Your pricing page is losing people at the comparison

Your pricing page probably does not have a price problem. It has a comparison problem, and in your analytics the two look identical, because both end with the person leaving without buying.

So founders cut the price. Nothing moves. Then they cut it again.

Why is my pricing page not converting?

Most often because the visitor cannot decide, not because they decided no. When tiers differ on six dimensions at once, ranking them takes real work, and the easy escape is to leave and think about it. That exit is silent and it looks exactly like a price objection.

A person who thinks your price is too high has actually made a decision. A person who cannot tell your Pro tier from your Business tier has made none, and they are the larger group.

Is it the price or the layout?

Two checks, five minutes. Look at where people leave and what they do just before. Price rejection tends to be fast and clean. Comparison failure is slow, involves scrolling up and down the tiers, and often ends with a tab closed rather than a back button.

1. Time on the pricing page. Under fifteen seconds points at price. Over ninety seconds with no click points at comparison.
2. Repeat visits. People returning to pricing two or three times without buying are stuck, not offended.
3. The stranger test. Ask someone outside your company which tier they should buy and why. If they hesitate, or ask you a question to decide, the page is not deciding for them.

Not sure which one you have? Revslip checks your pricing page free.

Does cutting to three tiers actually fix it?

Probably not, and the research behind that advice is far weaker than the people quoting it admit. The three tier rule usually traces back to one 2000 study about jam. The replication record since then has not been kind to it.

Here is the honest state of the evidence, because it matters more than the rule.

• 30% vs 3% jam purchase rates in the original 6 versus 24 experiment
• Virtually zero average effect across 50 later experiments
• 2025 re-test arguing both camps were measuring badly

So does more choice hurt sales or not?

It depends on the options, not the count. Iyengar and Lepper's jam table, published in 2000, found 30% of shoppers bought from six jams against 3% from twenty four. That single result became a law of marketing. Then people tried to reproduce it.

Scheibehenne, Greifeneder and Todd pooled 63 conditions from 50 experiments, 5,036 people in total, and found a mean effect size of virtually zero, with large variance between studies. Not a small effect. On average, no effect. That paper is also where the jam numbers above are quoted and checkable.

The story did not end there. Dean, Ravindran and Stoye argued in 2025 that the standard tests were underpowered, meaning they were too weak to detect overload even where it existed, and their stronger tests did find it.

Cutting from five tiers to three is not a fix. It is a guess wearing a citation.

Read those together and the useful conclusion is not a number. It is that overload turns on how hard the options are to tell apart. Six clearly different things are easy. Three nearly identical things are not.

What makes two tiers hard to compare?

Tiers become undecidable when they vary on more than one axis at once. If Pro has more seats but Business has more storage and a different support promise, there is no single question the buyer can ask themselves to choose. They have to model their own future usage first, and most people will not.

Four things do most of the damage:

• Multi axis differences. Seats, storage, support and features all changing between neighbouring tiers.
• Unowned tiers. No tier says who it is for, so every tier is for everyone.
• Repeated rows. Features present in all tiers, padding the table and hiding the two rows that matter.
• An unpriceable unit. Charging per credit or per workflow when the buyer has no idea how many they will use.

Which changes are worth making first?

Make one tier obviously correct for a named kind of person, then remove everything that does not help someone choose. This is layout and copy work, not discounting, and it is usually cheaper than a price change because it costs no margin.

1. Name the buyer on each tier. "For solo consultants" beats "Starter" because it answers the only real question.
2. Recommend one. Mark it, and say why in six words. Most pages leave the visitor to referee their own decision.
3. Cut rows shared by every tier. Move them to a single line under the table. They are reassurance, not comparison.
4. Reduce to one changing axis where you can. If seats is the real driver, let seats be the visible difference and fold the rest in.

Worth doing before any of that: price the leak first.

The formula we publish is traffic × CVR gap × AOV × mobile weight. A pricing page taking 3,000 visits a month at 1.4%, against 2.2% on your best converting page, with a €90 average order, is 24 sales and about €2,160 a month. That number tells you whether this is worth a week of work or an afternoon.

The uncomfortable part for us Across the 134 businesses Revslip has audited, comparison problems on pricing pages were common, but every one of those sites came to us already suspecting something was wrong. That is a biased sample, and it cannot tell you how often this happens on the web generally. Our tool also cannot judge whether your price is correct, only whether the page lets someone choose.

How long before you can trust the result?

Give it three to four weeks of normal traffic, and watch clicks on the buy button rather than the conversion rate. Pricing page traffic is usually too small for a fast read, and a change that helps people decide shows up as more tier selections before it shows up as revenue.

If your traffic is thin, judge the change on whether the stranger test now passes. That is a weaker signal than data and an honest one, which beats reading noise as a win.

When the price really is the problem

Sometimes it is, and no layout change saves you. If people reach pricing and leave in under fifteen seconds, consistently, across every traffic source, the number itself is doing that. Same story if your sales calls keep ending on budget rather than on fit.

The other case is when pricing is not the leak at all. If the whole site converts badly, the pricing page is just where the leaving happens to get recorded, and the twenty minute diagnostic will point you somewhere earlier.

Questions founders ask about pricing pages

Three come up more than the rest, and all three are really the same question about whether to change the number or the page.

Should I put pricing on the site at all?

Usually yes, and the deciding variable is deal size. Whether to show pricing on a B2B site covers where hiding it costs you the shortlist.

How many tiers should I have?

As many as are genuinely different. The evidence above does not support a fixed number, so use the count that lets each tier own a buyer.

Is a bad conversion rate here even unusual?

Nobody publishes a trustworthy benchmark for pricing page conversion on its own, so treat any number you are quoted with suspicion. Real benchmarks by sector and the conversion leak index are the closest honest context.