David Shepard Associates Blog

About two hours into our first meeting with a potential client we got around to asking The Big Question: what do you do when you get a new customer? And, the possible answers we were expecting to hear were:

(1)  We add new customers to the file and send them the same stuff we send everyone else until their lack of response causes them to drop off the file.

(2)  We put them into a “fixed sequence-new customer stream” then, after about six months to a year, we add them to the general file and promote them according to what our RMF or regression models suggest.

(3)  We evaluate all new customers (how did they respond, what behavior information came along with their response, what demographics do we know) and then place each customer into one of a number of segments, each of which has its own predetermined fixed steam of new customer offers.

(4)  We place each new customer into the most appropriate segment, as above, and then we treat each customer according to his or her response to individual promotions, i.e., there is not a fixed predetermined stream for each customer.

(5)  Before executing strategy 4, we place each customer into either a control group, or into one of a small number of test groups, and then execute strategy 4.

So, what are the pluses and minuses of each strategy? And, which answer were we hoping to hear? (Obviously, we were hoping to hear Strategy 1, so we could recommend doing something better.)

But, the first strategy has a lot going for it, even though it’s intuitively wrong, it’s easy to execute, another words it’s cheap. And, cheap is not necessarily bad. What’s more, anything else you do, including any of the other strategies suggested above, will cost more to execute, and their added costs will have to be cost justified.

Strategy 2 recognizes the need to treat new customers differently from the way you treat customers with greater tenure. Strategy 2 would suggest that might want to touch base with your new customers, perhaps a phone call, perhaps an e-mail, perhaps special offers to get the new customer to get into the buying habit. This might be particularly important for catalog companies, where the number of customers that fail to buy more than once, one-time buyers, is a very real problem.

Strategy 3 begins to sound like modern, one-to-one, relationship building direct marketing. Which means, it sounds right and might actually be right, but could be too costly to execute. So if you go this route you should test this strategy against strategies (1) and (2). Why does strategy 3 sound right? Because it should be intuitive that new customers come to you with different behavior information and possibly different demographic information.

What do we mean by different behavior information?

To start with what acquisition source did they come from (print, direct mail, the Internet, what specific medium)? What did they buy—for book and music clubs, what books or cds did they select; for catalogers, which products did they buy; for fund-raisers, how much did they donate; for mutual funds, how much did they invest? When did they come on the file, seasonality might influence the selection of future offers. And what about demographics? The zip code will tell you a lot, and depending on the medium and the product or service, it might make sense to ask for additional demographic information in the application itself, or in a welcoming package, or over the web if you’ve collected their email address.

If strategy 3 begins to sound like the new direct marketing, then strategy 4 will get us almost there. Strategy 4 differs from strategy 3 because it’s interactive. There is no fixed stream of predetermined offers or promotions. What the new customer receives is a function of what the new customer does. Assuming, as in strategy 3 we use all of the information available to place the new customer into the most appropriate segment, and then we send that customer his or her first promotion, P. If the customer responds to P1 we send them (mail or email) P2, if there is no response, we send P3. If there is no response to P2 we follow with P4, if there is a response to P3, we follow with P5 , and so on. This process can go on for anywhere from six months to a year (in theory it can go on forever) at the end of the process you will have collected enough transaction data to begin managing the relationship differently – depending upon the industry you are in. Which brings us to strategy 5.

How do you know that the sequence suggested in strategy 4, the business rules, if you prefer that term, is correct? Who says what the elements of P2 or P3 should be, how do you know if P2 should be service driven or price driven, or executed by direct mail, or email or by phone? And the answer is that without testing you don’t know. And that’s why strategy 5 suggests that while you must always decide to do something, and let’s call that something your best reasoned guess, or your control strategy, you need to be constantly thinking about and testing alternatives to your control strategy.

As a practical matter we suggest testing a significantly more expensive and a significantly less expensive strategy against the current control, so that you can work your way toward an optimal sending level. Once you get he spending level correct you can work on improving the components of the strategy.

One last observation. Strategy 1 is the very old direct marketing, strategy 4 is the new direct marketing, but to get to strategy 5 you need to combine the best of the old and the best of the new direct marketing.

This entry was posted on Wednesday, January 16th, 2008 at 6:38 pm and is filed under Marketing Strategy & Practices. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.