Every meaningful market has customers spread across a wide range of value. A payroll platform that saves a five-person consultancy an afternoon a month is the same product that saves a five-hundred-person retailer an afternoon a day - and yet a lot of SaaS companies charge both accounts the same rate per user, hoping the average works out.

It rarely does.

The small customer walks away because the price is too high for the value they get. The large customer stays, paying a fraction of what the product is worth to their business.

The fix here usually isn't a fence, though. A company doesn't jump from five employees to five hundred overnight, so a hard boundary drawn at some headcount would just get outgrown by the customers on either side of it. This kind of gradual difference is better solved with a value metric that scales smoothly with company size (see SaaS Pricing Metrics), or with tiers inside a single fence. Fencing, the tool this page is mainly about, is for splits that are categorical rather than gradual - the kind no customer drifts across simply by growing.

Price the customer, not the product.
‍
-Ulrik Lehrskov-Schmidt, author of The Pricing Roadmap

Pricing segmentation is the discipline of not letting that happen. It's how you divide the market into groups you can serve and charge separately, so revenue tracks the value each group gets rather than the value of your average customer.

Set up properly, it's the underlying structure of a pricing system that scales across SMB, mid-market, and enterprise - one that presents each customer with a plan that fits them, and gives the company a price list it can defend.

Pricing segmentation is a large part of what we do at WillingnessToPay - fencing sits inside almost every redesign we run, because it's the top-level decision the rest of the pricing system flows from. The framework below is what we use in practice (see also: our case studies at willingnesstopay.com/case-studies).

What is pricing segmentation?

Pricing segmentation is the practice of dividing your market into groups you can serve and charge differently - so revenue tracks closer to what each group values, rather than averaging it away with one number that fits nobody quite right.

It isn't a single tactic; it's a small hierarchy of decisions we work through in a certain order, starting from the top and working down through packaging, pricing metric, and modality.

Before we get into that hierarchy, it's worth clearing up a common source of confusion - because the way most teams first reach for segmentation, borrowed from their marketing team's playbook, isn't the same exercise as segmenting for pricing at all.

Pricing segmentation is not marketing segmentation

The instinct - especially in a scale-up where marketing has been running its own segmentation for years - is to lift the marketing team's segmentation straight into the pricing conversation.

It's a natural move. But marketing segments and pricing fences are different objects, built for different purposes, and using one where the other belongs is where a lot of pricing projects run into trouble later on.

Marketing segmentation is part of your acquisition model. It sorts customers by attributes - size, geography, sector, purchase history - into a statistical model that allocates sales and marketing spend. It's deliberately expansive in scope and finely sliced, and dynamic; it should create fine sub-categories and change constantly as you optimize spend.

Pricing fences are the opposite kind of object. A fence has to be discrete, stable, and non-arbitrary, because customers live inside it and pay according to it. You do not want your pricing boundaries shifting every quarter the way your marketing segments do - customers notice, and once they notice they start negotiating their way out.

It's precisely the discreteness and stability of fences that separates them from marketing segments. They aren't the same object, and importing marketing personas straight into pricing tends to leave you with boundaries you can't defend when the customer pushes back.

Fencing: the primary method

With that distinction cleared up, we can name the primary method for pricing segmentation. In our framework, it's called fencing - and everything else in the pricing system flows from the fences you draw.

A fence separates customers into distinct, non-jumpable groups you can run genuinely different products and prices for.

When you fence B2C from B2B, regulated from unregulated, or education from corporate, you can run different products and different prices on each side of the fence - and neither side resents it, because they don't see themselves as buying the same thing.

A bank doesn't feel cheated paying more than a university library for the same underlying database, because it isn't the same underlying purchase to them. That's the whole point of a well-drawn fence.

