What Is Conversion Rate Optimisation, Really?

Conversion rate optimisation is one of those terms that everyone in marketing says and almost nobody explains properly. When a founder asks what it is, the answer they usually get is a shrug and the word "testing". That is not wrong. It is also not useful.

This post is the version I wish someone had given me the first time I had to budget for a CRO programme. Plain English. Honest about what it can and cannot do. Separate playbooks for D2C ecommerce and B2B brands who convert through lead forms or distributor sign ups.

If you want the practical, step by step audit to find the conversion leaks in your own business, head to The Conversion Rate Optimisation Audit, step by step. If you want the full service page, it is here.

The plain English definition

Conversion rate optimisation is the deliberate practice of improving the percentage of visitors to your website who do what you want them to do.

That is it. The "what you want them to do" is the conversion. For a D2C ecommerce brand, conversion is usually a completed purchase. For a B2B brand, conversion is usually a qualified enquiry form, a demo request, a trade account application or a distributor sign up.

CRO is the discipline of moving that percentage from whatever it is today to something better, using evidence rather than opinion. It is not a tactic. It is a permanent function.

A simple worked example. A shop runs 100,000 sessions a month with a 1.5 percent conversion rate and a £70 average order value. Monthly revenue is roughly £105,000. If CRO moves conversion to 2.2 percent, same traffic, same AOV, monthly revenue becomes £154,000. Same spend on acquisition. Same team. £49,000 a month more revenue, £588,000 more revenue in a year, with no extra paid media. That is the quiet power of CRO.

Why CRO is the highest leverage spend most brands have

There are three ways to grow ecommerce or B2B pipeline. Get more traffic. Get a higher conversion rate. Get a higher value per conversion. Most brands over invest in the first and under invest in the second and third.

Paid media gets more expensive every year. In 2026 CPMs on Meta, Google and TikTok have risen across every category we track. Traffic growth is becoming the hardest of the three levers to pull. CRO, by contrast, is operating on traffic you already paid for. Every percentage point of conversion improvement drops straight to revenue without a corresponding rise in acquisition spend.

That is why the brands that grow the fastest through 2026 and beyond are spending a larger slice of their marketing budget on CRO than they used to, and a smaller slice on bid inflation. We dig into the commercial maths in CRO in fashion ecommerce: the 10.8 to 1 ROI multiplier most brands ignore and in The luxury CRO discipline.

What CRO is not

It is worth being clear about what CRO is not, because the term gets stretched to cover sins it should not.

CRO is not "making the website look nicer". Aesthetic is part of the work but only when the aesthetic is in the way of conversion. Most of the work is invisible; it is about friction, clarity, trust and speed.

CRO is not "A/B testing". Testing is a tool CRO uses. The discipline is bigger than the tool. Real CRO programmes spend more time on research, customer voice and hypothesis writing than on button colour tests.

CRO is not a one off project. The best CRO programmes run continuously, because customer behaviour, competitive context and technology all change. Brands that treat CRO as a once a year refresh fall behind brands that treat it as always on.

CRO is not magic. It will not fix a bad product, a mispriced offer or a wedge that does not differentiate. If the brand strategy is unclear, CRO will amplify the confusion. That is why we always run CRO in parallel with brand work, not instead of it. For more on that, see The Ultimate Guide to Building a Brand Strategy in 2026.

The four pillars of a proper CRO programme

Real CRO programmes have four pillars, not one. Most in house teams run one of them and call it CRO. That is how brands end up with years of testing activity and very little revenue to show for it.

Pillar one, research. Quantitative analytics (GA4, Shopify, Triple Whale, HubSpot, Amplitude), qualitative research (session recordings, heatmaps, surveys, user testing), customer voice (reviews, support tickets, sales call transcripts) and competitive context. The research is what tells you where the problems are. Without research you are guessing.

Pillar two, hypothesis. A hypothesis is a specific, falsifiable claim about why a conversion problem exists and what change would fix it. "If we add trust signals to the checkout, conversion will improve because cart abandonment survey responses repeatedly mention concerns about returns" is a hypothesis. "Let us change the button colour" is not.

Pillar three, experimentation. Split testing, multivariate testing, holdout testing and, increasingly, AI assisted personalisation. The experimentation is the part most people think CRO is. It is only useful when the first two pillars are solid.

Pillar four, implementation. The winning tests, the lost revenue discoveries and the friction fixes need to go into the site properly, with the development team, with measurement that proves the gain stuck. Implementation is the step where most programmes quietly lose the gains they claim on the reporting deck.

Without all four pillars, you do not have a CRO programme. You have a rolling series of opinions. We see this pattern constantly; we wrote more about why in Why CRO programmes fail.

What CRO typically costs

A proper CRO programme runs between £6,000 and £25,000 a month, depending on the size of the site, the volume of traffic and the ambition of the programme. That range buys research, hypothesis development, experiment design, a developer to build and ship tests, analysis, reporting and implementation planning.

Cheaper than that usually means you are paying for a testing tool and a junior analyst. More expensive than that usually means you are paying for layers you do not need. The Teylu model strips both problems out, which is why clients stay with us; our commercial structure is laid out on the CRO service page.

The honest conversation to have before you start is about the relationship between programme cost and traffic volume. If your site has less than 25,000 sessions a month, A/B testing will be slow and noisy. Most of the value in that case comes from qualitative research, friction removal and implementation of proven best practice, with tests reserved for the highest impact changes. If your site has over 200,000 sessions a month, the programme should be running multiple tests in parallel and compounding rapidly.

What CRO can realistically achieve

Honest benchmarks. These are drawn from our own portfolio across retail, FMCG, sports and B2B.

