Why conversion rate doesn't reveal the full story

CRO & UX Article

If you want to know how well your online business is doing, it is tempting to simply look at your conversion rate and make a conclusion based on its value. However, the success of your website should not be inferred through a single metric.

On its own, conversion rate is never the full story; there is a myriad of factors that can influence a conversion rate. An understanding of these constantly changing variables can wildly transform your interpretation of this key metric and, ultimately, the success of your business.

So, you’re looking at your analytics and want to improve your conversion rate. You’re thinking of methods that could drive sales, but what is already happening that might be affecting conversions? Before you go and start running optimisation tests, consider the following!

Time

Perhaps the most straightforward reason for a variable conversion rate is seasonality. Random fluctuations are always to be expected, but statistically significant peaks and troughs can be anticipated based on the time of year, month, week or even day - it doesn’t take more than common sense to understand why no one is buying anything at 4 am on a Tuesday in March.

Black Friday and the run-up to Christmas consistently help drive conversions. It also is the exact reason why, in January, protein supplements sell well, whilst alcohol sales drop, as people adopt a ‘new year, new me’ mantra. Such trends can be predicted, though a caveat here is the zeitgeist; the pandemic has also had a profound effect on digital businesses worldwide.

A drop in conversion rate is not necessarily an indication that your website is bad, it may just be the time of year. Understanding why these changes occur may also reveal routes that can be taken advantage of. Accounting for your conversion rate’s dependence on time can not only be reassuring but may also point you in the right direction when it comes to optimisation.

Data manipulation

Fundamentally, how is your conversion rate calculated? The context of the data from your analytics is always important; it is how you deal with the information that dictates your conclusions and measure of success.

A simple example is, are you looking at conversions from new or returning users to your website? Typically, the conversion rate will be much higher from returning customers who have had time to become more familiar with and trusting of your brand and then make a decision.

Are conversion rate variations simply an artefact of the data in your considered date range? This re-emphasises the aspect of time on conversion rates. To avoid adding unnecessary bias in your analysis, ensure that the date range you choose covers expected variations such as random fluctuations and weekdays vs. weekends to an extent that their impact does not produce a skewed data sample.

There are numerous ways that the data can be dissected. Splitting users based on their device (i.e. mobile vs. desktop vs. tablet) tends to show distinct behaviours when it comes to converting. The majority of browsing will occur on mobile, but desktop conversion rates are typically the highest. This does not mean that desktop is ‘better’. The number of conversions overall may still be greater on mobile.

Traffic

The number of people visiting your website can produce some unexpected results when it comes to conversion rate. Say your social media team shared a particularly popular post. Maybe you jumped on a trend and your business briefly went viral. This could lead to an unprecedented increase in traffic. However, even a boost in conversions may result in a lower overall conversion rate if the increase is not proportional to the traffic, e.g. if 1% of 100,000 users convert, this is still more favourable than 10% of 5000 users - I will let you do the math.

In this type of scenario, a conversion rate that is not proportional to traffic is not necessarily surprising due to another underlying factor - the type of traffic. People may be visiting your website who would never normally visit - they are not your standard target audience. It is also common that users who come to your website directly, or through a browser search, will be more likely to convert compared to those that arrive via social media. This reflects their intent.

Summary

The above list covers some of the main reasons why conversion rate on its own is not an adequate metric. The psychological reasoning for website users deciding whether to convert is highly non-trivial. An optimisation test will investigate the effects of a single variable amongst countless numbers. Some of these factors may even be out of your control, but this does not mean that you are at their mercy - if you can predict purchasing behaviour, e.g. by the time of year, then there will be a solution that could turn it into something advantageous.

Conversion rate can help you understand user behaviour on your website, however, it cannot give any insight into how or why traffic goes to your website in the first place.

It is easy to tunnel-vision on the likelihood that a user visiting your website will convert, but there will always be external factors influencing sales. Your conversion strategy may hinge on your optimisation tests, yet a conversion rate without a wider context is meaningless.