Would it have sold anyway?
A publisher borrowed a method from epidemiology to find out whether Shimmr’s advertising was causing sales or just standing next to them. He stopped the campaign and measured what happened.
Mensch Publishing · Getting Old: Deal With It by Lee Janogly · United States
The question every publisher actually asks
Not “did the book sell?” but “would it have sold anyway?”
It is the right question, and it is the one advertising almost never answers. A title goes on sale, a campaign runs, copies move – and the publisher is left unable to separate the advertising from the review, the season, the algorithm, or plain luck. Most case studies in this industry quietly skip past that problem. This one was designed around it.
Borrowing a method from epidemiology
The test was proposed by Richard Charkin, who runs Mensch Publishing and advises Shimmr on publishing. He has been President of the International Publishers Association and of the UK Publishers Association, Chief Executive of Macmillan, Executive Director of Bloomsbury for eleven years, and has held senior positions at Oxford University Press, Reed Elsevier and Pergamon. He was appointed OBE for services to publishing in 2024.
His reference point was Richard Doll’s British Doctors Study.
Doll’s problem in the 1950s was that plenty of data linked smoking to lung cancer, but correlation alone could not establish cause. His insight was to measure the reverse: he invited smoking members of the British Medical Association to quit, then compared them against those who continued. Watching the effect disappear when the cause was removed proved what watching it appear never could.
“I thought I might apply the same logic to Shimmr advertising results.”
Richard Charkin, Mensch Publishing
The title
Getting Old: Deal With It by Lee Janogly. A Mensch Publishing title, and exactly the kind of book the industry struggles to market: a good book by a working author, with no festival moment, no prize longlisting and no obvious hook to hang a campaign on. The sort of title that, in most houses, is published and then left alone.
The measurement.
Monthly rates, so the two periods compare like for like. Six months of advertising against the sixty days that followed.
Sold via Amazon while the campaign ran, then none at all after it stopped
Monthly US ebook sales fell by well over half once the advertising was withdrawn
No edition, price, publicity or seasonal change – only the advertising
Three phases, one variable.
Mensch changed nothing about the book. They changed only whether Shimmr was running.
Before
The title was on sale in the United States with no advertising behind it, performing as most unsupported backlist performs.
During
Shimmr ran autonomously for six months. Sales appeared, in both ebook and paperback, without any other marketing activity.
After
The campaign was stopped and the budget moved to another Mensch title, Tom Campbell’s Helen Hoskins. Sales stopped with it.
Because the effect appeared and disappeared in step with the advertising, the result cannot be explained by the book, the market or the season. Only by the campaign.
What the pattern shows
Sales began when the advertising began, and stopped when the advertising stopped. Nothing else about the title changed across either transition – no new edition, no price change, no publicity push, no seasonal event. The only variable that moved was the advertising.
Paperback is the cleaner signal of the two. Print sales ran through the campaign and then ceased entirely once it ended. Not a decline – a stop. Ebook sales did not vanish, but fell away sharply.
Notably, print was the effect Richard had considered least compelling during the campaign. It only became legible once the advertising was removed. That is the entire argument for measuring backwards.
“Sales of GODWIT have disappeared. QED the advertising worked independent of any other factors.”
Richard Charkin, Mensch Publishing
Why this matters if you are a small press
Independent publishers carry the discoverability problem in its harshest form. You have titles worth reading and no realistic way to give each one sustained marketing attention, because the arithmetic of doing it by hand has never worked at your scale.
What this test establishes is narrow and load-bearing. For a title that would otherwise have been left alone, the advertising was the thing producing the sales. When it stopped, they stopped. That is the mechanism, demonstrated rather than asserted – and demonstrated by a publisher who set out to disprove it.
Run the same test on your list.
We would rather you measured us than believed us. Bring a title, or a hundred, and we will show you what the Heat Pulse surfaces before any budget is committed.
Book a demo hello@shimmr.ai