How Many Pricing Tiers Should Your B2B Software Have?

How many pricing tiers should your B2B software have?

Three, unless you have a specific reason to differ. ChartMogul, which looked at over 600 SaaS pricing pages for its pricing report, found that three is still the magic number for tiers in B2B SaaS and that close to half of all companies in its report offer three separate tiers. That is the strongest default available.

Defaults are useful and they are also where thinking goes to die. Three works for a structural reason, and if that reason does not hold for your business, copying the number gives you the shape of a good pricing page without the function. I would rather you understand the mechanism and then choose three than choose three because everyone else did.

So this piece covers what the benchmark says, why three does the job it does, the two situations where fewer or more is genuinely correct, and how to decide for your own product without running a research project.

What does the benchmark data actually show?

A concentration around three, with real variation around it. ChartMogul reports that close to half of companies in its sample of over 600 pricing pages use three tiers. That leaves a majority of the remainder spread across other structures, which is a larger minority than the phrase magic number suggests.

The sample is worth knowing before you lean on the finding. ChartMogul reports that about half the companies included have fewer than 200 people, that more than a quarter have teams under 500, and that over half have not yet passed ten million dollars in annual recurring revenue. This is mostly small and mid sized software companies, which is probably your peer group, and that makes the benchmark more useful rather than less.

ChartMogul also flags a trend away from fixed tiers entirely, describing the emergence of modular pricing where companies offer so many features and add ons that each customer effectively gets a customised price for the exact features they want. That is a real structure, not a failure to decide, and it is growing.

Why does three work so well?

Because three is the smallest number that lets a buyer position themselves. One option is a price, take it or leave it. Two options is a binary that feels like a trap. Three lets a buyer say I am the middle one, or I am not big enough for that yet, which is a judgement they can make alone.

The mechanism is self selection, and self selection is what a pricing page is for. Your page cannot ask questions. It has to give a buyer enough structure to locate themselves in it and decide whether to continue. Three tiers, each with a clearly different intended customer, does that with less cognitive load than any other arrangement I have seen work.

This is also why three tiers that differ only by usage limits often underperform. If the tiers are the same product at different volumes, the buyer is not locating themselves, they are guessing a number, and guessing feels like risk. Tiers work best when each one names a different kind of customer rather than a different quantity of the same one.

When should you have fewer than three?

When you genuinely serve one kind of customer. Early stage products often have a single clear user with a single clear job, and inventing a starter and an enterprise tier to fill out the page manufactures complexity you then have to support. One honest price beats three tiers where two are fictional.

The failure mode I see here is the ghost tier. A company adds an enterprise option because the page looked empty, nobody ever buys it, and now every sales conversation includes a detour explaining what it would include if it existed. That is a tax you pay forever for a page that looked more complete for a week.

My test is whether you can describe each tier's buyer as a specific person with a specific job, without hedging. If you can only do that for one, ship one. You can add tiers later from evidence, and adding a tier because customers keep asking for something is a far better origin story than adding one because a template had three columns.

When do you genuinely need more than three?

When you serve genuinely different segments that cannot be ordered on a single axis. Three tiers assume a good, better, best progression. If your buyers split by industry, by deployment model, or by role rather than by size, forcing them into a ladder misrepresents the product and pushes people into the wrong plan.

The modular structure ChartMogul describes is the honest answer in that situation. If customers genuinely need different combinations rather than different amounts, a base plus add ons reflects reality. The cost is that your pricing page becomes harder to scan, and you take on the job of helping people configure rather than choose.

Be careful here, because this is also the most common rationalisation for a messy pricing page. Most companies that believe their buyers are too varied for three tiers have simply never done the segmentation work. The question to answer honestly is whether the variation is in what customers need or in what you have said yes to over the years.

Should the top tier be a custom plan?

Only if you will actually customise. ChartMogul defines a custom plan as what companies often mask as a separate tier but which is really an invitation to speak to sales for tailored pricing, usually with a volume discount in exchange for a longer commitment. It reports that the industry does not have a single opinion on whether to offer one.

That split tells you this is a business model decision rather than a best practice. A custom tier is correct when large deals genuinely differ in scope, security requirements, or contract terms, and those conversations produce different agreements. It is wrong when it exists purely to avoid publishing a number, because buyers read it as a price you are embarrassed by.

If you do include one, name who it is for rather than leaving it abstract. Contact us tells a buyer nothing. For teams over 200 seats, or for deployments requiring a security review, tells them whether to click. Vagueness in the top tier costs you qualified buyers who assumed it was not for them and unqualified ones who assumed it was.

Should you show prices at all?

Usually yes, and the exceptions are narrower than founders think. ChartMogul reports that companies generating between ten million and one hundred million dollars in annual recurring revenue are significantly more likely to keep their pricing private, which it connects to those companies being more likely to target enterprise customers.

Read that carefully, because it is easy to draw the wrong lesson. That finding describes companies in a specific revenue band with a specific buyer. If you are a seed stage company with a four figure contract value, the fact that a hundred million dollar enterprise vendor hides its pricing is not evidence that you should. You are not solving the same problem.

The cost of hiding prices has also gone up. Buyers increasingly shortlist before they ever contact anyone, and so do the AI systems they use to shortlist. A page with no number is a page that cannot be compared, and something that cannot be compared often simply is not. I went through this shift in more detail in SaaS pricing page design in the AI answer era.

What decides this for your company specifically?

Your segments, not your competitors. Write down the distinct kinds of customer you actually serve today, defined by what job they are doing rather than by how much they pay. If that list has three entries, you have your answer. If it has one, or seven, the benchmark is not your answer.

Then check the list against reality. Look at your last twenty deals and see which entry each one belongs to. Segments that never appear are not segments, they are hypotheses, and they should not occupy a column on your pricing page. Segments that keep appearing but do not fit any entry are the ones costing you money right now.

ChartMogul also reports that a majority of SaaS companies use free trials to acquire users, while fewer than 40 percent offer trials without requiring credit card details. Whichever tier structure you land on has to work with that entry path, because the tier a buyer sees during a trial shapes what they think they are buying. The structure and the motion are one decision, which is why I wrote about choosing between a self serve trial and a demo request.

What should you do next?

List your real customer segments this week, defined by job rather than by spend, and check that list against your last twenty closed deals. That comparison usually settles the tier count on its own, and it takes an afternoon rather than a pricing project.

If the answer changes your current structure, change it deliberately and communicate it properly, because existing customers are the expensive part of any pricing change. Getting the page right and the communication wrong is a worse outcome than leaving it alone. I covered that specific problem in changing pricing and packaging for existing customers.

And if your honest answer is one tier, ship one tier. The benchmark describes what companies do, not what is correct for you, and a pricing page that tells the truth about a simple business converts better than one performing complexity. If you want a second opinion on your structure before you publish it, reach out and let's chat.