A well run ecommerce programme should deliver a 15 to 40 percent lift in conversion rate inside 12 months, measured as a weighted average across tested surfaces. Larger sites at the lower end of that range; smaller, less optimised sites at the upper end. For luxury and premium categories the lift is often larger in absolute revenue terms because AOV is higher.

A well run B2B lead form programme should deliver a 25 to 60 percent lift in lead submission rate inside 12 months, with additional lift in lead quality measured downstream at opportunity and close rates. The lead quality gain is where the real money sits; many CRO programmes miss it entirely by measuring only submissions, not downstream revenue.

A well run distributor or partner programme (trade account applications, partner portal sign ups) should deliver a 20 to 45 percent lift in application completion inside 12 months, with a parallel reduction in application quality friction (missing information, rejected applications, manual follow up).

How CRO works for D2C ecommerce

D2C CRO lives on five surfaces: the home page, the category page, the product detail page, the cart and the checkout. Each has its own work.

On the home page the job is clarity of wedge, category entry points that match customer intent, and trust signals. Most D2C home pages try to do too much. A good CRO programme strips the home page back to three to five deliberate decisions per visit.

On the category page the job is browseability. Filters that match the way customers think about the category. Sort options that reflect real buying preferences. Enough product cards visible to feel credible without overwhelming the scroll.

On the product detail page the job is removing objections. Sizing, returns, delivery, reviews, payment options, comparable alternatives, sustainability claims. The product detail page is usually the highest leverage surface in the entire site. Small changes here compound across every inbound traffic source.

On the cart the job is eliminating friction and upselling responsibly. Sales tax and delivery cost need to be visible. Promotions must not create hesitation. Cross sells should feel helpful, not desperate. Guest checkout has to be available and obvious.

On the checkout the job is speed and trust. One page checkouts, with guest flow, with Apple Pay and Shop Pay and Google Pay all offered, with clear error states and a logo wall of known trust marks. Every additional field, every slow load, every unexpected cost costs conversion.

If that is you, read the step by step audit. It will give you the surface by surface inspection list you can run in a weekend.

How CRO works for B2B brands

B2B CRO has a different shape because the conversion is a lead, not a purchase. The funnel stretches post form submission.

Work happens in four surfaces. The homepage and service pages (establish the wedge and the credibility). The consideration content (blog posts, white papers, case studies that validate the brand). The conversion pages (the lead form page, the demo request, the trade application). The post form journey (the thank you page, the sales email sequence, the first call with sales).

The most common mistake B2B brands make is measuring conversion only at the form submission. The real conversion events sit downstream at the MQL, the SQL and the closed won. A CRO programme that does not track conversion through to revenue will happily optimise for more low quality leads, which makes the top of funnel look healthy while the pipeline deteriorates.

For B2B ecommerce and distributor brands, the CRO programme also has to account for the distributor as a customer. Are the trade application forms quick to complete. Is the distributor portal login memorable. Is the onboarding sequence engineered so that a new distributor is active and reordering inside 30 days. Each of those is a CRO problem hiding inside an operations question.

Running CRO in house versus with an agency

Both models can work. The practical question is whether your team has the research, experimentation and development capacity to run all four pillars continuously. Most do not.

In house CRO works when you have at least a mid to senior conversion specialist, a front end developer ring fenced for CRO work, dedicated access to analytics and testing tools, and executive sponsorship that survives a quiet month. If any of those is missing, the programme stalls.

Agency CRO works when the agency runs the full four pillar programme, embeds senior strategists with your team, and shares a reporting frame your CFO will sign off on. If the agency relationship is a junior analyst running A/B tests on a monthly cadence, you will lose money. We wrote more on what good looks like in the CRO service page.

Frequently asked questions

What is a good conversion rate?

Context matters more than the number. For ecommerce, 2 to 3 percent is average, 4 to 5 percent is strong, above 5 percent is excellent and usually category specific (for example, beauty and grocery can exceed 7 percent). For B2B lead forms, 2 to 10 percent is the normal range, with highly targeted campaigns sometimes exceeding 15 percent. The better question is not what is a good conversion rate but whether yours is trending in the right direction.

How long before I see results from CRO?

First wins (friction fixes, implementation of proven best practice, quick survey driven changes) land inside 30 to 60 days. Compounding gains from structured experimentation take 90 days to become visible and 12 months to earn their full return.

Do I need a lot of traffic for CRO to work?

No. Low traffic sites still benefit from research, qualitative insight and friction removal. A/B testing at low volume is slow and noisy, so low traffic CRO programmes lean on those other pillars until testing volume becomes viable.

What is the difference between CRO and UX?

Overlap, not synonym. UX is about designing good experiences. CRO is about maximising a specific commercial outcome. A brilliant UX designer may or may not be a good CRO practitioner. A good CRO practitioner is always attentive to UX but is ultimately judged on revenue.

Is CRO compatible with brand building?

Yes. Some people set them up as opposites, which is a mistake. Brand builds memory and permission. CRO harvests that memory at the moment of decision. Done together they compound. Done separately they conflict. The guide to doing both together is in the ultimate brand strategy guide.

Does CRO work for luxury brands?

Yes, and arguably it is where CRO earns its highest return, because AOV is high and small percentage gains turn into large cash figures. More on this in The luxury CRO discipline.

What tools does a CRO programme need?

A testing platform (Optimizely, AB Tasty, VWO, Convert or Shopify's native tools), analytics (GA4, Shopify, Triple Whale, HubSpot, Amplitude depending on stack), session recording and heatmap tools (Hotjar, Microsoft Clarity, FullStory), survey tools (Hotjar, Typeform or native platform) and a ticketing system to manage the backlog (Jira, Linear, Notion). The tools matter less than the discipline of using them together.

What Is Conversion Rate Optimisation, Really