How to choose a CRO agency

A good one does four jobs, in this order:

1. Audit. They look at where visitors drop off, using your analytics, heatmaps and the ads that send the traffic.
2. Build. They make a page or a change that answers what a visitor asks, in the order they ask it: is this for me, how does it work, can I trust you, what happens next.
3. Test. They split real traffic between the old version and the new one and let it run long enough to mean something.
4. Judge on sales. More enquiries that never turn into customers is not a win, so they check the result against what you actually sold.

Plenty of agencies sell only the first and second jobs. A redesign with no test is a guess with a nice finish. If the proposal has no test in it, you are buying design, and it should be priced and judged as design.

Ask for one result where the test was run properly, then check four things about it.

• Same traffic to both versions. Traffic was split at random between the old page and the new one at the same time. A before-and-after comparison across different months tells you about the season as much as the page.
• A real outcome. The measure is enquiries or sales, not clicks, scroll depth or time on page.
• The length of the test and the number of visits. A test on a few hundred visits can swing either way.
• What counted as a conversion. A form submission and a tap-to-call are different events, and a report that mixes them can inflate the number.

Here is what that looks like when we do it. For Jim's Mowing, we split their Meta ads at random for eight weeks between their main website and a quote page we built. 11.5% of visits to our page led to an enquiry (a form or a tap to call), against 4% on the main website, measured in Google Analytics. Same ads, nearly three times the enquiries. The full breakdown is on our CRO page, including the visit counts. We show the method with the number because the number alone proves nothing.

One result is not a guarantee for your business, and a good agency will say so. Treat any agency that promises a specific lift before they have seen your pages and your traffic as selling, not measuring.

It also helps to know your own baseline before any call. Our guide to what a good conversion rate is explains why a rate only means something against the same traffic and the same definition of a conversion, and that is the standard to hold an agency to. Then look at the pages they have built: our guide to landing page design lists the principles we judge a page on, and you can use the same list on theirs. A page with a menu full of links, a vague headline and a form that asks for ten fields has not been through a serious testing process, whatever the case study says.

These work because a vague answer is obvious.

• "What did you test last, and what did you learn when it lost?" Every real testing programme has losers. An agency with only wins is showing you a highlight reel.
• "How much traffic do I need?" The honest answer is that it depends on the size of the change. Bigger changes show up faster on less traffic. If they say you need no minimum at all, or that it cannot be done below a very high number, push on why.
• "Who builds the page?" Some agencies find the problem and hand you a list. Others build the fix. Know which you are paying for, and who writes the words, since the copy often decides the result.
• "How do you judge success?" The answer should be enquiries or sales, tied to what the lead is worth to you.
• "What happens to the page if we stop working together?" You should own it, and it should still work.

Often yes, and for a service business spending on paid traffic it is usually the better setup. The page and the ad are one sentence in two parts: the ad makes a promise, and the first screen of the page has to finish it. When two different companies own those halves, each blames the other when the numbers are weak.

If you already have an ads agency you are happy with, a standalone CRO agency can still work, provided they get read access to the ad accounts and the ad copy. A CRO agency that cannot see what the ad said cannot fix the gap between the ad and the page. If you want a second opinion on the ads side first, our free ads audit is built for that.

Our related guide on how to choose a digital marketing agency covers the wider questions on contracts, reporting and account access, and they apply here too.

Work out what an extra enquiry is worth to you, then ask whether the work can plausibly produce that. Our ad spend calculator shows how conversion rate changes the cost per enquiry on the traffic you already pay for.

The arithmetic is simple. If your page converts at 2% and the agency gets it to 4%, you get twice the enquiries from the same visits, and each enquiry costs half as much. That is a worked example, not a promise. The point is that conversion rate gains compound on every dollar of traffic you buy, which makes CRO worth more the more you spend on ads.

Prices for NZ CRO work vary widely with scope: a single landing page is a different job to an ongoing testing programme. Be wary of monthly retainers with no defined output. Ask what you will have at the end of month one and month three, in plain terms.

Sometimes the answer is not yet. These are the cases where we would tell you to wait.

• You have almost no traffic. If few people reach the page, you cannot test anything fairly. Fix the obvious leaks yourself (a slow page, a broken form, no phone number on mobile) and build traffic first.
• The offer is the problem. If people read the page and still do not want what you sell, a better page will not fix it. That is a pricing or positioning question, and no test will solve it.
• You cannot follow up leads. More enquiries into a business that answers slowly just means more wasted leads. Fix response speed first.
• You want a rebrand. That is a design project. Call it that and buy it as that.