Conversion Rate Optimisation for low traffic sites

If your ecommerce store doesn’t have enough traffic for AB testing you can still use CRO to improve your sales.

You can’t test anything without traffic.

While each website is different, there are some statistical rules at play. AB testing can be very complicated.The amount of traffic, amount of conversions and the uplift in conversions all matter.

The rule of thumb is approximately 500 conversions per month.

How low is low traffic?

Below is an AB test sample size calculator. As you can see, it gets complicated. But sample size tools can be helpful to find out if your site can handle AB testing.

Basically what the result below tells us is:

Our site has a 2% conv. rate. 1000 visitors per week. We are looking for an increase of 15% in conversions. We want the test to run for 4 weeks (the recommended time). Oh, and we want the winning variation to have an 80% chance of being actually the winning test.

Computer says no. Unfortunately for the above test we’ll need 57,806 users over the 4 weeks (14,451 per week). The alternative is to run the test for 57.81 weeks at 1000 users per week. Over a year! That is out of the question.

But if the increase in conversions we are looking to achieve is 63% we can test with 1000 users per week over 4 weeks.

So if we want to move the needle on a low traffic site we’ll need some big wins. You aren’t going to get big wins by AB testing button colours. Many meaningful changes will need to be made to the online store.

So let’s start looking for those big wins….

Everything starts with research

Optimisation is research driven. Research about your users. What they want and need from your site to buy. We can research in two ways.  Quantitative research and Qualitative research.

Quantitative research is about the data. It tells you what pages they visited and for how long. Whether they made a purchase or not and where they left your site. All very important, but it doesn’t give you the why.

Quantitative research includes:

  • Heatmaps – Where users are clicking on your site.
  • Recordings – Watching people use your site.
  • Analytics analysis – Bounce rate. Shopping funnels. Do mobile devices convert less than desktop computers?

Qualitative research includes:

  • User testing – Watch users interact with your site. Does it make sense? Can they find what they are looking for?
  • Surveys – Asking why people bought from you. Or why they didn’t.
  • Heuristic  analysis – Checking pages for relevancy, motivation and friction.
  • Qualitative research is often the most important tool in improving low converting sites. Results from heatmaps or Google Analytics can be more difficult to interpret with low traffic.

Find out more about the Conversion Optimisation process here

Use the research to look for insights

Identify as many issues as possible and roll out all the changes at once. Small changes will barely make a difference on the low traffic or conversions sites. Whereas big changes can have a large difference on conversions.

If the research is good then the overall average of the changes will make a difference. If we make 30 changes to the site. 15 might improve conversions, 9 might not have any effect and 6 could make conversion rates worse. Overall you’re winning. It’s not ideal, but you have to play with what you’ve got.

Making big changes can make a high enough uplift in analytics to be measurable. 30%, 50% 80% or more conversions. If conversions go up then most of the changes were successful and you’re on the right track.

AB testing results will let you detect much finer increases. 5%, 10%, 20% increases can be detected. With more certainty that increase isn’t down to business cycles, the weather or buying trends.

An agency I have worked with asked me this about a low conversion site:

Rather than just recommending big changes, why can’t we be implementing smaller experiments that can still move the needle, and doing it in a structured way where we can see the impact of the changes and show this to the client. This would be a much better way to build up their confidence in what we’re doing and position larger changes down the track, and to be honest I thought that’s what CRO was about; implementing well structured changes for incremental improvements.

This is putting the cart before the horse.  Any CRO plan, whether the site is big or small, should start with the most impactful changes first. This is especially important for sites with lower sales.

When considering conversion rate optimisation for your online store you’ll need to get comfortable with making sweeping changes to get those first wins.

If people say they can increase your conversion rate on a low traffic site with AB testing they are wrong. The sample size is simply too low. Test times are too long. This leads to the AB test being easily influenced by other factors such as false positives.

Make the most of the traffic to your ecommerce store with a CRO / UX audit.

Creating a roadmap to AB testing

CRO for low traffic sites is about finding and fixing problems that give the biggest uplift in conversion rates.

If the uplift in conversions is large enough the site can become a candidate for AB testing.

If sales are still too low, more research can be conducted. The basis of a good AB test starts with a well research hypothesis. We get ideas for increasing sales from research, AB testing only confirms whether the idea was right.

If you have conducted CRO research, implemented changes and are getting more conversions, then great. This article goes into more depth about AB testing. What it is and what it can do for your store.

FREE VIDEO TRAININGHow to increase sales and profit for your WooCommerce store

FREE VIDEO TRAINING How to increase sales and profit for your WooCommerce store

A simple 5 step plan to increase sales and profit for your WooCommerce store without wasting money on ads

How much money are you leaving on the table?
Find out the 10 key conversion killing mistakes you might be making today.
Learn how I help my clients avoid them and make more sales.
Please enter your name.
Please enter a valid email address.
Something went wrong. Please check your entries and try again.