How Much Does Conversion Rate Optimization Cost?

“CRO” Is Three Different Purchases

Conversion rate optimization means getting more of the people who already visit your site to call, book, buy or send an enquiry. Vendors use the one label for three things that cost very different amounts, and most confusion over CRO pricing comes from comparing a quote for one against a quote for another.

• The audit. Someone studies your analytics, session recordings, forms and pages and writes down what is stopping visitors from acting. It is a diagnosis, delivered once.
• The fixes. Someone changes the pages: rewrites the first screen, shortens a form, moves the phone number, rebuilds a slow template. This is design and development work, and its cost depends entirely on how much the audit found.
• The testing program. Someone runs controlled experiments month after month, showing different versions of a page to different visitors and keeping the winner. It is a retainer, and it only works above a certain volume of traffic.

A small business is usually sold the third and needs the first two. That gap is where most of the wasted money in this category goes.

The Price Ranges, Line by Line

These are typical 2026 ranges for small and mid-sized businesses in the United States. Enterprise programs run far higher and are not covered here.

• CRO audit: $1,000 to $5,000. The low end covers one key page or funnel reviewed by a specialist; the high end covers the whole site with recordings, form analysis and a call-tracking review.
• Implementing fixes: $500 to $5,000. Copy and layout changes on a few pages sit at the bottom. Rebuilding a template, a booking flow or a slow mobile experience sits at the top.
• Ongoing testing program: $2,000 to $10,000 a month. Usually a minimum term of three to six months, covering test design, building variants and reporting.
• Software: under $100 a month at small-business traffic. Session recording and heatmap tools often have entry-level plans that cost little or nothing. Dedicated testing platforms cost considerably more and are only worth it when you can actually run tests.
• Performance-based deals: a share of the measured increase. Less common, and only as fair as the baseline written into the contract.

If you are pricing a single page built for an ad campaign rather than improving an existing site, that is a different purchase with its own ranges, covered in how much a landing page costs.

The Traffic Floor: Why Most Small Businesses Should Not Buy Testing

An A/B test only tells you something when each version has collected enough conversions that the difference between them is unlikely to be luck. That number is set by statistics, not by the agency, and it is much larger than most owners expect.

The Arithmetic, With Illustrative Numbers

Take a service page that converts 3 percent of visitors into enquiries. Using a standard sample-size rule of thumb at conventional confidence levels:

• To reliably detect a lift from 3 to 4 percent, a third more enquiries, you need roughly 4,700 visitors on each version, about 9,400 in total and around 280 enquiries.
• To detect a lift from 3 to 3.3 percent, a 10 percent improvement, you need roughly 50,000 visitors on each version.
• A page that gets 1,500 visitors a month would take more than six months to finish the first test, and years to finish the second kind.

Most small-business pages get hundreds or low thousands of visitors a month. That is why a quick rule works: if a page produces fewer than about 100 enquiries a month, buy fixes, not tests. Below that line a testing retainer mostly pays for reports on experiments that never reach a conclusion, or worse, reports that declare a winner early because the numbers happened to wobble the right way.

The good news is that low-traffic sites rarely need testing to improve. Their problems tend to be obvious once someone looks: a phone number that is not tappable on mobile, a form with nine fields, a first screen that never says which town the business serves. Nobody needs a statistical test to justify fixing those.

What an Audit Should Hand You

Since the audit is the purchase that matters most for a small business, judge it by what you hold at the end. A useful audit produces four things.

• A ranked fix list. Each item names the page, the problem, the evidence and the effort, ordered by likely effect divided by cost.
• Evidence you can see. Recording clips of visitors abandoning the form, the field where most of them stop, the scroll depth where the call button disappears.
• A baseline. Enquiries, calls and bookings per month by source, measured before anything changes, so the fixes can be judged honestly.
• A clear stop point. What the business can do itself, what needs a developer, and whether a testing program is justified at current traffic. Usually it is not, and a good auditor says so.

Red Flags in a CRO Proposal

A retainer proposed before any audit. A promised percentage lift before anyone has looked at your data. A testing plan for a site with a few hundred monthly visitors. A deliverable described as “best practices” rather than findings about your pages. And success measured in clicks on a button rather than in enquiries that reached you.

Measure Enquiries, Not Button Clicks

The most common way CRO money disappears is a report that shows a higher click rate on a button and no change in the number of jobs booked. Clicks on a form button are not enquiries until the form arrives somewhere a person reads it, and calls from the site are invisible unless something counts them.

Before you pay for any optimization, make sure every enquiry is stamped with where it came from and lands in one place. That setup is covered step by step in how to track where your leads come from. Without it, neither you nor the vendor can say whether the money worked.

Check the Leak Is Actually on the Page

CRO only fixes what happens on the website. Before buying it, rule out the two leaks it cannot touch. The first is the wrong visitors: if the traffic is coming from searches that were never going to hire you, no page change converts them. The second is what happens after the enquiry: a form that emails an inbox nobody checks, calls that go to voicemail, or quotes sent two days late. Both are common, and both show up as “the site does not convert” when the site is fine.

We walk through separating those causes in my website gets traffic but no leads, and the phone-specific failures in why a website is not converting on mobile. If the diagnosis there points at the page, CRO is the right purchase. If it points somewhere else, spend the money there.

Performance Pricing: Read the Baseline Clause First

Some vendors offer to work for a share of the increase in conversions or revenue. It sounds like the vendor is taking the risk. In practice the risk moves into how the increase is measured. A slow month before the contract starts makes any later month look like a win; a seasonal peak gets credited to the vendor; and if the vendor’s own dashboard is the measure, the vendor is marking its own homework.

If you consider one, insist on three clauses: a baseline period fixed in writing, ideally the same months last year; enquiries counted from your own records; and a cap on the total fee with an end date. If the vendor will not agree to all three, the offer was never really about performance.

When the Answer Is a Rebuild

Optimization assumes the site is basically sound and needs adjustment. Sometimes an honest audit finds the opposite: a template so slow on phones, so hard to edit or so dated that every fix fights the platform. When the fix list costs more than half of a new site, rebuild instead of patching. A new build designed around one clear next step per page, with calls, forms and bookings tracked from the first day, often outperforms years of small tweaks to the old one.

That is the work behind our AI website building service, and the full ranges are in how much a small business website costs. We build with Claude Code, which is why a conversion-focused rebuild can often be live in days and costs closer to an audit-plus-fixes budget than a traditional agency redesign. Pages built to convert also need visitors to convert, which is where SEO services come in.

Six Questions to Ask Before You Sign

1. How many monthly enquiries does my main page produce, and is that enough to test? (A vendor who does not ask about your traffic before quoting a testing program has answered this for you.)
2. Is the audit priced separately from implementation and from any retainer?
3. What will I hold at the end of the audit, page by page?
4. How will we count enquiries, calls and bookings, and from whose records?
5. Who makes the changes, and is that included or billed separately?
6. At what point would you tell me CRO is not worth continuing?

The first question does most of the work. A good vendor answers it with your numbers, and if the answer is “not enough to test,” recommends the audit and fixes and stops there.