|
The Merriam-Webster site (www.m-w.com) defines loyal as "unswerving in allegiance."
People are unswervingly loyal to their utility companies due to lack of choice. People are
unswervingly loyal to their life insurance companies through a lack of interaction. But there
are very few industries where absolute, unswerving loyalty can be hoped for.
The customers with the greatest lifetime value are generally those who are not only loyal to
your products, but also loyal to your company. They are the ones who are willing to promote
your firm and act as references to other prospects. They run user groups and fan clubs. They
tattoo your company logo on their bodies.
You know you have a strong brand when a significant portion of your customer base tattoos
your logo on their chests and forearms. Just ask Harley Davidson.
On the Web, loyalty also refers to site visits over time. People who visit your site more than
once a week may be more valuable than those who visit once a month. Prospects who are
considering the purchase of an automobile are more likely to buy if they show up on a daily
basis for a month. But there is no expectation they will return month after month. Once the
buying decision has been made, they usually will stop visiting car sites altogether.
"Loyalty to me is repeat customers—who's visiting our Web site the most, who is actually
purchasing product, data, or whatever..." (Technology company)
Loyalty, then, means very different things to different sites. Those selling advertising space to
generate income are completely dedicated to increasing the number of site visits and page
views. Yahoo!, AltaVista, MSN, and other portal sites are constantly looking for ways to get
people to come back more often and stay longer.
"We're starting to look at defining loyalty in terms of our promotions. Are our customers loyal
because of coupons or are they loyal because of the brand?" (Retail company) Loyal
customers come back frequently, buy often, recommend your company to others, and readily
try out new things. They may even come looking for products or services that you do not offer.
Perhaps a customer is a regular buyer at BarnesandNoble.com, but does she also buy at
Amazon.com? What if BarnesandNoble.com started selling travel? How quickly would she
start buying travel from them? She might be a loyal book buyer, but will she be a loyal CD
buyer as well? Will she be devoted and buy just about everything she can from Barnes &
Noble? Will she recommend Barnes & Noble to her friends?
Recency, Frequency, and Monetary Value (RFM) analysis helps to answer the most
fundamental question in database marketing: Who are my best customers? Using past
transactions, each customer is viewed simultaneously in three different dimensions:
• Recency. Has the customer made a purchase—or visited your site—recently?
• Frequency. How often has the customer placed orders—or visited your site—
historically?
• Monetary value. What is the customer's total spending and profitability?
Each dimension provides a unique insight about a customer's purchasing behavior:
• Recency. Decades of statistical analysis have shown that customers who have
made a purchase recently are more likely to purchase again in the near future.
• Frequency. Frequent purchasers are likely to repeat purchasing into the
future.
• Monetary value. Customers with high spending in the past might spend again
in the near future. This dimension is different from frequency in that it
identifies customers who place infrequent but high value orders and, therefore,
could be highly profitable.
Dividing customers into a number of segments using RFM-based clustering methods helps
identify and profile customer segments that are not intuitively obvious or visible from reports,
yet represent significant opportunities.
Recency
When was the last time a particular visitor came to your site. Does that have a bearing on their
loyalty? Oh, yes.
From the e-metrics white paper:
Recency is a core measure that describes how long it has been since your Web site recorded a
customer event (e.g., visited the site, purchased a product, etc.). Recency is generally
considered to be the strongest indicator of future behavior. According to RFM, the most likely
users to purchase tomorrow can be readily calculated from past experience. A loyal luggage
buyer may buy a suitcase once every three years. Milk, bread, and egg buyers tend to shop
weekly. When browser-based cookies first came on the scene, they were used to welcome
people back and let them know how the site had changed since their last visit. More than just
a parlor trick and more than just a convenient way of keeping people up to date, knowing
when somebody was last at your site is an important part of user profiling. As recency
diminishes—as the time since the last activity or event increases—the potential for future
purchases decreases. Eventually, a predetermined amount of time lapses and the user has
officially attrited. In an attempt to reactivate customers, different offers might be targeted to
different users as recency fades. If you have shopped at Amazon.com with any regularity, you
may have received a "we miss you" gift certificate. The Amazon system notes the consistency
of your visits and purchases and sends off an email enticement should you fall outside of your
normal purchasing pattern.
Frequency
From the white paper:
Users may visit hourly, daily, weekly, monthly, or less. Here are three scenarios where
frequency means different things to different sites.
The Retail Experience
A user who only comes to a florist Web site four times a year may be considered a very loyal
customer. A wedding anniversary, Valentine's Day, a spouse's birthday, and Mother's Day
are the major flower-giving occasions. A one-time-only user can be encouraged to come back
for another holiday as can the user who only comes twice a year. But the user who comes four
times needs special enticement to increase his or her frequency. A dollars-off coupon, a
bouquet-of-the-month club, or a "buy ten, get one free" offer may all appeal. Offers can be
tested on each level of frequency to increase response rates.
