The Ad That Worked

Dear friend,

The riddle first, as the postmark demands. Your century today, and the subject is the happiest morning a trader ever has, and what he usually does with it.

Two sellers find an ad that sells,
Both hear the same delighted bells.
One triples what he spends by noon,
One climbs by thirds, and not so soon.
By spring, one ad lies dead and cold.
What killed it? Not the words, I'm told.

Hold your answer. Story time.

The Story

Two firms selling the same sort of thing, and the goods, all together now, were equally good. Both advertised. Both, in the same month, found the thing every advertiser is looking for: an advertisement that worked.

Forty pounds a day going out, and sales coming back at twelve pounds apiece. The first seller looked at that on a Sunday evening and reasoned exactly as you or I would have reasoned, which is the trouble with it. If forty pounds buys me this, then four hundred buys me ten times as much. He was not being greedy. He was being arithmetical, and he had a coal fire in his head: more coal, more heat.

On Monday he put four hundred a day behind it, and by the following Sunday his sales cost him forty-six pounds each, and he could not for the life of him understand what had happened to a machine he had not touched.

Three things had happened, and none of them were about the advertisement.

The first: the forty pounds had been reaching the keenest few thousand people in the country, the ones already half looking for exactly this. There are only so many of them. The extra money could not conjure more; it could only go and find the next hundred thousand, who were colder, vaguer and further from their wallets, and who cost more per sale because they were worth less.

The second: frequency. The same faces now saw the same advertisement nine times in a week, and an advertisement seen nine times by an uninterested person is not persuasion, it is wallpaper, and slightly irritating wallpaper at that.

The third, and the one that catches the careful: to spend ten times the money in the same day, the machinery must place the advertisement in ten times as many places, and the additional places are, by definition, the ones nobody was bidding for.

So he switched it off, told everyone at the trade evening that advertising doesn't work any more, and went back to nothing. And here is the real loss, larger than the wasted fortnight: he had spent the whole of his good season running the one advertisement and making no others. When it died he had an empty cupboard, and had to start again from a standing position with no idea which of his ideas was any good.

The second seller did two dull things. She raised the spend by about a third at a time, waited to see what the cost per sale did before raising it again, and stopped climbing the moment that number got worse rather than when the money ran out. She was not watching her sales, which always rise when you spend more, and always flatter. She was watching what each sale cost her, which is the only figure that tells the truth.

And while the good advertisement was running, she spent her afternoons making the next three: a different opening, a different proof, the same offer put to a different sort of person. Two were worse. One, eventually, was better. That is the trade, and it never changes: she was not trying to preserve one lucky advertisement, she was building a cupboard, exactly as the knife man built his by testing one thing at a time and keeping what won.

The Lesson, Since You've Earned It

You cannot buy more of the people who wanted it. You can only buy more of the people who didn't. That is the whole of scale, and it is why the winning advertisement so reliably dies at the moment of its promotion. It never was a machine. It was a match: a particular message meeting a particular pocket of people who were ready for it, and the pocket has a bottom. Every extra pound spent past that bottom is spent on strangers who fit the message a little less well than the last lot, and the cost per sale rises for reasons that have nothing to do with your words.

Which means the number you watch decides whether you keep the thing or kill it. Total sales will rise while you are ruining yourself, because you are buying them; cost per sale will tell you the day the pocket ran dry. Watch the wrong one and you will scale enthusiastically all the way down.

And when the pocket is genuinely empty, the answer is not a bigger budget, it is a bigger pond, which is slow work done honestly: a new angle for a different sort of buyer, a new proof for a warier one, a fresh advertisement made while the old one still pays. The man who makes nothing during the good months is not being efficient. He is betting his whole trade on one lucky sentence lasting for ever, and no sentence ever has.

Three questions, and open the numbers rather than the memory before you answer:

1. What does a sale cost you, week by week? Not the total, not the sales, the cost of each one, written down in a line so you can see the direction of it. If you cannot produce that figure for the last eight weeks, you are not advertising, you are hoping in public and paying for the privilege.
2. How fast do you raise the money? A third at a time, then wait for the number to settle. Any faster and you will not know whether what you are seeing is a bigger pond, a tired audience or an accident, and you will draw a permanent conclusion from a fortnight's noise.
3. What is in the cupboard? Name the next three advertisements you would run if today's stopped working tomorrow. If the answer is none, that is the most urgent thing in your business, and it wants doing this week, while things are going well and nobody feels the need.

The Answer to the Riddle

What killed the ad? Not the words.

The money killed it, poured in faster than the audience could grow, until the advertisement was being shown to people it was never written for. It was the same advertisement on the Sunday it worked and the Sunday it failed, unchanged, unaltered, every syllable identical. The first seller concluded that advertising was broken. What was broken was his belief that a match between a message and a pocket of people is a machine you can feed.

The verse:

The words that worked at forty pound
Are still the words when nothing's found.
You bought the keen ones out, that's all,
Then paid to shout at strangers' walls.

Until next time. Go and write down what a sale cost you each week this quarter, and make one new advertisement while the current one is still paying.

P.S. And if an advertisement of yours has been running untouched for a year and still pays, do not tinker with it out of boredom, which is the other way trades kill their best work. Leave it be and build the next one beside it. Boredom is the advertiser's disease, and the customer has never once caught it.

Also in the archive: The Chandler's Two Jars, What You May Spend and The Leaking Coach.