Personalised pricing and what you have to disclose

Adjusting what a player is offered based on their behaviour is standard practice. Adjusting what they are charged based on their behaviour has been moving into regulated territory across several markets, and the two are easy to conflate internally while being treated very differently in law.

The distinction regulators are drawing

The line, broadly, is between offering different things and charging different amounts for the same thing.

Showing one player a starter bundle and another a large pack is a merchandising decision. Both prices are the listed prices of those products, and any player could buy either. Nothing is personalised about the price itself.

Showing two players the same product at different prices, because of something you inferred about them, is personalised pricing. That is what disclosure regimes in several markets have been aimed at, on the basis that a consumer should know when the price they see was set for them specifically.

The uncomfortable middle is a discount applied to one player and not another. That is technically a different price for the same product, and whether it falls inside a disclosure requirement depends on the market and on how the discount is framed. A promotional offer available to a defined group is usually treated differently from an individually-computed price, but the boundary is not sharp everywhere.

Why this is easy to do accidentally

Most studios that are doing personalised pricing did not decide to.

A system that selects offers per player, and where offers include their own prices, produces different prices for different players as a side effect. Nobody wrote a rule that said charge this player more; the selection logic picked a different item and that item costs a different amount.

That is usually defensible — it is offer selection rather than price personalisation — but only if you can demonstrate it. The demonstration requires a record of what was selected and on what basis, which is precisely what an opaque selection system does not produce.

The related accident is a discount that is generated per player. A system that computes how much of a reduction a given player needs is doing personalised pricing regardless of what it is called internally, and that is the version most likely to attract attention.

What to have in place

The first thing is a record. For any player who saw a price, being able to say what they saw, when, and under what rule. This is the answer to almost every question that arrives, whether from a regulator, a platform, or a player asking why their friend saw something different.

The second is a stated policy about what varies. Writing down internally that prices are fixed per product and that variation happens through which products are shown — if that is what you do — turns a defensible practice into a documented one. If it is not what you do, writing it down surfaces that before someone else does.

The third is the disclosure itself, where required. The markets that mandate it generally want a clear statement, at the point of sale, that the price was personalised. Building the capability to show that in the purchase flow is much easier before you need it than after a deadline is set.

See how we keep offer decisions recorded and reviewable →

The design consequence

Beyond compliance, there is a product argument for keeping prices stable and varying the offer instead.

Players talk to each other. In any game with a community, price differences between players are discovered, and the discovery is uniformly negative regardless of who got the better deal. The player who paid more feels cheated; the player who paid less wonders what else is being decided about them.

Offer variation does not have this problem. Two players seeing different bundles is unremarkable — it reads as the store showing relevant things, which is what players expect. The same two players seeing different prices for the identical item reads as manipulation.

So the practice that is easiest to defend legally is also the one that survives community scrutiny, which makes the decision straightforward: vary what you show, keep what things cost stable, and be able to show which of the two you did.

This is general commentary rather than legal advice, and the specific obligations differ by market — worth confirming with counsel for the markets you operate in.

There is a practical reason to settle this before it becomes urgent. Retrofitting a record of what each player was shown, into a system that never kept one, means the period before the change is permanently unaccounted for. If a question arrives about last year, the honest answer is that the data does not exist — which is a considerably worse position than having a record that shows something imperfect. The record is cheap to start keeping and impossible to backfill, which puts it in the category worth doing before anyone asks.

Talk to us about compliant offer personalisation →