Is It Fair to Charge Different Customers Different Prices for B2B Software?

TL;DR Charging different customers different prices is the normal condition of a negotiated B2B market, not the exception. The real axis is architected variance versus accidental variance. Architected differentiation traces to structure the buyer can see: Customer Groups derive value differently, editions carry different capability sets, and the value metric scales the bill with value received. Accidental differentiation is two similar customers at materially different net prices because one negotiated harder, and it is self-punishing. Pricing to each buyer’s individual willingness to pay is that same defect run one customer at a time: the buyer cannot verify the number, so they spend the cycle shopping it and hunting for where they are being gamed. Consistency runs the other direction. It puts trust back into the ecosystem of buyers, and trust is what produces efficiency in your marketplace. Fairness has never meant an identical number for every buyer.

Somewhere in a renewal call this quarter, a procurement lead is reading a number off a screen that came from another company’s contract. The route varies: peer networks, sourcing consultants, an executive who moved from one of your customers to another. They know what a company like theirs pays, and they want to know why your number is different.

If your answer takes more than a sentence, what needs work is the architecture behind the call, not the call itself.

You Are Already Charging Different Customers Different Prices

The ethics question usually arrives as a binary: either every customer pays the same price, or you are doing something you should feel bad about. Neither branch survives contact with a real enterprise business.

Differential pricing is charging different buyers different prices for the same product when cost to serve does not explain the difference. In negotiated B2B markets it is unavoidable: volume differs, configuration differs, term length differs, the mix of Customer Groups differs, and each is a legitimate reason for two invoices to carry different numbers.

Price discrimination is the economics label for the same thing, and it carries a pejorative charge inherited from physical-goods and reseller markets where the statutory history sits. In B2B software the practical exposure is relational rather than statutory, because buyers talk. Statutory questions vary by jurisdiction and belong with your counsel.

The useful axis is not uniform versus discriminatory. It is architected versus accidental. Architected differentiation is a design decision you made, can explain, and would repeat. Accidental differentiation is variance nobody chose, discovered after the fact and defended in the moment. Both produce a spread of net prices. Only one survives a buyer who compares notes, and building the first is what value-based pricing is for.

The Three Structures That Make Differential Pricing Legitimate

Legitimacy comes from visibility. When a buyer can trace their price back to something they can see, verify, and predict, the difference reads as structure. When they cannot, it reads as treatment. Three structures do that work: Customer Groups, editions, and the value metric paired with volume.

1. Customer Groups

Clusters of customers who derive value from the product in similar ways, regardless of company size or vertical. A group deriving deep operational value from a capability is buying something materially different from a group using it at the edges, and pricing them identically is not neutrality. The difference between Customer Groups and buyer personas turns on exactly that: value derivation, not demographics.

2. Editions

Different capability sets at different price points, published, with the contents of each visible before purchase. Decades of peer-reviewed economics on self-selection menus and product versioning make this the textbook-sanctioned form of differentiation, including the counterintuitive result that a deliberately lighter version can serve buyers who would otherwise go unserved while protecting the full product’s price. Editions let buyers sort themselves without anyone negotiating individually. Our treatment of software packaging covers how the capability sets are composed.

3. The Value Metric and Volume

The value metric scales the bill with value received, so a customer extracting more pays more without a separate negotiation. Volume pricing attaches price to commitment level. Both are arithmetic the buyer can run themselves before the call.

Each is verifiable by the buyer, which separates them from the fourth thing that also produces price variance and explains nothing.

The Form Buyers Actually Punish

The indefensible form is negotiating-skill pricing: two similar customers at materially different net prices for the same value, because one had a sharper procurement team, arrived in the last week of a quarter, or drew a rep who had already booked their number.

Nobody designs this. It accumulates. A concession made once to close becomes the reference for the next deal in that territory, then the floor for the region, then the norm. Net-price variance across similar deal shapes is the most common architectural defect we find in the pattern library.

What makes it self-punishing is the discovery. Peer-reviewed behavioral research on reference price effects finds that buyers judge a price against a reference point rather than against your cost structure, and what we observe across decades of patterns in our corpus is that a discovered lower reference does not fade. The reference resets, and every published price you hold becomes an opening bid. The discount envelope widens and does not come back. Our piece on discounting approaches that slow SaaS growth traces how that widening compounds.

The same research tradition carries a second finding on renewals. Fairness entitlements attach to an existing relationship, and the sharpest objection in that work is reserved for the vendor who raises a price after learning the buyer cannot easily leave. Repricing a renewal off switching cost is the one move buyers treat as a breach rather than a negotiation.

