Conversion Rate Optimisation Strategy: Fix the Funnel, Not the Form

A conversion rate optimisation strategy is a structured, evidence-based process for identifying where potential customers drop out of your funnel and systematically removing those barriers. Done well, it compounds returns from traffic you are already paying for. Done poorly, it becomes a series of button colour tests that consume time and prove nothing.

The distinction matters more than most teams realise. CRO is not a tactic. It is a discipline that sits at the intersection of data, psychology, and commercial judgment, and it rewards the teams who treat it that way.

Why Most CRO Strategies Fail Before They Start

When I walked into my first agency CEO role, one of the first things I did was pull apart the P&L. Not because I was looking for drama, but because I wanted to understand where the business was actually making money and where it was haemorrhaging it quietly. The same instinct applies to CRO. Before you test anything, you need to understand where value is being destroyed in your funnel, and most teams skip that step entirely.

They go straight to execution. They pick a page that feels important, write a hypothesis that is really just a preference, and run a test with a sample size that would not satisfy a GCSE statistics class. When results are inconclusive, they call it a learning and move on. Nothing changes.

The problem is structural. CRO gets treated as a conversion task rather than a commercial one. Teams measure whether the variant beat the control, not whether the improvement moved revenue. Those are not the same question.

If you want to understand how the conversion funnel actually works as a system before building your optimisation strategy around it, CrazyEgg’s breakdown of the conversion funnel is a useful place to orient yourself. The mechanics matter because your strategy has to map to them.

What a Proper CRO Strategy Actually Looks Like

A functioning CRO strategy has four components: diagnosis, prioritisation, experimentation, and iteration. Most teams have a version of the third one and not much else.

Diagnosis means understanding your funnel quantitatively and qualitatively before you form any hypotheses. Where are people dropping off? At what rate? From which traffic sources? On which devices? The quantitative picture tells you where the problem is. The qualitative picture, session recordings, heatmaps, user interviews, tells you why.

Prioritisation means ranking your opportunities by potential impact, implementation effort, and confidence in the hypothesis. There are frameworks for this, PIE and ICE being the most common, but the principle is simple: work on the things most likely to move the needle first, not the things that are easiest to build.

Experimentation means designing tests that can actually answer the question you are asking. That requires a clear hypothesis, a defined success metric, a minimum detectable effect, and enough traffic to reach statistical significance in a reasonable timeframe. If your site does not have the volume to support A/B testing on a specific page, you need a different approach, not a smaller test.

Iteration means treating each test as an input to the next one, not a standalone event. The teams that compound gains from CRO are the ones who build institutional knowledge over time, not the ones who run one test per quarter and wonder why the needle is not moving.

For a broader view of how CRO fits into the full performance marketing picture, including how it connects to paid media, SEO, and analytics, the CRO and Testing hub on The Marketing Juice covers the discipline end to end.

How to Diagnose Your Funnel Before You Test Anything

Diagnosis is where most of the strategic value sits, and it is also where most teams spend the least time. There is a temptation to jump to solutions because solutions feel productive. Diagnosis feels slow. But running the wrong test confidently is not progress.

Start with your funnel data. Map each stage from first touch to conversion and calculate the drop-off rate at each transition. You are looking for the stage where the largest volume of potential customers exits. That is your primary constraint, and that is where your strategy should start.

Segment that data before you draw conclusions. Drop-off rates that look acceptable in aggregate often reveal serious problems when you break them down by traffic source, device type, or geography. I have seen paid search campaigns with a blended conversion rate that looked fine until you separated mobile from desktop and found that mobile was converting at a third of the rate. The aggregate number was masking a significant problem.

Once you know where people are dropping off, you need to understand why. Heatmaps and session recordings are the most direct route to that answer. Hotjar’s approach to funnel optimisation gives a practical overview of how to layer qualitative tools onto your quantitative analysis. The combination of both is what separates a grounded hypothesis from an educated guess.

