---
title: "Conversion Rate Optimisation: What Most People Get Wrong"
description: "What CRO actually is\n\nCRO is the practice of systematically understanding why people don't do the thing you want them to do, and removing the reasons.\n\nNote what that definition doesn't say. It doesn't say testing. It doesn't say landing pages. It doesn't say buttons. Testing is one method CRO uses to validate a hypothesis, but a test without a hypothesis is just gambling with your traffic."
---

[blog](https://demo.hubspotasia.com/blog)

# [Conversion Rate Optimisation: What Most People Get Wrong](https://demo.hubspotasia.com/blog/conversion-rate-optimisation-what-most-people-get-wrong)

 Written by [Account Management Team](https://demo.hubspotasia.com/blog/author/account-management-team) | Aug 26, 2026, 7:44:46 AM

Conversion rate optimisation is one of the most misunderstood disciplines in digital marketing. Not because it's complicated — the core idea is simple — but because the name itself is misleading. "Conversion rate optimisation" sounds like it means *make the conversion rate number go up*. That framing is responsible for most of the bad CRO work in the industry.

Here's what CRO actually is, and the seven things people most commonly get wrong about it.

## What CRO actually is

CRO is the practice of systematically understanding why people don't do the thing you want them to do, and removing the reasons.

Note what that definition doesn't say. It doesn't say testing. It doesn't say landing pages. It doesn't say buttons. Testing is one method CRO uses to validate a hypothesis, but a test without a hypothesis is just gambling with your traffic.

Good CRO is closer to research than to design. You investigate, you form a theory about what's blocking people, you change something specific, and you measure whether the block was real. Most of the value is in the investigation.

## Misconception 1: "CRO means A/B testing"

Testing is the validation step, not the work.

If you run a test without knowing *why* you expect the variant to win, you learn almost nothing when it does. You can't generalise the result, you can't build on it, and you can't tell whether the win was real or noise. You've bought one number at the cost of several weeks of traffic.

The research that precedes a test — session recordings, form analytics, funnel drop-off analysis, customer interviews, support ticket themes, search query reports — is where the insight lives. Testing just confirms whether you read the evidence correctly.

A useful heuristic: if you can't finish the sentence "we believe this will work because we observed that…", you're not ready to test.

## Misconception 2: "A higher conversion rate is always better"

Conversion rate is a ratio, and you can improve any ratio by damaging the denominator.

Cut your paid traffic to branded search only and your conversion rate will jump. Cut your prices by 40% and it will jump again. Add a low-friction "download the brochure" goal and it will jump dramatically. None of these made the business better, and two of them made it worse.

The metric that matters is **revenue per visitor**, or better still, profit per visitor over the customer lifetime. Conversion rate is one input to that. Average order value and retention are the others, and they frequently move in the opposite direction to conversion rate.

The classic example: adding a discount code field to your checkout usually increases conversion rate and decreases revenue, because people leave to hunt for a code and either don't return or return with a discount they'd never have asked for.

## Misconception 3: "Best practices will work for us"

There's a well-known case where changing a checkout button from "Register" to "Continue" produced an enormous revenue increase. It gets cited constantly. What gets left out is that the win came from a specific insight about a specific audience's specific objection to account creation.

Copy the button label without the insight and you get nothing.

Best practices are compressed conclusions from other people's contexts. They're a decent source of hypotheses and a terrible source of decisions. Trust signals help when trust is the barrier. Shortening a form helps when form length is the barrier. If the actual barrier is that your pricing is unclear, you can optimise the form forever and move nothing.

The uncomfortable version of this: your competitors don't know what they're doing either. Copying their page is copying an untested guess.

## Misconception 4: "The test reached significance, so we're done"

This is where most CRO goes quietly wrong, and it's worth being specific about the numbers.

Statistical significance tells you the probability of seeing your result if there were no real difference. It does not tell you the test is finished, and it does not tell you the effect is large enough to matter.

Three failure modes:

**Peeking.** Checking a running test daily and stopping the moment it crosses 95% dramatically inflates your false positive rate — in practice to somewhere around 30%, depending on how often you look. If you decide the end point after seeing the data, the maths no longer holds. Fix the sample size and duration before you start, and don't stop early.

**Insufficient traffic.** To detect a 10% relative improvement on a 2% baseline conversion rate at conventional power, you need roughly 30,000 visitors *per variant*. Many sites simply cannot run meaningful A/B tests on primary conversions. This isn't a reason to give up on CRO — it's a reason to use different methods: qualitative research, funnel analysis, and shipping well-reasoned changes without testing them.

