Pricing tiers should split buyers, not features

A pricing tier is a buyer, not a bundle of features.

Most SaaS pricing pages have three columns. Starter, Pro, Enterprise. The columns differ by a long list of checkmarks, and nobody on the buying side can tell which column is theirs without a call. That is not packaging. That is a feature spreadsheet with prices on top.

Packaging is the decision about who each tier is for. Get that right and the page sorts buyers for you. Get it wrong and every tier competes with the one next to it.

What does splitting tiers by buyer look like?

Look at how Meta packaged its AI subscriptions in September.

TechCrunch reported that Meta One launched with consumer plans, Core at $7.99 and Premium at $19.99, built around expanded Muse image and video features. Separately, creator and business tiers ran from Essential, starting at $14.99, up to Max, starting at $499, with Meta Business Agent capacity. Whatever you think of Meta, the structure is easy to read. A person making images and a business running an agent are different buyers with different jobs, so they get different ladders.

Compare that with the typical B2B page, where Pro adds a few integrations and Enterprise adds "advanced analytics." The buyer has to work out which checkmarks matter. Most will not. They pick the cheapest tier that looks safe, or they leave.

Why do feature-count tiers fail?

Three reasons.

They make buyers do your segmentation. If the page does not say who a tier is for, the buyer has to guess. Guessing is friction, and friction on a pricing page costs revenue. It is the same standard as the grunt test for messaging: if someone cannot tell quickly whether it is for them, the page failed.

They cannibalize. When tiers differ by small feature gaps, buyers pick the cheapest one that covers today and plan to upgrade later. Later rarely comes, because nothing in the product tells them they have become a different buyer.

They hide the upgrade trigger. A good tier boundary sits where the customer's situation changes. More people involved. A second team. A security review. Money flowing through the product. When the boundary is one more toggle, there is no moment where upgrading feels obvious.

How do I pick the buyer for each tier?

Start from your ICP and the jobs inside it, not from your feature list.

1. List the distinct buyers who pay you today. Not personas. Buyers with budget and a job: a solo operator, a team lead, a department head with procurement rules.
2. Write the one job each buyer hires you for. If two buyers have the same job, they probably belong in the same tier.
3. Find the value metric that grows with each job. Seats for a team tool, volume for a usage tool, outcomes if you can defend them. I made the outcome case in price the GTM outcome, not the AI seat.
4. Put the boundary where the buyer changes. The upgrade should happen when the customer becomes a different buyer, not when they want one more feature.
5. Name tiers after the buyer or the job. "For solo founders" and "For sales teams" sort buyers faster than "Plus" and "Pro."

Then test each column: could a buyer read only the header and first line and know it is theirs?

What about AI products with usage costs?

The same rule holds, with one extra constraint.

AI products carry real costs per use, so tiers need usage limits. The mistake is letting the limit become the only difference between tiers. If Starter gets fewer credits and Pro gets more, you have one product sold at two volumes. That is fine for a usage add-on. It is weak packaging.

Put the buyer first and the limit second. A team tier should include what a team needs: a shared workspace, admin controls for billing and users, and more usage. A solo tier should be priced for a solo budget. The credit math still has to be readable, because tiers do nothing if a buyer cannot estimate the bill.

How do I test new tiers without a full reprice?

Change the page for new buyers first. Leave existing customers where they are.

Rewrite the tier headers around buyers. Then track three things: which tier new signups and new deals choose, how often prospects ask sales which plan fits, and how many accounts upgrade when their situation changes. If "which plan is right for me" questions drop and the mix shifts toward the tier built for your best customers, the packaging works. If not, your buyer segments are wrong, and that is a positioning problem before it is a pricing problem.

Adapt or fail. Price the buyer, not the checkbox.

Start Signals, Convert, Grow

FAQ

What is good-better-best pricing?

A three-tier structure where each tier costs more and does more. It works when each tier maps to a distinct buyer and job. It fails when the tiers are just feature counts.

How many pricing tiers should a SaaS product have?

As many as you have distinct paying buyers, usually two to four. More tiers than buyers means your tiers compete with each other.

What is a value metric?

The unit your price scales with, such as seats, usage volume, or outcomes. It should grow as the customer gets more value from the job they hired you for.

Should I grandfather existing customers when I change tiers?

Usually yes, at least for a while. Test new packaging on new buyers first so you can read the result without a wave of complaints from current customers.

How do I know if my tiers are confusing?

Count how often prospects ask which plan fits them, and check whether most buyers pile into the cheapest tier. Both mean the page is not sorting buyers.