Your price is read before your copy, and it makes a claim

Three numbers, one product

A founder I worked with last year sold the same monitoring tool at three prices in eighteen months: $19 a month, then $490 a year, then $2,290 a year. Same code, near enough. Same two engineers. What changed each time was not the feature list. It was who turned up.

At $19 the median customer was one developer with a side project, the support inbox ran about 90 tickets a month, and monthly churn sat near 9%. At $490 the buyer was a team lead with a company card, tickets fell to roughly 30 a month against a smaller base, and churn dropped to about 4%. At $2,290 the buyer was a head of engineering, procurement got involved, the sales cycle stretched from two days to five weeks, and churn was under 2% annually. Revenue at the top price came from 60% fewer customers and was 3.4 times larger.

The interesting part is not the revenue. It is that at each price the product was understood to be a different kind of thing. At $19 it was a utility. At $2,290 it was infrastructure with a person behind it. Nobody rewrote the homepage to say so. The number said it.

What a number tells a buyer

A price is the fastest-parsed element on any commercial page. It is a single token, it needs no reading comprehension, and buyers use it to answer three questions before they engage with anything else.

The first is is this for someone like me. Price is a budget-authority signal. A $29 tool is bought by an individual out of discretion. A $2,000 tool is bought by someone who can sign for $2,000, which in most companies is a different human with different concerns. If your copy is aimed at the practitioner and your price is aimed at the VP, the page argues with itself, and the buyer resolves the contradiction by leaving.

The second is what happens if this fails. Buyers correctly assume that price correlates with the cost of failure being underwritten. Cheap means you are on your own. Expensive means someone answers the phone. This is why a security product at $15 a month is not a bargain, it is a warning. If the vendor priced their own risk that low, they have not thought about the risk.

The third is how much of this is a person. Anything above roughly $1,000 a year implies onboarding, a human in the loop, a named contact. Below a few hundred, buyers expect to self-serve and are annoyed when they cannot.

The price is a sentence. If you have not decided what it says, it is still saying something, and probably not what you would have chosen.

Cheap is more expensive than it looks

The standard argument for a low price is that it lowers friction and buys you volume, and volume buys you learning. The first half is true. The second is usually false, and the reason is who volume brings.

Low prices recruit low-commitment buyers. Low-commitment buyers do not implement. They sign up, poke around, never move their real workflow across, and churn in month two without ever generating a usable signal. You get a lot of data about people who did not care and almost none about people who would have paid.

Run the arithmetic on your own numbers rather than trusting mine. Take support minutes per account per month and multiply by a loaded cost for whoever answers. On the $19 plan above, the average account consumed about 34 minutes a month of founder time. At any sane valuation of founder time, that plan lost money on every account that stayed. It was not a growth engine; it was a subsidy paid to people who were never going to be customers.

There is a second cost that founders miss. A cheap price attracts feature requests from buyers whose problems are not your market's problems, and those requests feel like product signal. I have watched two teams spend a quarter building for the $19 cohort while the $500 cohort quietly left. Your roadmap follows your inbox, and your inbox follows your price.

The shape of the number matters too

Beyond the magnitude, the form of the number carries information.

• Round numbers read as considered and negotiable. $2,000 says a human set this and a human may change it.
• Ragged numbers read as computed. $1,847 says a model produced this, which makes it harder to argue with and harder to believe as a considered position.
• Prices ending in 9 read as retail and consumer. In B2B they cost you seriousness without buying much conversion.
• Per-seat pricing tells the buyer this is used by many people and that adoption is the risk. Per-usage pricing tells them the cost scales with success. Flat pricing tells them you want predictability more than upside, which reads as confidence.

The unit is a positioning claim as much as the number. When we set the founder-weekend seats at $1,490 and $2,290, the choice of a per-seat, per-weekend price rather than a monthly retainer was deliberate. It says: this is a bounded thing with an end, not a relationship you will have to cancel later. Buyers who want a retainer self-select out at the pricing page, which is where I want that conversation to end.

Raising the price is a repositioning

Founders treat a price rise as a revenue lever. It is a positioning change, and it should come with the rest of the change or it will fail.

If you double the price you are telling a new buyer that a new promise applies. That promise needs supporting evidence on the page: a named implementation process, a response-time commitment, a case with a number in it, the removal of the free tier that undercut your own claim. I have seen a price double and conversion hold because the page changed with it. I have seen a price double alone and conversion fall by two thirds.

The practical sequence I use looks like this.

1. Set the new price for new customers only and grandfather everyone else for twelve months. This removes the fear that stops most founders from testing at all.
2. Change the page in the same deploy: the buyer named in the hero, the failure mode you underwrite, the onboarding promise.
3. Hold for a full sales cycle plus 50%. If your cycle is three weeks, that is five weeks of data. Judging a price rise in nine days is how people talk themselves out of a good decision.
4. Watch two numbers, not one. Conversion will fall. What matters is revenue per session and the quality of the accounts that convert, measured by whether they finish onboarding.

If revenue per session holds and onboarding completion rises, the higher price is correct even if your conversion rate looks worse on the dashboard. That trade is uncomfortable to watch and it is almost always right at the early stage, because the constraint is not customers, it is your attention.

What to do with this

Open your pricing page and write down, in one sentence each, the three claims your current number makes: who it is for, what you underwrite if it fails, how much human is included. Then open your homepage and check whether the copy makes the same three claims. If the two documents disagree, one of them is lying to your buyer, and the number always wins the argument.

Then calculate support minutes per account per month for your lowest tier and multiply it out honestly. If that tier is negative on founder time, it is not a funnel, it is a hobby. Kill it or price it to survive.

Finally, pick your next ten inbound conversations and quote the higher number. Not a test, not a discount, just the number said plainly and then silence. The thing you learn from the pause is worth more than the deal.