The Migration Matrix

Recently, there’s been a push to build sophisticated regression models using the latest technologies and techniques. And rightly so! The gains in both processing speed and capacity have enabled statisticians to incorporate large amounts of catalog, retail and Internet data, as well as demographics, to generate results specifying the best names to mail.

A marketing professional is considered remiss if he or she isn’t testing a regression model with various vendors. While these analyses reflect some of every cataloger’s database marketing efforts, there are some areas that commonly are neglected. I’ll discuss one of these: measuring the success of your recent marketing efforts using available data. How do you ensure that you’re effectively managing and mailing your customer file? By looking at what I call “a migration matrix.”

Many catalogers generally look at response rates, average orders, demand per catalog and profitability. But few look beyond those things. If you use data from your marketing database, you can identify opportunities to increase demand and orders by developing programs to increase customer activity, thereby keeping buyers more recent. In the past, when catalogers wanted to increase demand, they usually increased circulation or added pages to their catalogs. Now, with frequent increases in paper and postage costs, you must be more careful about circulation increases by also focusing on retaining current customers.

In addition, mailing smarter boosts your company’s overall profitability and customer lifetime value. Successful strategies include improving your ability to convert single buyers to multi-buyers at a faster rate; increasing the rate of repeat purchases from multi-buyers; and boosting conversion rates of prospect/inquiry names into buyers.

This article will address how to measure the success of your marketing efforts by calculating how well you convert single buyers into multis — that is, how to get customers to go from first base to second.

Steps to Creating a Migration Matrix

If the rate of conversion of singles to multis has remained static or increased during a sustained period of time, you’re doing a good, consistent job of marketing to single buyers. But if the rate of conversion has declined, it’s a red flag to revisit your marketing program’s effectiveness. Following are some steps that can help you create your own migration matrix.

Collect data. First, you’ll need the number of buyers acquired — grouped by the year of their acquisition — over a period of time, say 1999 through 2002. (All marketing databases should include the year of first purchase.) This data gives you a denominator; create a marketing statistic by filling in the numerator.

In addition, your marketing database should have the date of all subsequent purchases. The date of the second purchase is the critical element here, as it’s this, in conjunction with the original purchase year, that enables you to determine the rate at which your customers convert to multi-buyers. Collecting these data may require some programming from your service bureau or your MIS staff.

Determine the productivity relationship between single- and multi-buyers. For as long as we’ve kept track of productivity by recency, frequency and monetary cells, multi-buyers always have been much better to mail to in terms of both response rates and average orders. The chart “Single- to Multi--buyer Improvement” highlights the superiority of multis to single buyers using average actual figures from some well-known mailers. As you can see in the chart, zero-to-six-month multis are three times more productive than zero-to-six-month singles. Clearly, you must push those one-time buyers to buy again.

Create the matrix. Getting back to our conversion rate (or our migration rate), create a table that lets you compare one year to the next. While I’m referring to years as the basis for creating a performance measure, you certainly can use quarters or months as the basis for calculating a metric for comparison.

On the left side of the “Migration Table,” under Acquisitions, I’ve identified the number of customers per year in the cell where the column year and row year coincide. In the cell for the acquisition year of 1999 and the year of second purchase of 2000, you’ll see the number of buyers who came on board with their first purchase in 1999 and who made their second purchase by the end of 2000 (8,250).

Similarly, in the cell for the 1999 acquisition year and conversion year of 2001, you’ll see the number of buyers acquired in 1999 who made their second purchase by the end of 2001 (13,530). This figure must be greater than the prior one, as it’s a cumulative number of buyers with their second purchase.

Devise a similar chart for your catalog. After you plug in the number of acquisitions by year and cumulative number of conversions in each subsequent year, you’re ready to determine the success of your marketing efforts by calculating your conversion rates.

What Do We Have?

The Migration Table shows that the sample catalog has experienced a significant decline: 32 percent = (1 - [22,500/33,000]) in acquisitions from 1999 through 2001. The decline stopped in 2002 as acquisitions stayed at 22,500 for that year.

But the rate at which this company converted singles into multis has steadily declined from 25 percent in 2000 to 22 percent in 2002 — a cause for concern. Total acquisitions decreased during the past four years, and the rate at which single buyers are converting into multi-buyers also declined.

In the second year after acquisition, the rate at which single buyers converted also declined, in this case by 2 percentage points (a drop from 41 percent to 39 percent). So only 15 percent (39 percent minus 24 percent) of the buyers acquired in 2000 converted in 2002, the second year after their first purchase. This is down from the 16 percent (41 percent minus 25 percent) rate from 1999 acquisitions in 2001.

What does this mean? Based on the data, the quality of prospects could’ve declined. It could be the way these prospects were acquired. How were they promoted? Was there a discount offered on a first purchase when the prior year’s acquisitions weren’t?

Many catalogers are reluctant to offer discounts to prospects, because such offers can condition new customers to buy only when there’s a sale. As a result, housefile performance will suffer and consequently, you’ll find yourself in a vicious cycle of having to frequently promote just to stay even with your past performance. Productivity-wise, this is troubling, since discounts hurt your gross margins and put pressure on profit and loss contributions.

Nevertheless, when measuring your marketing success, keep track of the number of contacts you’ve made to your housefile’s single buyers. If you’ve increased or decreased the number of contacts, this will affect how your migration rates have changed from year to year.

A migration matrix is a good way to keep your finger on the pulse of your catalog’s marketing program. Perhaps you’ve had less contacts with customers than in prior years, or your merchandise isn’t as fresh as before. It also could mean your creative needs to be reviewed, or your lists aren’t working as well as they previously did. Or it could be because your page counts have dropped, reducing your catalog’s product offerings. While the migration matrix doesn’t tell you what happened, it does alert you to a potential problem that needs attention and action.

Michael I. Grant is president of Michael I. Grant & Associates Inc., a direct marketing consulting company based in Scarsdale, NY. He specializes in catalog marketing, circulation, database and analysis issues. He has been an independent consultant for more than 11 years. Grant wrote this article at the request of the Catalog Success editors. He can be reached at (914) 722-4177, or via e-mail: michaelgrant12@optonline.net.