April 11, 2024

Can't We Do Marketing Better?

Can't We Do Marketing Better?

When I was Pitney Bowes' CEO and also the Co-Chair of the Mailing Industry Task Force, I often heard the phrase "junk mail" applied to direct marketing mail. Today, we refer to junk email as "spam" and it gets diverted away from our primary inboxes.

As an attorney who approved marketing campaigns and the copy associated with them and an operating executive to whom marketers reported, I am astounded at how marketers and operational executives define "success" in advertising campaigns If they did a direct letter mail campaign and it was "profitable" in terms of its revenue yield at a 1% response rate, they considered the campaign "successful."

But I look at it differently. If 1% of recipients respond favorably, that means that 99% of the recipients either don't want to receive the mail or email or they find the channel, timing, the price, or the product or service offered not to match their needs. Moreover, that means that Pitney Bowes or any other firm with comparable results was spending too much to secure a new order

Pitney Bowes spent a lot of effort using controlled testing through which we sent out letter mail pieces using different marketing channels, timing, pricing, messages, and products to get side-by-side comparisons of results. Our goal was to increase our response rate and reduce the cost of securing an order. The process took 12-18 months when we tried to improve the success rate for a newly-introduced Personal Post Office postage meter, and we succeeded in increasing response rates and reducing significantly the cost per order. We were never satisfied and continually sought to increase our response rates and reduce the cost of securing new orders.

Yet what we did was not typical of how many marketers operated. They got enamored of doing things a certain way, and, once they found a profitable approach, kept doing it, as opposed to trying to improve on their results.

Before 1996, Pitney Bowes spent money on TV, radio, and national print magazine advertising for the Personal Post Office postage meter and used John Ratzenberger, who played Cliff the mail carrier on the TV series Cheers, as our celebrity spokesperson. Outsiders complimented us and our employees and directors were excited that we were getting national visibility for this new product.

There was only one problem: the economics of this marketing campaign were not great. Too many customers signed up, received their postage meter, and then cancelled within a few weeks or months of receiving the equipment. We had the highest cancellation rate ever as a result of this campaign. We transitioned to much less glamorous and more targeted marketing and, over time, were successful enough that our head of small business marketing, Neil Metviner, on behalf of our small business customer division, accepted an award as B-to-B Marketer of the Year by the Direct Marketing Association

Many marketers like the idea of doing memorable campaigns and getting awards for their creativity, but, ultimately, the fact that non-customers admire the campaign and reward those who designed and implemented it is beside the point. The simple question for any marketing message, channel, offering, or timing is this: did it succeed in delivering the highest lifetime customer value? This is true, whether the product or service is sold to business customers or consumers or even to employees that self-insured employers are trying to engage for health-promoting activities.

For example, as we think about the role of generative artificial intelligence (AI) for the production of marketing copy, its most compelling value proposition is not that it significantly outperforms human marketing copy writers the first time around, although it might achieve that result. With generative AI, marketing copy can be created in less than a minute for pennies, which, for many customers, will make it appealing.

But, when properly designed, it learns far faster than do humans. AI is far more effective and infinitely faster in delivering multiple versions of a message, pricing plans, timed outreaches, and channels and in analyzing what combinations work best for what target audiences. The gap between the human marketer and intelligent use of AI widens rapidly. Our MoveFlux platform learns faster than human marketers, but it is designed to create marketing copy consistent with the client's goals, brand guidelines, style of communication and data protection and security requirements.

But the business marketplace uptake is far slower than it should be, because many marketers aspire only to deliver a "profitable" campaign, that is, a campaign in which the economic return for the campaign exceeds the cost of designing and executing it. Marketers often get insufficient funding to test whether their campaigns can be improved. All too often, response rates decline, but they keep doing the same things.

Some marketers experiment, but they rely on human intelligence, as opposed to testing their instincts and insights against alternatives that AI might generate. They are highly confident that they can discern what works and why it works, even though they do not test their instincts systematically.

John Wanamaker, a pioneer in creating the retail store formats that dominated how Americans shopped for most of the 20th century once said: "Half the money I spend on advertising is wasted; the trouble is I don't know which half." That no longer has to be the case, but many marketers act as if it were still impossible to assess what advertising and marketing spending is accomplishing or as if their judgments, which are not based on good data, enables them to determine whether and to what extent their marketing activity is working.

Many Americans who reach high levels in organizations want to settle into predictable, consistent processes that produce enough positive results to help them stay in their jobs, but are not driven to improve on those processes.

Artificial intelligence upends all of this. It is partially because it produces different or better results than humans can, but also because it challenges humans to employ AI to improve continually. Its likelihood of improving outcomes is not guaranteed. At least right now, human intelligence is needed. But it is inevitable that firms that integrate generative AI into their marketing processes will outstrip those that fail or refuse to do so.