User interviews are underused in CRO. Most teams rely entirely on behavioural data and never ask a customer what confused them, what nearly stopped them from converting, or what they were comparing you against. Fifteen conversations with recent converters and fifteen with people who abandoned can surface insights that no analytics dashboard will ever show you.

The Prioritisation Problem: Where to Start When Everything Feels Important

When I was growing the agency from around twenty people to closer to a hundred, one of the hardest things to manage was prioritisation. Everything felt urgent. Clients wanted things immediately. The team had more ideas than capacity. The discipline was in deciding what not to do, not just what to do.

CRO strategy has the same problem. Once you have diagnosed your funnel properly, you will have more potential tests than you can run. Prioritisation frameworks exist to solve this, but they only work if you are honest about the inputs.

The PIE framework scores each opportunity on Potential, Importance, and Ease. Potential is how much improvement is possible if you fix the problem. Importance is how much traffic or revenue is affected. Ease is how complex the test is to build and run. Each is scored from one to ten and averaged. The highest-scoring opportunities go to the top of the queue.

The ICE framework is similar but replaces Importance with Confidence, which is how certain you are that the change will have the predicted effect. This is useful if your team tends to overestimate impact and underestimate uncertainty.

Neither framework is perfect. Both require judgment. But both are better than picking tests based on what is easiest to build or what the highest-paid person in the room thinks will work. Unbounce’s breakdown of the right and wrong way to approach CRO makes a similar point about the danger of skipping structured prioritisation.

One additional filter worth applying: ask whether fixing this problem will move a metric that matters to the business, not just a metric that matters to the CRO team. Improving the click-through rate on a secondary CTA is not the same as improving revenue per visitor. Keep the commercial outcome in frame throughout.

Building Tests That Actually Answer the Question

A hypothesis is not a preference. This distinction sounds obvious, but a significant proportion of the tests I have reviewed over the years were built on preferences dressed up as hypotheses. “We think a shorter form will convert better” is not a hypothesis. “Reducing the form from seven fields to three will increase submission rate because our session recordings show users abandoning at the address field, which is not required for the initial enquiry” is a hypothesis.

The structure matters because it forces you to connect the change to a specific observation and a specific mechanism. If you cannot articulate why the change should work, you are not ready to test it.

Sample size is the other place where tests fall apart. Running a test for a week and declaring a winner because the variant is up 12% is not CRO. It is noise. Before you start any test, calculate the minimum sample size required to detect your target effect at a meaningful confidence level. If your traffic volumes cannot support that within a reasonable timeframe, the test is not viable on that page. Either find a higher-traffic entry point or accept that you need a different methodology.

When I was managing significant paid media budgets across multiple markets, the discipline of not reading results early was genuinely difficult. There is pressure to show progress. But calling tests early is one of the most reliable ways to make decisions that feel confident and are actually wrong. The cost of a false positive in CRO is not just a wasted test. It is a change shipped to production that may be actively hurting performance while you assume it is helping.

If your team is newer to structured experimentation, this Unbounce piece on how CRO experts would spend four hours optimising a site is a useful illustration of how experienced practitioners think about test selection and sequencing.

The Funnel Stages That Most Teams Under-Optimise

Most CRO effort concentrates at the bottom of the funnel: checkout pages, contact forms, pricing pages, and landing pages. That focus is not wrong, but it is incomplete. The stages that tend to receive the least attention are often where the most value is being lost.

The top of the funnel is where audience quality is determined. If you are driving the wrong visitors to your site, no amount of landing page optimisation will fix your conversion rate in a way that matters commercially. You might improve the rate at which unqualified visitors submit a form, but that creates a different problem downstream. Semrush’s guide to TOFU, MOFU, and BOFU is a useful reference for thinking about how optimisation priorities should shift at each stage.

