Pricing: the changes that failed all changed what was being bought
New Coke lasted seventy-nine days. Unity's per-install fee lasted twelve months. Fifty-nine pricing decisions on record, and the ones that blew up were not the ones that asked for more money.
Raising the price is not the risky move
Fifty-nine pricing decisions are on record. Forty were never reversed, six were undone, and ten held only in part. Read the six and the ten together and a line appears that does not follow price at all.
The decisions that survived asked customers for more money. The decisions that failed changed what the money was for.
Coca-Cola reformulated its cola in April 1985 because it was spooked by Pepsi's sip-test momentum. The company fielded over 400,000 calls and letters, and was still taking 1,500 a day in June, four times normal volume. Coca-Cola Classic returned seventy-nine days after launch. Nothing about the price had moved. The product had.
The unit is the product
Unity switched from a seat-based subscription to a per-install runtime fee in September 2023. It revised the fee within ten days, removing it for Personal-tier users and grandfathering older engine versions, and canceled it entirely a year later, returning to seats with Unity Pro up 8% to $2,200. The eventual price was higher than before. The reversal was not about the amount.
Broadcom killed VMware's perpetual licenses to extract revenue from a base it described as captive. Subscription revenue rose 275% year over year, and roughly half the midmarket base let licenses lapse around the acquisition. Again, the number moved and so did the thing being sold.
A customer who is asked to pay more evaluates a price. A customer whose unit of purchase has changed has to re-evaluate the relationship, and reopening that question is far more expensive than whatever the increase was worth.
The other failure mode is pricing to land and never raising
Disney priced Disney+ deliberately cheap in April 2019 to maximize reach in its first serious streaming attempt, intending to raise the rent later. By 2024 the direct-to-consumer businesses reached a combined $47 million operating profit against a $512 million loss the year before, and Disney+ with Hulu still lost $19 million without ESPN+.
Kering tried the opposite at Gucci, trading entry-level volume for higher-end positioning by subtraction. Revenue fell 23% as reported in 2024, and the reset's architect was fired before the higher-end trade materialized.
Both are the same problem seen from either end. A price set to acquire is not a price that works, and the move to the working price is a separate decision that nobody schedules.
How to run this decision
Drawn from what separated the records that stuck from the ones that were undone. A step nobody in the record can be seen taking is not here.
- 01
Decide whether you are changing the number or the unit
Write it down in one line. Per seat to per install is a unit change. Perpetual to subscription is a unit change. Ten percent more per seat is a number. The two carry completely different risks and the paperwork usually does not distinguish them.
- 02
For a unit change, run it as a product launch
Migration path, grandfathering, a named owner for the customer conversation, and a communications plan that explains what is being bought. Unity shipped a unit change as a pricing update and spent a year undoing it.
- 03
Compute the volume you can afford to lose, before you set the number
At the new price, how much of the base can leave before the increase stops paying? Broadcom's subscription revenue rose 275% while about half the midmarket base lapsed, which is a good trade or a bad one depending entirely on a number someone should have written down first.
- 04
If you price low to land, put the increase on the calendar now
With the date, the step and the person accountable. Disney's launch price was a deliberate land grab and the raise was left to later; five years on, the core service was still losing money. A price that is never raised was not an acquisition price, it was the price.
- 05
Separate the pricing decision from the product decision
Coca-Cola's problem was diagnosed as competitive pressure on taste and answered with a reformulation. If the pressure is on price, change the price. Changing the product to solve a price problem forfeits the one thing customers were not arguing about.
- 06
Set the withdrawal trigger and the date you read it
Coca-Cola took 79 days, Unity took 12 months, Delta took a few. The cost of a wrong pricing move is mostly a function of how long you defend it, and that is decided in advance or not at all.
What it looks like when this is going wrong
- The change alters what the customer is buying, not what it costs.Unity, 2023. Broadcom, 2023.
- The case describes the customer base as captive or locked in.Broadcom and VMware, 2023.
- A competitive threat about taste, quality or features is being answered with a price or product change.Coca-Cola, 1985.
- The launch price is described as temporary and no raise is scheduled.Disney+, 2019.
- Volume is being traded for positioning with no date by which the positioning must pay.Kering and Gucci, 2023.