Fencing sits at the top of a small hierarchy, and the rest of the pricing system is built inside it in a specific order:

• Fencing comes first. It separates the customers into non-jumpable groups, and everything below gets designed inside a fence rather than across the whole market.
• Tiering and laddering come next. Within each fence, offers get structured around jobs-to-be-done, features, and customer size (see SaaS Packaging: willingnesstopay.com/saas-packaging).
• Pricing metric and modality are decided last - what you charge for, and how you charge for it (see SaaS Pricing Metrics: willingnesstopay.com/saas-pricing-metrics).

The opening question we ask on a design engagement is a simple one - is there an obvious way to separate our customers into distinct categories that makes every packaging and pricing decision downstream easier?

If the answer is yes, that's where we start. In practice the answer is often yes, though not automatically - a genuinely single-vertical business may already be operating inside its fence by definition (the vertical it serves versus everyone else) and doesn't need to invent a second one on top of it. The more common problem we see is different: nobody has properly asked the question, and teams either miss a fence that's sitting in plain sight or manufacture one that doesn't hold up (see the five rules below).

How to draw a fence that holds

Not every fence holds up. In our engagements we've seen a lot of fences that seemed obvious in a workshop and dissolved the first time a customer or a sales rep pushed on them.

Over time we've distilled the fences that stick down to five rules, and we use them as a test on every fence in a redesign.

The five rules of a good fence

The fast test is simpler than the rules themselves - if you can't determine which side of a fence a customer is on in about ten seconds, the fence is ill-defined.

"Startups vs. corporates" and "digitally advanced vs. not" both fail it, which is why neither survives contact with a real sales conversation.

And "on-prem vs. cloud" looks like a fence but usually isn't - customers migrate between them or run both, so it's better treated as a delivery mode than a segmentation boundary.

The three types of fences

The fences you draw usually fall into three kinds, and having names for each helps you add flexibility without the pricing conversation descending into chaos.

Each type addresses a different commercial question, and the three of them together let you serve very different customers well from the same underlying product.

Segment fences separate customers by buying profile, not by size alone. A pure revenue threshold is jumpable - companies grow through it, which is exactly what Rule 2 above rules out - so the fence that actually holds is the categorically different way each profile buys and gets served: self-serve with no sales touch, inside sales with light customization, or a dedicated team handling bespoke procurement and compliance. Thirdfort is the case we point to here: their enterprise buyers needed a fundamentally different sales motion and support model, not just a bigger number, so we built a value-based enterprise tier around that different buying process and drew the fence there. Enterprise deal closure sped up by 96%.

Commitment fences reward longer commitments. Monthly costs more per month than annual, annual more than multi-year - you trade price for revenue predictability, and both sides feel fairly treated at each rung.

Risk fences protect you from outliers. Caps, minimum commitments, and overage charges keep a usage spike from becoming an unlimited liability, and give the customer a legible ceiling on their exposure.

Together these three let you flex commercial terms by who the customer is and how they buy, while keeping a stable public list price.

This is what "dynamic pricing" really means in B2B - dynamic fences, not a sticker price that changes by the hour. That other version is a different game, and rarely the right one in enterprise.

Take SafeEx and Monta - a compliance and inspection platform, and an EV-charging platform, both scale-stage B2B businesses.

Both came to us during international expansion, and both ended up on geographic and segment fencing that let them run different structures per region and roll the new pricing out with 100% compliance across the whole organization.

The fences did the work of segmentation without opening a case-by-case negotiation with every existing customer, which is the outcome the right fence design should produce.

Why segments pay different prices: the Behavioral Pricing Matrix

Fencing and its three flavors tell you how to separate customers. Separation on its own, though, only gets you to the next question - what will each segment end up paying? The framework we use to answer that is the Behavioral Pricing Matrix, and its usefulness lies in what it tells you can't be value-priced as much as what can.

The starting point is that prices are never absolute, only relative. What a price is relative to is set by two forces:

• Competition and commoditisation - how available the same solution is elsewhere in the market.
• Customer sophistication - how much information asymmetry sits between you and the buyer. Sophistication itself is built from four factors: frequency (how often the buyer buys this kind of thing), insight (how well they understand the value), data (their access to competitor prices and products), and priority (whether they care enough to shop around).

