Services consulting
Conversion rate optimization: the cheapest growth you already own
CRO means increasing the percentage of visitors who complete the action you want — without buying more traffic. Because the traffic already exists, a 20% relative improvement is typically the cheapest growth available to a business.
What is conversion rate optimization and what is it not?
Conversion rate optimization is the systematic improvement of the percentage of visitors who complete a defined action — buying, booking a call, submitting a form, starting a trial. It is a multiplier on traffic you already have, not a source of traffic.
It is also not, despite how it is often sold, a redesign exercise. Redecorating a page because it looks dated is not a hypothesis. The discipline is closer to engineering than to design: identify a specific mechanism causing loss, form a testable assumption about it, change one thing, and measure.
What should you fix first?
In almost every audit we run, the largest losses come from a small number of mechanical faults rather than from subtle design questions. In approximate order of typical value:
- Broken or unreliable tracking. If you do not know your conversion rate, every subsequent decision is guesswork. This is always first.
- A mismatch between the promise and the page. Someone arrived from an ad promising one thing and landed on a page about something else. This is an ad-to-page problem, not a page-design problem, and it is frequently the single largest recoverable loss.
- Friction in the form. Every unnecessary field, every unclear label, every mobile-hostile input costs completions. Ask only for what you need to take the next step.
- No clarity about the next step. Visitors who cannot tell what happens after clicking do not click.
- Unaddressed objections. Price, risk, switching cost, security — the reasons people hesitate are usually known and usually unaddressed.
- Slow pages on mobile. Checked last in the list but routinely worth several percent.
Design and visual polish come after all six. Not because they do not matter, but because a beautifully redesigned page with a form mismatch underneath it will convert worse than the ugly page that answers the objection.
Which research methods actually produce useful findings?
| Method | What it reveals | Where it fails |
|---|---|---|
| Session recordings and heatmaps | Where attention actually goes, friction points, dead clicks | Small samples mislead. Heatmaps show where people look, not why they left |
| On-site surveys | Stated reasons for abandonment, in the visitor's own words | Self-selection bias — the 3% who respond are not the 97% who left |
| Usability testing | Task completion rates, where comprehension breaks | Five users find most usability problems; they cannot find statistical outliers |
| Funnel and drop-off analysis | The actual quantitative shape of the loss | Tells you where, never why. Needs qualitative work to explain it |
| Support and sales call reviews | Real objections in customers' own language | Only covers people who raised their hand, not those who left quietly |
| Search query and landing page analysis | Intent mismatch between query and page | Requires good data on which query produced which session |
The productive sequence is quantitative first to locate the loss, qualitative second to understand it. Running a heatmap tool on a site with no funnel data produces interesting images and no decisions.
Why do most A/B tests fail?
Two reasons, and both are avoidable. First, statistical power. Most sites do not receive enough traffic to detect a realistic improvement in a reasonable timeframe.
The arithmetic matters. To detect a 20% relative lift on a baseline of 2% with 95% confidence and 80% power, you need roughly 25,000 conversions per variant. For a site converting 50 leads a week, that is about 500 weeks. No amount of patience fixes that — the test is not runnable, and running it anyway produces a result you should not believe.
Second, discipline. Changing multiple elements at once makes the result uninterpretable. Declaring victory on a metric that moved for unrelated reasons — a seasonality effect, a press mention, a tracking change — is how organizations conclude that removing the phone number doubled revenue.
Which page patterns consistently matter?
Across the sites we have reviewed, the following patterns recur often enough to be worth checking first on any page. These are starting hypotheses, not rules, and each should be validated against your own data.
- One clear primary action per page. If there are three equally-weighted calls to action, none of them convert as well as one would.
- The primary action repeated at natural decision points, not just above the fold and in a sticky header.
- Form length reduced to only the fields needed for the next step. Every extra field measurably reduces completion.
- Specific, concrete benefit language rather than abstract claims. "Cut your invoicing time by roughly five hours a month" outperforms "streamline your finance processes".
- Objections handled visibly near the action — price, contract length, data security, what happens after submitting.
- Proof placed where the doubt occurs, which is usually at the action rather than at the top of the page.
- Mobile inputs with the correct keyboard type, autofill attributes and adequate touch target sizes.
Do not test before your tracking works
The most common CRO failure is testing on top of unreliable measurement. Before any experiment, confirm that conversion events fire once per session, that test traffic is excluded from reporting, that offline conversions such as phone calls are captured at all, and that consent mode is not silently dropping a meaningful share of European traffic.
It is worth remembering that phone calls are frequently the primary conversion path for service businesses and are invisible to GA4 without explicit tracking. If that describes you, your conversion rate is currently wrong by a large factor and every decision made on it is compromised. The GA4 guide covers the measurement layer in detail.
Related services
Frequently asked questions
How long does CRO take to show results?
A well-identified mechanical fix — a form shortened, an objection answered, an ad-to-page mismatch corrected — can show within two to four weeks. Structural work takes a quarter or more. A properly powered A/B test needs however long it needs, which may be longer than you are willing to wait; in that case use before-and-after with seasonality controls and say so honestly in your reporting.
How much can conversion rate realistically improve?
It varies enormously and anyone quoting a universal figure is selling. Businesses with a clear ad-to-page mismatch or an over-long form commonly find 20–40% relative improvement available. Well-optimized sites with good traffic quality might find 5–10%. The audit tells you which situation you are in — and if the answer is that little is available, that is useful to know before you spend money finding out.
Do I need enough traffic to run A/B tests?
For conventional tests on your main funnel, usually yes, and this surprises people. Detecting a 20% relative improvement on a low-conversion page needs tens of thousands of conversions per variant. Where traffic does not allow it, use qualitative testing, before-and-after comparisons with seasonality controls, and patterns with prior evidence behind them.
Should we redesign our whole website?
Rarely as a first step, almost always as a way of avoiding a diagnosis. A redesign bundles dozens of changes into one event, so whatever happens you learn very little about why. It also resets accumulated signals and consumes budget that could fund incremental improvements. Find the specific faults first.
Is CRO worth doing if our traffic is low?
Especially then, because it is the cheapest growth available and traffic is the scarce input. But be realistic: if you get 200 visitors a month, even a large relative improvement is a small absolute number, and the highest-return use of your time may be fixing the acquisition problem instead. We will tell you which we think it is.
What is the difference between CRO and landing page optimization?
CRO is the discipline — a measured, evidence-driven process of finding why visitors do not convert and testing changes. Landing page optimization is one technique within it, usually focused on a single campaign or traffic source. The term is often used loosely to mean either.
Do heatmaps and session recordings tell me what to fix?
They tell you where attention goes, which is a useful input but not a diagnosis. Heatmaps show where people looked, not why they left — and a click on a dead element means confusion, not interest. Use them alongside funnel data and user interviews rather than as a decision engine.
Can you improve conversion rate without changing the design at all?
Frequently, yes. The largest losses we find are not visual: a form asking for information you do not need, a landing page that does not match its ad, an absence of clarity about what happens after submission. Those are copy, structure and offer changes. Design is one lever among several, and rarely the largest.
Need conversion rate optimization for your business?
Get a free, no-obligation review of where you stand today on conversion rate optimization — and the three changes that would move the number that matters. No pitch deck.
Book a free 30-minute audit