**Too-short duration.** A test that runs for four days misses your weekly cycle. B2B traffic behaves differently on Tuesday than on Saturday. Run for whole weeks, minimum two, ideally through a full purchase cycle.

## Misconception 5: "Most tests win"

They don't. Depending on whose data you believe, somewhere between one in five and one in ten tests produces a meaningful positive result.

This is normal and it's not a sign of failure. It's a sign that the discipline is doing its job — filtering out the changes that felt right but weren't. The failed tests are how you avoid shipping things that would have quietly cost you money.

The practical implication is about expectations. If you promise a client a win every month, you'll end up manufacturing wins from noise. Set the expectation that most tests will be inconclusive, that the value comes from the cumulative learning, and that the occasional real win pays for the whole programme.

## Misconception 6: "The win we got is permanent"

Two things erode test results after launch.

**Novelty effect.** Returning visitors react to change itself. A new layout can lift engagement for a fortnight simply because it's unfamiliar, then settle back. Segmenting new versus returning visitors usually exposes this.

**Regression to the mean.** If you stopped the test at a favourable moment, the "true" effect is smaller than the one you measured. Almost always.

Validated wins should be re-measured a quarter later. A meaningful share won't hold up, and knowing which ones didn't is more valuable than the original result.

## Misconception 7: "CRO happens on the landing page"

The page is where the conversion is measured, not where it's usually lost.

The most common cause of a poor conversion rate isn't the page at all — it's a mismatch between the promise that got someone there and what they found when they arrived. An ad that implies a free trial landing on a page that demands a credit card will convert badly no matter how good the page is. A keyword attracting research-stage traffic to a bottom-of-funnel page will convert badly no matter how good the page is.

For most accounts, targeting and message match are bigger levers than page design. Which means CRO and paid media can't sensibly be run as separate disciplines by separate people who don't talk to each other.

Similarly, the funnel continues past the button. Checkout, onboarding, delivery, follow-up — all of it affects whether a conversion becomes revenue.

## The one that costs the most: aggregate blindness

A test comes back flat. Nothing happened. Into the failed pile.

Except the aggregate result was flat because mobile improved by 15% and desktop declined by 14%, and they cancelled. That's not a failed test — it's two findings hiding behind an average.

Always segment results: device, new versus returning, traffic source, geography, high versus low intent. Be careful about doing it fishing-expedition style after the fact (slice enough ways and something will look significant by chance), but decide your important segments in advance and look at them properly.

## What good practice looks like

- **Research first.** Know what you're testing before you test it, and know why.
- **Combine quantitative and qualitative.** Analytics tells you where people leave. Recordings, surveys, and interviews tell you why.
- **Pick the metric that reflects the business.** Revenue per visitor over conversion rate, wherever you can measure it.
- **Decide sample size and duration before launch.** Then leave the test alone.
- **Expect most tests to fail.** Budget for learning, not for wins.
- **Fix tracking before optimising.** Every conclusion depends on the data being right, and broken tracking is far more common than anyone admits.
- **Treat it as ongoing.** Audiences, competitors, and expectations shift. There's no finished state.

## The short version

CRO isn't about making a number go up. It's about understanding the gap between what your visitors want and what your site asks of them, then closing it — deliberately, with evidence, one validated insight at a time.

The teams that get results from it aren't the ones running the most tests. They're the ones asking the best questions before they run any.

[View full post](https://demo.hubspotasia.com/blog/conversion-rate-optimisation-what-most-people-get-wrong)

```json
{
  "@context" : "http://schema.org",
  "@type" : "BlogPosting",
  "author" : {
    "@type" : "Person",
    "name" : "Account Management Team"
  },
  "dateModified" : "2026-08-26T07:44:47.003Z",
  "datePublished" : "2026-08-26T07:44:46Z",
  "headline" : "Conversion Rate Optimisation: What Most People Get Wrong",
  "image" : {
    "@type" : "ImageObject",
    "height" : 911,
    "url" : "https://24311380.fs1.hubspotusercontent-na1.net/hubfs/24311380/Modern%20wp.webp",
    "width" : 1727
  },
  "mainEntityOfPage" : "https://demo.hubspotasia.com/blog/conversion-rate-optimisation-what-most-people-get-wrong",
  "publisher" : {
    "@type" : "Organization",
    "logo" : {
      "@type" : "ImageObject",
      "height" : 60,
      "url" : "/hs/hsstatic/content_shared_assets/static-1.4092/img/default-amp-logo.png",
      "width" : 60
    },
    "name" : "blog"
  }
}
```