From talking to people about insights gained about digital marketing from Pete McCarthy and learning new things both by having the conversations and then ruminating about them. it has recently become obvious that as people learn Pete’s lessons, they’re going to encounter a new problem they don’t have a solution for. This must already be apparent in some quarters. (Actually, if it’s not, it shows your digital marketing efforts are probably in need of improvement.)
A core tenet of Pete’s approach is that you define the audiences for your book (and one big jump is to get from “topics” in the book to “audiences” for the book) and then you use a variety of tools to find out what words these audiences use when they’re searching, where they hang out, how they get recommendations, and what else they like, believe in, and do. All this is done with the objective of aligning marketing efforts with true consumer intent and behavior.
Once you have defined the audiences and found the right terminologies, places, and times, the next question is which terms work best to drive engagement, and then sales. You often find that out by doing some experimenting: finding ways to market to the audiences (a Facebook interest, a Twitter hashtag) and testing the different words (and sometimes imagery) you think will resonate with them. This is classic A/B testing.
So if you have, as you might for many books, five audiences and six search terms under each, you could have 30 different “experiments” to conduct for that book from just your initial research. That’s a lot of A/B testing to keep track of.
One very quickly realizes that a major publisher using this approach should have tens of thousands, perhaps hundreds of thousands, of experiments taking place simultaneously.
This ties into another point Pete has made to me and which will be the subject of the final panel at our Marketing Conference on September 26. You can know whether digital marketing is “working” or not. “Working” means “positive ROI”: you spend a dollar on the marketing effort and you get more than a dollar back. As long as that is true, you keep doing it. If it isn’t true, you (probably) stop.
In other words, the old practice of setting a “budget” for a book’s marketing effort is an anachronism in the digital world. (That’s the main topic of that wrap-up panel.)
This idea also recalls a concept that was the topic of a post here three years ago: that marketing spending should be seen as “investment” because what is learned or gained in the way of customer attention helps the marketing efforts of the future.
I spoke with a big house CEO in the UK last week about this and he wholeheartedly agreed that the need I identified was a pain point. But, he said, at this point the experiments in his shop, and he believes in competitors’ shops as well, are not even being systematically measured for ROI.
While it is a tall order, the benefits of doing this — making all marketing spend an “experiment” and measuring its ROI — are too great to ignore and the alternative is choosing to guess when the answers are out there to be gotten.
This leaves a big gap between the marketing that is actually possible today given our ability to learn about audiences and search terms and social media engagement on one hand and what publishers can practically do across their lists on the other. That is, we’ve learned about some things that can really work, but which are devilishly difficult to scale.
The number of experiments taking place in most houses should likely exceed the number of active titles in the marketplace. Some of these might be no- or very low-cost. You can learn a lot buying $50 worth of keyword exposure in Facebook. But if you don’t continue it (or even increase it) if it is working (or turn it off it isn’t), then you don’t gain the benefits you should and possibly a good practice can turn into a disaster.
Once publishers more broadly learn and understand the marketing techniques Pete has developed (many of which will be explained at the Marketing Conference), they’ll discover the need for the next tool: a way to automate the management of these techniques on a broad scale. They’ll want technology that measures whether something is working and turns it off or continues or even extends it based on ROI. (Indeed, before they can do that they’ll need to develop their own “attribution” model that ties marketing spend to results and then to profit; it is not as simple as direct online conversions, though that is the logical place to begin.) They’ll want exception reporting that bubbles up what humans need to look at on a regular basis.
This is a complex technology problem. If it is seen as primarily an IT requirement, publishers might be reluctant to fund a solution. If it is seen as essential for marketing, it might be looked at differently. This is a dichotomy Pete put his finger on, which he calls the awkward dance between marketing and IT.
Right now, the chances are good that most houses aren’t doing enough research (into audiences and search terms) or enough experimenting. The research doesn’t pay off if you don’t use it, and the use doesn’t pay off if you don’t measure and manage it. (Or as Pete says, and will say at the conference, “rinse and repeat”.)
A lot of this is common practice outside the publishing industry and many “marketing automation solutions” do exist, some of which have been adopted by some houses for some specific uses. The trick is to find the tool that fits the need; like a pair of pants, it has to fit. And with the sheer number of small and diverse experiments required for publishing — so many ISBNs and so many retailers and so many marketing venues — the right comprehensive tool hasn’t been created yet.
I wonder if any of the established systems providers or start-ups looking to help publishers are working on this problem. If the formula for success is “see a need and fill it”, it might be a good idea. As publishers develop their competence at digital marketing, the need will be very apparent.
Also on September 26 and in the same midtown Manhattan venue as our Marketing Conference, Publishers Launch Conferences (in conjunction with our partners for this event, Digital Book World) will be presenting a Publishing Services Expo. The Expo consists of three mini-conferences designed to help publishers figure out how best to get help from service providers in three distinct areas: 1) Editorial/Production, 2) Rights and Royalties, and 3) Digital Asset Distribution. The Services Expo is priced low to make it affordable for attendees who want only one or two of the three tracks. Ticket-holders for the Marketing Conference are also allowed access to the Services show for anybody who wants to duck out for a particular piece of it.