Chipmunk Marketing

A broker tells a cautionary tale about animal wear and the search for list results This is an article about list recommendations, in story form. The characters have had their idiosyncrasies changed to maintain their anonymity and protect me from their wrath. In fact, there are no villains in this article, just a very knotty set of marketing problems.

A few years ago when I was still a list broker, I received a call from Yuri Heep, a mailer I had known for over 20 years. At one time I had brokered for his company, The Animal Imprimatur, which sold animal wear with the animals’ or pet owners’ initials or names on them. TAI specialized in chipmunks, and was in fact the largest chipmunk-blanket manufacturer and distributor in the United States.

After a two-year association, Yuri and I had fallen out. The reason, quite simply, was that after four consecutive terrifically successful mailing campaigns, his firm’s spring effort fell apart. He blamed me, and decided to punish me by withholding payment for the lists. Eventually, he sent checks for equal amounts over 24 months, but I was forced to explain to the list owners why payment was so slow.

Now he had returned, because TAI was in deep trouble. Despite having many successful years, the company couldn’t make its lists work – neither its house list nor its rented files. For the past five years, results had gone slowly but steadily down – not in any specific area, but across the board. Now the firm hovered on the brink of disaster; the entire chipmunk market was falling into disarray, and even the beaver market, a subsidiary set of products, was dropping steadily. Yuri had come to ask for my marketing help.

“I want you to take a close look at how we use outside lists and our house files,” he said. “I’ll reward you handsomely.”

“How will you reward me, Yuri?” I asked, acutely aware of the way our relations had ceased two decades earlier. “By paying me in equal installments over a two-year period, as you did before?” The past weighed heavily upon me. I was not going to be duped again.

He hesitated. “Win or lose,” he said gravely, “if you give this matter your most serious consideration and make a comprehensive marketing recommendation, I will give you the body of my brokerage over an extended period of time. I really do mean it, Bob.”

Once a fool, always a fool. I sometimes exhibit an enormous lack of common sense. And the challenge was formidable. I would examine his mailings; I would find a way of increasing his response; millions of chipmunk-lovers would continue to have their garbage cans ravaged by the lovable little animals with blankets on them emblazoned with words like “Chucky, we love you” and “Happy Scavenging, Johnny!”

I asked for all of Yuri’s material – his list histories – for the previous year. He promptly refused. “I will not give you specific results,” he said. “I’ll index the results for you.”

From a brokerage standpoint, the difference between real and indexed results – figures doctored to substitute for actual results, rather than the results themselves – is often the difference between good recommendations and mediocre ones. Show me a mailer who doesn’t want to give his broker real results, and I’ll show you a mailer who would rather retain his ego and data than improve his mailings. “No way, Yuri,” I replied. “The real results or no deal.” He eyed me. “OK,” he sighed. “I’ll send them to you this afternoon.” Then he left.

Before looking at my friend’s results, I reflected on the depressing situation described to me in the meeting. If product mailing results plateau at a particular point and then suddenly go downhill, one can reasonably guess that an aberration of nature or commerce has taken place. For example, the computer maintenance house has screwed up the merge/purge something awful, or the wrong geographical areas have been selected in both list rental and house list orders, or something like that. But a steady decline over a five-year period indicates something inherently wrong with the offer or the market to which it has been mailed. Perhaps all the chipmunks in the United States already had little blankets of their own, or lawn-owners had tired of outfitting them in the winter; maybe most of his customers felt that paying $149.35 for a sweater that said “Scavenger U.” or “Arise, Garbage Lovers!” was too much (except that TAI had done extensive price testing); perhaps chipmunks were out and horses or yaks were in. In any event, it was clear that the situation was not going to be resolved easily.

At the same time, it certainly would be worthwhile to look at the company’s list use. I might find a solution that would moderately increase this company’s results, and this could radically improve Yuri’s results and frame of mind. So I set to work.

There are five time-honored methods of bringing up list results. They are:

1. Find new markets. This would be unrealistic in TAI’s case, because it had been mailing for twenty years; if it hadn’t found all of the good lists, it didn’t deserve to be in the list-rental business. But of course I would have to check this out.

