The discount that never ends.

Circana published its back-to-school read a fortnight ago, and there is one line in it I have not been able to put down. Buried in the commentary, almost apologetically, was the observation that promotions on their own may no longer be enough to move demand.

That is a very polite sentence. Sit with it for a minute and it becomes an obituary. While that report focuses on the US market, the story in India isn't very different.

The consensus reading of the July numbers went the way these readings always go. Retail revenue down one per cent, units down two. Discretionary general merchandise softer still. Everyone simply wrote "cautious consumer", filed it, and went back to building the next promotional calendar.

I think that reading is comfortable, and I think it is wrong.

The consumer did not get tired. Our selling did.

A promotion is not a price. A promotion is an argument, and the argument is: this is worth more than you are paying today, and today is special.

Run that argument twice a year and it works. Run it every fortnight for five years and the customer eventually does the sensible thing. They stop believing the first half of the sentence and start believing the second. The special price becomes the price. The listed price becomes an opening bid that nobody is expected to pay.

We see it in our friends and family. They defer purchases knowing "oh, they'll have a discount at the start of next month, I'll get it then."

This is measurable, and it has been measured. Kalwani and Yim showed in 1992 that both the frequency and the depth of promotions move what shoppers expect to pay, and that the movement is asymmetric. A promotion that fails to arrive costs you more than an unexpected one earns. You are not running a discount. You are renegotiating your own reference price, permanently, inside someone else's head — and the only one keeping notes is you.

I have started calling it renting your own demand. You did not create the sale. You borrowed it from a customer who was going to buy anyway. And the rent will be due again next month, at a slightly higher rate.

Nobody put the reference price on the ledger

Here is what makes this so hard to stop, and it has nothing to do with consumers.

Every promotion has an owner, and that owner is measured on the promoted period. Lift during the window. Revenue against the same week last year. Everybody in that meeting is graded on the fortnight.

Nobody in the room is graded on the reference price. Nobody's dashboard has a row for "what our customers now believe this product is worth". The cost lands two years later, on a different person.

The measurement counts one side of the ledger. It always has. An incentive structure that only counts the upside of a decision will keep producing that decision forever, regardless of what anybody in the room actually believes. This is an operating problem before it is a pricing one — the same category of failure as the handoffs that quietly become marketing's most expensive line item.

You cannot verify a peptide

This hurts far harder in some categories than others.

In 1970, the economist Phillip Nelson drew a line between two kinds of goods. Search goods, where you can assess quality before you buy: a spanner, a kilo of rice, a plane ticket to Delhi. Then experience goods, where you cannot, because the only way to know is to consume the thing and wait.

Almost everything winning in consumer right now is an experience good. You cannot verify a peptide serum on the product page. You cannot verify a magnesium supplement, a sleep protocol, or a scalp treatment. You certainly cannot verify a B2B platform during a forty-minute demo.

When a buyer cannot assess quality directly, they reach for whatever proxies are available, and price is the loudest one in the room. Rao and Monroe's meta-analysis of thirty-six studies found the effect strongest precisely when people can compare prices side by side — a fairly good description of every product page ever built.

Which means that in an experience-goods category, your price is not sitting next to your quality claim. Your price is a quality claim. A thirty per cent discount is not a marketing decision. It is you, publicly, revising your own evidence of quality.

I have been watching this in the data. Across a panel of US direct-to-consumer storefronts this month, the single largest health and beauty brand realises about ninety-two per cent of its list price. Its ambitious challengers, the ones running the promotional calendar, sit in the sixties and seventies. The discipline is not a reward for winning. It looks a great deal like the reason.

Everybody learns how to enter

It's like Abhimanyu learning how to enter the chakravyuh while still in his mother's womb, but never learning how to leave.

That is the promotional spiral, and it is nearly the whole story. Every founder learns the entry. The entry is easy, it is well documented, and it works the first time. Almost nobody learns the exit, because the exit is not a tactic. It is a repositioning, and it costs a quarter. And repositioning, as I keep saying, is not a tagline.

The JCPenney objection, which is a very good one

In early 2012, Ron Johnson removed coupons and sales from JCPenney and replaced them with everyday fair pricing. It was intellectually correct. It was honest. It was, by any reasonable reading of the psychology, the right diagnosis.

Sales fell twenty-five per cent in the year to February 2013, taking about $4.3bn of revenue with them, and he was gone by April.

I do not think that case disproves the argument. I think it demonstrates that the exit is real and it is narrow. JCPenney removed the reason to buy today and put nothing in its place. The customer did not lose a discount. They lost the only story the brand had been telling them for a decade, and were handed integrity instead — which is admirable and does not go in a shopping basket.

The exit is not "stop discounting". The exit is: give people a reason to believe the product works, and then hold the price so the reason stays credible. Those two moves are one move. Doing the second without the first is what killed JCPenney.

For a B2B founder reading this and feeling comfortably distant: the same mechanism runs on your end-of-quarter desk. Fifteen per cent to close by Friday, twice, and you have taught your entire buying market what your software costs. Procurement takes better notes than any consumer ever has.

What I do not know

I do not know whether a brand with genuinely no differentiation can get out at all. If your product really is interchangeable with the one beside it, price may be all you have, and this essay is not much use to you. That is a real limit.

What to do on Monday

Take last year's promotional calendar and put one extra column next to it: what did we tell the market our price actually is. Not margin. Not lift. The number a returning customer now waits for.

Most teams have never written that number down, and writing it down changes the conversation more than any deck about brand equity ever has.

Then the harder question, the one that column always ends up asking. If you took the promotion away next quarter, what would be left to say?

I would genuinely like to hear from anyone who has run that exit and survived it. The failure is famous and the successes are quiet, which is exactly the wrong way round for everybody trying to learn from them. If you have got out, or watched someone get out, I want the story.

Sources: Circana, "Back to School Signals a New Reality", August 2026. Kalwani & Yim, "Consumer Price and Promotion Expectations", Journal of Marketing Research, 1992. Marketing Science Institute report 09-213. Nelson, "Information and Consumer Behavior", Journal of Political Economy, 1970. Rao & Monroe, Journal of Marketing Research, 1989. Realised-price figures from Particl's US DTC storefront panel, trailing 30 days to mid-August 2026. JCPenney figures as reported at the time of Ron Johnson's departure, April 2013.