Quantitative Methods

Personalized Pricing and the Legal and Reputational Limits

Segmented pricing is universal and uncontroversial. Individually inferred pricing is a different product with a different rulebook — and for most businesses the arithmetic says build the fence and skip the model.

  • Advanced
  • 12 min total
  • 13 chapters

What decision this helps you make: Whether to ship a price that varies by who the customer is, and if so on which axis — the operative test being whether you could state the rule out loud to the person paying more.

What this topic is

Personalised pricing sets the price from attributes of the individual buyer rather than from a published rule they could choose to satisfy. It sits at the far end of a spectrum that begins with one price for everyone, passes through segmented pricing with published conditions — student rates, volume tiers, prepay discounts — and ends with a model inferring what this particular person would pay. Everything up to the last step is ordinary commerce. The last step changes the legal exposure, the disclosure obligations and the reputational risk all at once.

Why it matters

The modelling is now easy and the data is already in your systems, so the question arrives as an engineering proposal rather than as a decision. It is a decision. The gains from individual inference over a well-built published fence are usually modest and often do not cover the cost of building and maintaining the model, while the downside is concentrated, sudden and hard to reverse. And the legal position is genuinely unsettled in places, which means the correct posture is not "is this legal" but "which of these axes could I defend in public, in writing, to a regulator and to the customer who paid more."

Who should learn it

Owners, pricing and growth leads, and product managers being asked to approve a model that varies price by customer — plus anyone who has to write the disclosure or answer for it afterwards.

What you will understand

  • Where the line sits between a published fence and an inferred price, and why that line is the one that matters
  • Which legal regimes actually govern consumer price differences, and which one everybody cites by mistake
  • How a model that never sees a protected characteristic can still price on one
  • How to compare the gain from personalisation against a plain two-tier fence, in money, before you build anything

Prerequisites

Common misconception

"Charging different customers different prices is illegal under the Robinson-Patman Act." That statute reaches price discrimination in the sale of commodities of like grade and quality to buyers who compete with each other — that is, to resellers.[1] It does not reach sales to end consumers and it does not reach services at all. This is the most confidently misapplied piece of law in commercial pricing conversations, and getting it wrong in both directions is costly: teams either abandon perfectly lawful segmented pricing on a phantom risk, or they ship individually inferred consumer pricing believing the only relevant statute does not apply to them, while the regimes that actually govern it — civil rights law, unfair and deceptive practices law, credit and insurance regulation, and consumer disclosure rules in the EU — go unexamined.