How satisfied are you with your online conversion rate? When asked this question, only 1% of companies responded ‘very satisfied’.

It’s not surprising. The average conversion rate of large UK e-commerce sites is around 5%. This means that for every 100 visitors to your site, only five will buy from you. Smaller sites and lesser-known brands are likely to have much lower conversion rates.

An entire industry has sprung up to address this problem. Conversion Rate Optimisation (CRO) can be defined as the process of increasing online sales by following a data-driven approach. The aim is not to make large-scale changes, but rather to shoot for a series of incremental gains. In the long run, all those “almost nothings” stack up.

Some people are attracted to CRO by claims of simple changes that led to massive results, such as the famous button change that produced an additional $300 million in the first year. Be warned. What you don’t read about is the thousands of cases where those small changes didn’t have any impact.

To be successful, website optimisation should be approached from a different angle. Conversion rate is a reflection of what visitors are doing on the site. If you think of it that way, you should not be concerned with making changes to a page, but rather with influencing people’s behaviour. These are the questions that should drive your optimisation agenda:

• How can you get more of those “non-converting” majority of visitors to buy from you?
• How do you get customers to buy more regularly?
• How do you get them to spend more when they do buy?

This approach requires you to get under the skin of your customers, understand their motivations, their reasons for buying from you and, importantly, reasons for not buying from you. Here are four easy things you can do to help you achieve this:

1. Get to know them with an email survey

An email survey is a great way of getting to know your users quickly. It’s easy to launch, cheap or inexpensive and can generate a large number of responses.

The questions you ask will depend on what you are trying to learn from the exercise, but here are some ideas:

• How would you describe our site/product/brand to a friend?
• At which shops or online stores have you shopped for the product you most recently purchased from our site?
• Why did you buy from us and not another website or shop?
• What do you like about Competitor X?
• What do you like about us compared to Competitor X?
• What has been the best / worst thing about your experience as our customer?
• Thinking back to the last time you shopped on our site, what concerns did you have?

SurveyMonkey is a popular service, but some alternatives include Typeform and Google Forms.

2. Understand their on-site behaviour

Email surveys are good at gathering demographic and attitudinal data, but weaker in terms of behavioural data and exposing usability problems. On-site polling is a better way of sourcing that information. Usually, it takes the form of a question that slides in from the bottom of the webpage. Tools to consider are Hotjar, Qualaroo, Queryz and Informizely.

These polls can be annoying, so limit it to one or two at a time. Open-ended questions are often more insightful. One of our favourites is posted on the payment confirmation page: “Was there anything that almost put you off buying from us today?”

3. Usability testing

Some of our most powerful insights have come from watching users interact with the website. This is known as usability testing.

Use an online panel-based service such as or to get started quickly. You define a task for the user to perform, and within hours you should have videos showing how people use your site, as well as where and why they may be getting stuck.

4. Draw insights from a/b testing

An a/b test shows the original page to half of your visitors, while the rest of the audience sees the page with your changes. This side-by-side comparison allows you to determine with a high degree of confidence whether the changes are having the desired impact.

What many marketers don’t realise about a/b tests, is that it’s a great way to learn about your visitors. Whether your variation performs better or worse than the original, you can learn something from the experiment. Ask yourself why the change you introduced may have caused the detected shift in behaviour, and what that tells you about those visitors.

Tools to look at include VWO,, and ABTasty. On the enterprise side, consider Optimizely, Qubit and Sitespect.

get in touch

six + = 12