CRO AGENCY – CONVERSION RATE OPTIMISATION

UX design is about making the experience understandable and usable. Conversion rate optimisation is about increasing the share of visitors who follow through. They often overlap – clearer UX generally raises conversion – but the starting point differs. CRO starts from data on where visitors drop off and prioritises changes by business value. Good conversion rate optimisation requires UX expertise, but not all UX work is conversion rate optimisation.

Rarely in the button colours. On B2B websites, friction comes from what the website does not say and from what it asks of visitors too early. The same patterns recur in every industry.

• The form asks too much, too early. Company registration number, budget and phone number – before the visitor knows what happens once the form is sent. Fewer fields and a clear description of the next step often remove more friction than any design change.
• The offer is unclear. The homepage describes the company, not what the buyer gets, who it is for and what problem it solves. A visitor who has to read three pages to work out whether you are relevant rarely bothers.
• There is no pricing page. B2B buyers want to qualify themselves before getting in touch. "Request a quote" with no indication of price filters out everyone who doesn't have time to book a meeting just to get a figure. A starting price, a pricing model or a worked example is often enough.
• There are too few ways to get in touch. A form is the only way in – no phone number, no named contact and no way to book a call directly. Different buyers want to contact you in different ways.
• The proof is in the wrong place. Case studies and reviews exist, but on a separate page, far from the point where visitors hesitate.

We work on the whole journey from first visit to submitted form, not just a single landing page. On B2B websites with a long sales cycle, we also measure the steps in between – downloads, booked meetings, visits to the pricing page – because a single end goal gives too little data to steer by.

In an online shop, the conversion rate is decided in three places: the product page, the checkout and the delivery and returns information visitors look for before they decide to buy. That is why conversion rate optimisation for e-commerce is less about the homepage and more about the pages where the purchase actually happens.

On the product page, the job is to answer the questions customer service hears most often – fit, dimensions, delivery time, stock levels – on the page itself, not in an FAQ three clicks away. In the checkout, the usual problems are forced account creation, delivery costs that only appear at the last step, too many steps and missing payment methods. Delivery and returns terms should be visible where hesitation arises – in the basket and on the product page – not just in the footer.

Mobile performance matters here too. A slow online shop loses visitors before they have even seen the product, and Core Web Vitals affect both how the shop feels to use and how Google rates it.

If your shop runs on Shopify, we work directly in the theme and the checkout settings. We also help you choose the platform, theme and apps when the shop is being built or rebuilt – read more about how we work as a Shopify agency. How to bring organic search traffic to the shop is covered separately in the guide to SEO for e-commerce.

The first thing we deliver in a CRO project is rarely a new page. It is tracking you can trust. Google Analytics often turns out to be miscounting the very things that matter most, and then there is nothing to optimise against.

The errors are usually mundane but decisive. The conversion is recorded when someone clicks the submit button rather than when the form is successfully sent, so spam and abandoned attempts count as leads. The booking tool sits on another domain and loses the session along the way. The consent banner is configured so that an unknown share of visitors never appears in the statistics. Forms submitted via JavaScript show no thank-you page, so tracking based on the thank-you page records no event at all.

That is why we usually start by reviewing events, key events and consent settings (Consent Mode v2) before drawing any conclusions. We recommend treating that review as part of the deliverable, not a detour: a conversion rate that is measured wrongly leads to the wrong priorities in everything that follows. If you would rather not use GA4 at all for privacy reasons, the same baseline can be built with one of the tools in our comparison of GDPR-friendly analytics tools.

As a rule of thumb, for an A/B test to detect a 20 per cent improvement, each variant needs around 400 conversions – more if the improvement is smaller. The test should also run for at least four full weeks. With two variants, that works out at roughly 1,000 conversions a month in the flow being tested – conversions, not visitors. If you have ten orders a month, a test never becomes reliable before conditions change. That does not make conversion rate optimisation the wrong investment. It means you need a different method.

With low traffic, we replace the test with qualitative methods that show why visitors hesitate, not just that they do:

• Session recordings and heatmaps, which show where visitors stop, scroll past or click on things that are not clickable.
• Form analytics, which show the field where people give up.
• User testing with a handful of people from your target audience, who carry out a concrete task on the website while we observe.
• Conversations with sales and customer service about which objections and questions keep coming up – they often already know what the website fails to explain.
• A structured review of the offer, flows and copy against established principles – the kind of review we describe in the guide to what a UX audit includes.

Unlike a test, this approach measures the effect against the baseline over time, allowing for seasonality and other changes, rather than in a controlled comparison. That is why we prioritise bigger changes – offer, page structure, form flow – over minor tweaks: without a test, small differences cannot be separated from the noise. If an improvement cannot be shown statistically, we report it as a trend, not as a certain result. We explain when conversion rate optimisation is not the right investment at all – for example when the website is going to be rebuilt anyway – in the section on when CRO is not worth it.

A common objection to CRO projects is that they end in a report nobody has time to implement. We recommend keeping analysis and implementation in the same team, and that is how we work: the people who analyse the friction also rebuild the website – in Webflow, Shopify or a coded website with a headless CMS – and follow up afterwards.

It matters for two reasons. The first is pace: an action list that has to be interpreted by another agency or an in-house developer with other priorities often sits untouched for months. The second is quality: whoever framed the hypothesis knows what can and cannot be changed when the change is built. Recommendations that get simplified in the handover often lose exactly what made them work.

Changes that have proved their worth are built permanently into your platform, not left as variants in a testing tool. You should be able to manage whatever we build yourselves afterwards, without waiting for a developer. Where that is not possible – structural changes to a coded website, for example – we say so in advance and help when needed.

Yes – that is the usual starting point. Conversion rate optimisation is normally done on a website that already exists and already has traffic; that is the whole point. We work on most platforms, most often Webflow, Shopify and coded websites in Next.js.

In some cases, though, rebuilding is more economical. When technology, design and messaging are all lagging behind, every individual improvement is expensive and makes little difference. A new website on a modern foundation solves everything at once and makes further work more efficient. If the analysis shows that this is the case, we recommend a new website rather than optimising details on a foundation that will not hold.

We are based in Stockholm, and most of the work is done remotely. We work from data, recordings and your website. Reviews and check-ins happen over video. Your location affects neither the analysis nor the implementation.

What does need your time is knowledge of the business: margins, sales cycle and which leads actually become customers. We gather that at kick-off and in regular check-ins, wherever you are.

Mainly three things: how many flows need analysing, whether tracking has to be fixed before the work can start, and whether you buy a fixed-scope analysis or ongoing work with implementation and follow-up. A one-off analysis is quoted at a fixed price after an initial call. Ongoing work is usually part of an agreement with no lock-in, where the plan is reprioritised based on what is working. We cover market prices and pricing models in Sweden in the guide to what a CRO agency costs.

The first changes – a clearer offer, fewer form fields, faster pages – can have a measurable effect within a few weeks. A structured testing programme takes longer, because each change needs enough data before the result can be trusted. Expect two to three months before you see a meaningful trend, and treat conversion rate optimisation as ongoing work rather than a one-off project.