Cross those two axes and you get four pricing conditions:

The lesson we take from the matrix in practice is that value-based pricing is only available in certain conditions - broadly, where competition is low.

Where competition is high and customers are sophisticated, they know your cost line and shop you down toward it, and no amount of value selling changes that. So the practical approach isn't to set value-based pricing as the goal in itself. It's to set making money as the goal, and use the matrix to work out where value-based pricing is available to you at all.

Segmenting customers by the conditions on this matrix, not just by company size, is what separates pricing that captures value from pricing that hopes to.

Vertical vs. horizontal: how far to segment

With the matrix in mind, one question always comes up in the same shape. When the same product creates very different value for different verticals - chain restaurants versus banks, say - do you fence them apart and charge each differently, or run one horizontal offer for everyone?

The answer is one of the more useful rules of thumb we carry into a pricing project.

You can fence verticals, and it works for a while.

Fencing has a ceiling, though. Maintaining several separate vertical schemes adds real operational load, and as a rule of thumb you shouldn't be running five verticals until you're past $1B in revenue - before that, the complexity costs more than it captures. That's a caution against running many vertical fences in parallel, not against vertical fencing altogether - if one vertical's value and willingness to pay are genuinely far apart from everyone else's (banks versus chain restaurants, say), a single vertical fence (this vertical / everyone else) can easily be worth drawing well before $1B in revenue. What doesn't hold up is five or six of them running at once.

Short of that case, the alternative is to stay horizontal and let a well-chosen value metric do the discrimination for you within a single fence. If banks get more value from the product and use more of the metric, they pay more without needing another fence built around them.

In our engagements, the cleanest structure is usually a small number of durable fences plus a value metric that scales within each - segmentation through the metric rather than through ever-more bespoke schemes (see SaaS Pricing Metrics: willingnesstopay.com/saas-pricing-metrics).

Segment on attributes, not personas

One last note, and it's the sort of thing that only comes up once the structural fencing is settled and the team turns to optimizing specific price points.

At that stage, the best price modeling doesn't predict what a customer will buy based on their segment or persona. It predicts based on their composition of attributes and the preferences attached to them.

Work with the real underlying distribution (1,000 customers each with a different user count) rather than a flattened average ("1,000 customers with an average of 10 users"), because that average hides enormous variation. A large customer can easily be 12x as valuable as a small one, because they have more users, bring more complexity and compliance needs, and pay a higher price per user as a result.

The two ideas aren't in conflict. Use a few discrete fences to give your pricing structure. Use attribute-level distributions to optimize the price points inside that structure. Personas are a useful shorthand for go-to-market, but they're a blunt instrument for setting prices.

Where pricing segmentation fails

Even a well-designed segmentation framework unravels in predictable ways when the discipline behind it slips. When we walk into a redesign and find pricing segmentation not doing its job, it's usually one of four patterns.

• Importing marketing segments into pricing. Personas built for ad targeting get promoted into pricing boundaries, and they're too fluid to hold - customers argue their way across them, and the fence stops functioning as a fence.
• Arbitrary fences. "Small vs. big," "startup vs. corporate," revenue thresholds you set internally - all of them invite negotiation, because the line is yours to move.
• Over-fencing. Too many bespoke schemes - five verticals at $30M ARR is the classic - generate more operational drag than captured value, and each exception compounds into commercial debt.
• No segmentation at all. A single price that forces SMBs to overpay or enterprises to underpay - and usually both at once.

Of the four, over-fencing is the one we see catch teams out most often in engagements. Every bespoke scheme you can't maintain cleanly eventually becomes commercial debt, and commercial debt is the slow way to stall growth from the inside. The discipline the framework is really trying to enforce is straightforward - segment enough to capture the value that's on the table, and no further.