What a three-way tie on my book cover taught me about reading buyers
Picture this, you poll your customers. The results come back split down the middle. No clear winner, no clean story to tell leadership. So you shrug, pick the one you liked anyway, and file the exercise under inconclusive.
That shrug is the mistake.
Last week I put three potential cover designs for my first book (called The Buyer in the Loop) in front of two audiences at once. My LinkedIn network voted. So did a grounded synthetic customer panel of four personas, built from real buyer evidence. The public count came back almost perfectly tied. And the tie was the least interesting thing in the data.
The headcount hid the story
When I added up the visible votes, the three options sat within a whisker of each other. By raw count, nothing won. Stop there and you have a tie and a problem.
It turns out the aggregate is the wrong unit of analysis. A poll result is a sum of very different people pulling in different directions, and the sum erases the one thing you need to see: who wanted what, and why.
So I went back and classified every visible commenter by the role they hold, then looked at how each role had voted.
Segment before you count
Sorted by role, the tie dissolved into a pattern with a spine.
- Leadership, the founders and CEOs and senior executives, leaned toward option 2 - a version with a Gartner-esque blue background. Their language clustered around authority, credibility, and premium positioning. The phrase that kept surfacing was a version of "the book I should have on my shelf."
- Senior marketers split almost evenly across all three. The people whose actual job is choosing between options were the least decisive group of all, which is its own quiet finding.
- Commercial and sales leaders showed the strongest relative pull toward option 3, it was the only one that pictured a "buyer" and their reasons were practical: distinctiveness, and the kind of cover that stands out in a buyer's hand.
- Junior and mid-level marketers broke hard for option 3. Small sample, but the cleanest preference of any segment.
Same three covers. Four audiences. Four different answers.
That is the market telling you it is more than one market.
The synthetic panel called the split
Here is the part I find hard to stop thinking about.
Before a single human voted, my synthetic customer panel had already mapped the terrain. The leadership voice landed on option 2. The two marketing voices split, one on option 1 and one on option 3. The sales voice went to option 3.
Then the humans voted, and the shape held. Leadership leaned to 2. The senior marketers split almost evenly, the same fault line the panel had shown. The operators and commercial voices broke for 3.
I want to be precise, because the claim is easy to oversell. The panel did not pick a winner. With four personas it could not, and across the contested middle it disagreed with itself. What it predicted was the structure of the disagreement, which kind of buyer would pull which way and where opinion would fracture. My LinkedIn community then filled that structure with texture the panel could never have invented on its own.
This is what a grounded synthetic customer is for. Fast directional signal, before you spend a penny on the real thing. The humans stay the ones who decide.
Am I measuring the wrong thing?
Now the uncomfortable part. A vote measures what people will say in public. It captures what they want to be seen choosing. Whether that matches what would make them buy is a separate question, and a poll cannot answer it.
Look at the recurring words. Option 2 drew premium, sophisticated, authoritative. Flattering, and also exactly the vocabulary a senior marketer uses to perform good taste in front of peers. The vote may be measuring social positioning rather than purchase intent.
So the count being tied matters less than it first appears. I was never going to learn the winner from a public poll. What the poll gave me instead was the language each segment reaches for and the trade-offs each one weighs. That is the asset. The vote was just the wrapper around it.
A fourth option, never offered
The strongest signal in the whole exercise was a design that did not exist.
Several commenters, independently and without prompting, described the same thing: the authority of option 2 carrying the human element of option 3, integrated more subtly than the busier version on the table. They had moved past my three. They were describing a fourth design that did not yet exist.
That beats a constrained vote, because people are telling you the design they actually want rather than picking the nearest available compromise. So I am not running this vote again. My next conversation is with my book designer Jim MacLeod (author of the Visual Marketer - a great book I suggest everyone in marketing read), about building the option nobody got to vote for. (Can you tell I just love building up a sense of suspense around this book?)
The tie was the map
Go back to the shrug, the tied result that looked like a failed experiment.
It was the opposite of a failure. The moment your audience splits evenly is the moment the aggregate stops helping and the segments start talking. Read who is choosing what and listen to the words they reach for. Pay close attention to the option they keep describing that you never offered them.
A tie is a map. You just have to read it by who is holding it.
