to be solved
- Solution: what do you think the outsiders can do to help? Give your
ideal answer
- Goal: a one-paragraph statement
- Strategy: the big picture and how this project fits into it
- Customers
- Size of the work: how big a solution are you looking for right now?
Don’t think to big, start with small successes.
- Project phases
- Quantitative measures: give your prospective partners some idea
of the numbers they are dealing with
- Organization: who is involved in the process
- Timing
- Pricing scheme: give them some sort of numerical measurements
to price out; otherwise you will be comparing apples and pears.
- Budget: if asked for it, only provide a range
- Competition: provide a small test to let them show what they can
do
- Evaluation criteria: criteria including any weighting that you want
to include
- Confidentiality agreement
Rules of the RFP:
- Due date
- Questions (where and how to ask)
- Digital submission (allowed or not)
- Where to send
- Extensions (will extensions of time be permitted)
- Bidders list (sometimes it is a good idea that all bidders know what
they are up against)
“In our business, you have to have a champion at the top, or there will be
no database marketing program. This is the story of corporate America.
You lose your champion, you lose your program”.
(Case: OMC, manufacturer of the Johnson & Evinrude outboard motors)
“Just as a CFO can look at a balance sheet and an income statement and
determine a profitable direction for company activity, so can an alert
marketer look at the information stored in a customer database, and learn
of profitable marketing opportunities. Such knowledge in the hands of a
skillful marketer is market power”.
Steps to knowledge:
- Build a customer marketing database with complete purchase
history and demographics
- Develop an active marketing program
- Build a website that involves the customers
- Determine LTV
- Use the database and the website to analyze marketing activities
- Use the knowledge to obtain sufficient resources