The Considered Purchase
Deciding on the acquisition of an expensive item creates a decidedly different rhythm of site
visits. The occasional click-through at the start of the process gives way to a steadily
increasing number of visits up to the moment of purchase. If these traffic patterns are
properly modeled, they can lead to a clear indication of when the sale may occur. With this
knowledge—and some clever data mining techniques— a company can build predictive
models to be more proactive, launch opt-in email campaigns, dynamically alter the site, or
have a salesperson call on the prospect. Manufacturers can use information about the
frequency of visits to notify their distribution chain about potential sales. Service
organizations can watch customer activity to determine the right moment to up-sell and crosssell.
Training departments can track frequency to decide when to offer additional courses.
The Business-to-Business Bond
Frequency becomes even more important when the relationship between parties is
longstanding. When extranets are used in the place of electronic data interchange (EDI), the
pattern of visits and orders can be very telling as Web site traffic becomes the pulse of the
buyer/seller relationship. If a steady customer with a predictable pattern of visits changes her
browsing and buying habits, it is a good sign that human intervention can increase the
spending potential. Frequency information can yield insight into a customer's displeasure,
expose a shift in customer personnel, or signal the possibility of increased business.
Monetary Value
From the white paper:
The monetary value of a Web site user can only be estimated until a purchase is made. The
user who comes once a day for a week is assigned a much higher probability of purchasing
than one who comes once every three months. As soon as a user becomes a customer, actual
monetary value can be derived from spend and profit margin data. Over time, how much does
the customer buy per month? How profitable are those sales? What are the characteristics of
a high spender versus a low spender?
Clearly, different Web sites will have different indexes for purchase probability and
profitability. But historical ratings of actual customers are of great value when spread across
the users of a single site. These are the figures that help sites recognize which users are most
likely to become profitable buyers.
Later in the white paper, we discussed how the measurement of these factors would be
different, depending on whether you were running a site aimed at a quick purchase, a lengthy,
considered purchase, or a business-to-business long-term relationship.
The Retail Experience
Customer loyalty here is measured in purchases. How much do they buy? How often do they
buy? Are they a profitable customer? The formula for loyalty will include the following
variables:
• Visit frequency. Scored based on number of visits per month.
• Visit duration. Scored based on number of minutes per visit.
• Visit depth. Scored based on number of page views per visit.
• Purchases per visit.
• Number of items purchased per visit.
• Total revenue of purchases per visit.
• Profitability of purchases per month.
If additional marketing programs are implemented, the customer might be evaluated on
factors such as:
• Number of referrals per month: Did the customer refer others?
• Value of referrals per month: Did those referrals buy? How much?
• Questionnaire propensity: How willing is the customer to answer survey
questions?
• Contest participation: How willing is the customer to participate in contests?
• Reward points program: How willing is the customer to participate in affinity
programs?
The Considered Purchase
Loyalty can be measured on a short-term basis to try to clinch the sale. It can also be
measured on a much longer-term time scale.
If the customer is buying a refrigerator, chances are excellent that she will not need another
one for years. You can keep her in your database for those years while waiting for the right
opportunity to remind her of your quality and value. Insurance companies keep information
on newborns in their database for decades in order to offer additional auto insurance in
fifteen-and-a-half years.
Most loyalty calculations will revolve around how the user browses your site. Here we begin
with the same variables as above in the Retail Experience, but with a twist:
• Visit frequency. This is scored based on visits per decision period and mapped
to a decision-making curve. Buyers of one type of product will visit a certain
number of times in the first period, a certain amount in the middle of the
process, and signal that a buying decision is actively being made when they
increase (or decrease) to a different number of visits in a set time span.
• Visit duration. Scored per session. This is another indication of how close to a
decision the user may be.
• Visit depth. Page views per visit are as revealing as frequency and duration.
To this list, we add the important analysis of how the user traverses the site:
• Site path. How well is the user following an optimal site path?
• Contact. How often does the user send email, engage in a chat session, or fill
in a form on the site? What sort of questions does the user ask?
• Product configurator. How many times does the user run the configurator and
which features are selected?
The Business-to-Business Bond
The metrics change again when it comes to extranets. In the business-to-business
environment, the emphasis on selling is replaced with a focus on service. Taking orders and
solving problems are paramount, while a less aggressive eye is kept open for up-selling and
cross-selling. Loyalty comes in many shapes and sizes, and hence loyalty metrics naturally
tend to differ greatly between different sites with different business models.
• Visit frequency. In this environment, the watchword is consistency. Is the user
coming to the site at set intervals and doing what is expected?
• Visit duration. Are there any changes in the amount of time it takes the user to
place the order?
• Visit depth. Is the user looking at products above and beyond his or her norm?
• Visit tenure. Time elapsed since first visit.
• Purchase tenure. Time elapsed since first purchase.
• Purchase frequency. Number of purchases per quarter (or month).
• Total lifetime spending. Total spending since first visit.
• Visit recency. Time elapsed since most recent visit.
• Purchase recency. Time elapsed since most recent purchase.
• Required clicks to first purchase. Minimum number of clicks required to
complete the first purchase in a visit. The first purchase may require more
clicks than repeat purchases.
• Required clicks to repeat purchase. Minimum number of clicks required to
make a repeat purchase.
• Actual clicks to first purchase. Actual number of clicks until the first purchase
was made.
• Actual clicks to purchase. Number of clicks until a repeat purchase.
Tracking where they went is part of the loyalty factor; the other part is tracking what they do. | 
Disclaimer