Fairness Is the Surface, Not the Number

Fairness in B2B pricing does not mean every customer pays the same price. A single price overcharges the Customer Groups deriving modest value and undercharges the ones extracting the most, which is its own unfairness with nobody named in it.

The operational line is narrower and more useful. Market Fairness Pricing is SPP’s framework for keeping net price consistent across Customer Groups, channels, and configurations, so that similar buyers receive similar scheduled net prices for similar value, with every deal prepped uniformly for the deal desk. Same surface, governed gates, not the same number. The economic statement underneath that surface is worth making plainly: you price for the profitability of the Customer Group, and more broadly for the profitability of the customer mix, rather than for the ceiling of the individual buyer in front of you. No single deal carries the whole business, which is precisely why no single deal has to be squeezed.

A scheduled net price is what makes that testable. It is the target the pricing surface produces at any given commitment: the price a customer should land at if the deal is run as designed. Without one, “get paid fairly for your value” has no operational meaning. With one, a spread becomes measurable rather than anecdotal, and margin-calibrated discounting is the practice that produces the surface those targets come from.

Discount governance exists in every company that holds a surface. What it should never be is the thing doing the explaining after the fact.

Pricing Customer by Customer Is the Slowest Way to Sell

Value-based pricing has never been the same thing as pricing to willingness to pay. Pricing to each customer’s unique willingness to pay means pricing customer by customer rather than pricing to the Customer Group and the blended customer mix the business runs on. It presents as precision. What it produces is a number the buyer cannot verify.

The cost nobody puts on the ledger is cycle time, and what drives it is trust. The buyer gets the quote quickly enough. What they cannot do is believe it. A price set one customer at a time has nothing behind it they can check, so the sensible move is to go find out whether they are being gamed: shop it competitively, call peers, put procurement to work finding where the deal takes advantage of them. That search is your sales cycle, and it is the most elongated one of any approach.

You pay for the distrust twice, once in the weeks it takes and again in the concession you make when the buyer comes back holding comparisons. Their job has changed from deciding whether the product is worth it to auditing whether they are being taken advantage of, which is when building it themselves starts to look attractive, because a build has no counterparty to distrust.

Consistency runs the other way, and this is the part sellers underweight. A uniform architecture, where a deal of the same shape lands in the same place regardless of who negotiates it, puts trust back into the ecosystem of buyers, and trust is what produces efficiency in your marketplace. In our pattern library that efficiency appears in three places: lower sales and marketing cost, higher average selling prices, and more closed-won deal flow. Velocity is part of that, not the whole of it. Software companies know that step change from their own engineering teams, where coding assistants moved the constraint off the keyboard and onto judgment. A consistent pricing architecture does the same to a sales cycle, leaving the rep the part that genuinely needs a human.

What About AI-Driven and Personalized Pricing?

The version arriving now is whether models trained on deal history can set a price per buyer, and whether that differs in kind from what a good negotiator has always done.

In kind, no. In scale and speed, considerably. A model setting a price per buyer is customer-by-customer pricing with the human taken out, so it inherits every cost of that approach and adds one of its own: nobody in your company can explain an individual number. Willingness to pay is an observation about a buyer, not an architecture. Value derivation is what should be doing the sorting, and willingness follows from it rather than the other way around, which is also why willingness-to-pay surveys fail in B2B software.

The consumer debate about personalized pricing built on behavioral data is a separate animal, and importing its vocabulary confuses the B2B question. The test that survives both is the same: can the buyer see the structure that set their price?

Peer-reviewed research on fairness-constrained algorithmic pricing finds that fairness constraints generally reduce revenue relative to the unconstrained optimum, which is why they have to be built in as inputs rather than expected to emerge from an objective function pointed at revenue. Anything trained to maximize price per buyer will find the buyers who cannot compare notes, and in B2B those buyers eventually can.

The Test to Run on Your Own Pricebook

Take a set of recent deals with similar shapes, comparable in volume, configuration, and Customer Group, and ask whether your own architecture explains the net-price spread across them. That is the same read as a pricebook deviation diagnostic, run on your own book. Every dollar of difference should trace to an edition, a volume commitment, a value metric reading, or a documented Customer Group distinction.

Whatever does not trace is what you would defend in the room, and the share of the spread that survives is your real position. If most of it maps, you have architected differentiation and an answer for the procurement lead reading a peer’s contract off a screen. If most of it does not, calling it a price discrimination strategy flatters it, and what you have is chaotic discounting with the deal desk setting policy by default.

The rebuild from there is per-company design work: the Customer Groups, the capability sets, and the value metric are specific to your product. Our talk to an expert form is where that conversation starts.

Send us ten similar deal shapes and what your pricebook says they should have been, and a pricing expert will read it and reply.

FAQs