The middle of the funnel, where visitors are evaluating your offer against alternatives, is chronically under-optimised. Most teams treat this stage as a content problem rather than a conversion problem. But the pages that sit between initial awareness and final decision are where trust is built or lost. Case studies, comparison content, social proof, and objection handling all live here, and they rarely get the same testing rigour as the checkout flow.

Post-conversion is the stage teams almost never optimise. What happens after someone converts shapes whether they return, refer, or expand their relationship with you. For e-commerce, that means the post-purchase sequence. For B2B, it means the handoff from marketing to sales and then to onboarding. If your CRO strategy stops at the confirmation page, you are leaving a significant portion of the available value untouched.

CrazyEgg’s overview of the website conversion funnel covers how to think about each stage as a system rather than a series of isolated pages, which is the right frame for building a strategy that compounds.

How to Use Qualitative Data Without Getting Lost in It

Qualitative data is the part of CRO that most analytically-oriented teams are least comfortable with. It is messy, it does not aggregate cleanly, and it is easy to cherry-pick. But it is also the only data source that tells you what people were thinking when they made a decision, and that is irreplaceable.

Session recordings are the most accessible starting point. Watch recordings of sessions that ended in abandonment at your highest drop-off point. Look for patterns: where do people pause? Where do they scroll back? Where do they click on things that are not links? Each of these behaviours is a signal that something on the page is creating friction or confusion.

Heatmaps show you where attention concentrates and where it does not. If the element you believe is your primary value proposition is getting no heat, that is worth understanding. Either it is not visible, not credible, or not relevant to the visitors landing on that page.

On-page surveys and exit surveys can surface objections you would not have anticipated from behavioural data alone. A single well-placed question, “What, if anything, stopped you from completing your purchase today?” can generate insights that reshape your entire hypothesis backlog. Hotjar’s CRO tool overview covers how these different qualitative methods work together in practice.

The discipline is in using qualitative data to generate hypotheses, not to validate them. If a session recording shows people abandoning at the shipping cost reveal, that is a hypothesis about pricing transparency, not proof that free shipping will fix your conversion rate. Test the hypothesis. Do not assume the qualitative observation is the complete answer.

When CRO Strategy Connects to Broader Commercial Decisions

The most commercially significant CRO work I have been involved in was not about button colours or headline copy. It was about pricing presentation, offer structure, and the sequencing of information in the buying process. These are decisions that sit at the intersection of marketing and commercial strategy, and they require a different level of seniority and sign-off than a standard A/B test.

When I was working with clients managing hundreds of millions in ad spend, the conversations that moved the needle most were about what we were asking people to commit to at each stage of the funnel. Reducing the initial commitment required to take the first step, whether that was a free trial, a sample, a consultation, or a smaller initial purchase, consistently produced larger conversion improvements than any page-level optimisation.

This is where CRO strategy has to connect to product and commercial decisions. If your offer structure is creating friction that cannot be resolved by page design, the answer is not a better headline. It is a different offer. That conversation requires CRO practitioners to have credibility in commercial terms, not just testing methodology, which is why the discipline benefits from being led by people who understand P&Ls, not just dashboards.

The compounding effect of CRO also changes the economics of paid acquisition in ways that most teams do not model explicitly. A 20% improvement in conversion rate on a paid search campaign does not just reduce your cost per acquisition by 20%. It changes your maximum viable CPC, which changes who you can outbid, which changes your impression share, which changes your volume. The downstream effects are significant and they are rarely captured in the way CRO results are reported to senior stakeholders. That reporting gap is worth closing.

If you are building out a broader CRO capability and want to understand how testing strategy, analytics, and funnel thinking connect as a discipline, the CRO and Testing hub on The Marketing Juice covers the full landscape, from foundational principles to advanced experimentation frameworks.

About the Author

Keith Lacy is a marketing strategist and former agency CEO with 20+ years of experience across agency leadership, performance marketing, and commercial strategy. He writes The Marketing Juice to cut through the noise and share what works.

Frequently Asked Questions