2. Find new lists in the markets that have worked for the mailer. Ditto here, for the same reasons.

3. Find new list segments that have traditionally worked for the mailer. A good idea, and worth a try. If a mailer is using a lifestyle questionnaire and making “gardening” interest work, it’s quite possible that he can make “home repair” interest work because there’s been a traditional relationship between people who garden and people who make home repairs. And so forth.

4. and 5. Find ways of tightening up the selectivity for lists that have been successful for the mailer. If one adds a segment to a list rental order, the net effect will be to strengthen the affinity between product and list, and results inevitably go up – although the universe of available names will go down. A good way to demonstrate this would be to add an age select to a list where one had not been taken before. So Fingerhut last-three-month sweeps entrants would get an additional segment: those age 40-plus.

This concept takes into consideration our strong suspicion or certainty that there is a segment or segments – usually demographic – that, if plugged in, would work for the mailer. In the Fingerhut example above, (1) Fingerhut would need a 40+ segment, and (2) we would have to be certain that a 40+ segment would strengthen the mailer’s offer. Obviously, the latter point is more important than finding the required segment in any particular list. And since many or most behavioral lists do not offer demographic segmental selectivity, the first task in this matter would be to affix with certainty the relationship between a demographic segment – age, income, and, as in the case of TAI, single-family dwelling units or SFDUs, because chipmunks are less prone to climb tenement stairs to find rotting bananas – and a product offer. And that can be done only through demographic appending – of both the house list and the most recent mailing stream.

One of the most talked-about ways to accomplish these goals is to run a multiple regression analysis. In fact, this is a very effective method, as a rule. But old-timers in the direct mail business tend to run shrieking into the hills when confronted with a graph that has lots of dots and axes. Some of their objections are not at all valid: I don’t want to go back to school; Who needs re-education when postage is rising to 34 cents; and, If I wanted to look at pictures I would reread my old Captain Marvel comic books.

Of course, some of their objections are quite valid, the primary one being that multiple regression results tend to be quite volatile: What works in November stops in December. Very often, mailers feel like they’re riding in a biplane during a storm. So in the case of our animal-clothing clients, I thought it appropriate to recommend the appending of files as indicated here. Would it be expensive? On the basis of total dollars, while not mind-boggling, the cost would certainly capture everybody’s attention. On the “per thousand” basis, it would not be terribly expensive at all. And the demographics affixed would be the mailer’s permanent property.

So, after a lot of night reading and giving up no less than two Met games – a true sacrifice indeed! – and one episode of “Law and Order,” I made these recommendations to my client:

1. I couldn’t readily find any new markets. Scratch that.

2. I found one or two new lists in the old markets that had not been tested by my client. Not too great, but worth testing.

3. I actually found a few lists with segments that might have some promise for the client. Again, worth testing.

4. A very strong suggestion that Yuri append the files, as indicated previously, to find the demographic densities that worked best, with age, income and SFDU as the three centers of attention. I offered the client a choice of three appending houses that could do the job for him in a responsible fashion. I specified the nature of the tests that should occur after the appending had taken place.

I sent the recommendation off to Yuri, and turned to the sports pages.

Doesn’t all this sound efficient and professional? In fact, like most brokers, I hadn’t the faintest certainty that any of these solutions would work. List recommendations are a crapshoot, in which brokers are certainly somewhat more proficient than mailers. That’s what they get paid for. Brokers are considerably less influenced by costs when giving this advice than they are by the desirability of getting results. To put it another way: We proposeth, but mailers disposeth. So the statement I was really making to Yuri was: This is the best way of doing it that I know, and since I work in this field eight hours a day I have a somewhat better chance of being successful at it than you do. But I make no guarantees!

And Yuri took it that way. He promptly dumped my proposal for demographic appending as too expensive; he accepted my pitifully few suggestions for new testing; he gave me an order for 800,000 names, told me they probably wouldn’t work – why should they? And walked away from me forever. Or, at least, I assumed, for another 20 years. In doing so, he broke his original promise about working with me for a long duration; but he had never been willing to live up to that promise in any event, and I knew it, and he knew I knew it. So what’s the harm? he reasoned.

Two years later, The Animal Imprimatur was sold to a multiconglomerate specializing in waste disposal that made a comprehensive study of TAI’s list data, appending age, income and SFDU information to each record. They used the data results to make a set of recommendations to local governments about how to stamp out chipmunks.

And they never mailed another name.