Conversion Rate Declines Demystified

Ralf Hiemisch/Getty Images

Only two numbers tell e-commerce companies how well their site is performing: conversion rate and average order value. But behind those two seemingly simple figures are dozens of variables and thousands of data points. Setting aside average order value for a moment, consider just the conversion rate, which is influenced by traffic levels, the traffic source mix, the proportion of new vs. returning visitors, click-through rates, the number of pages in the purchase funnel and even changes to page design.

So on the day that a web analytics manager walks into their office and sees that their conversion rate has dropped, they might feel overwhelmed, like they’re about to jump into an abyss. To know how to fix the problem, they need to know what caused it. To know what caused it is to solve a very complicated puzzle. But it doesn’t have to be so hard…

1. Start the investigation by examining traffic sources. Pull a report from your analytics platform on the amount of traffic coming from each source—natural search, referring sites, bookmarks, direct navigation and any pay per click or banner ads—and their individual conversion rates. The date range of the report should be broad enough to get a sense of historical norms, so that any recent shifts can be identified. From this dataset, you can learn whether conversation rate per channel has dropped off across the board or if, instead, the mix between these traffic sources has changed. Traffic mix changes are important. No two traffic sources attract the same exact level of conversion. So if the overall mix between traffic sources has shifted, the conversion rate for your site would have shifted as well.

2. Next, pull the same report but focus instead on new and returning traffic numbers. Look to see if conversion rates have gone down for both categories or if the traffic mix between the two has changed. It’s likely the conversion rate for new visitors is different from that of returning ones. If the mix has changed, you might be getting a lift in the lower-converting of the two.

If you find that conversion rates have gone down across all traffic sources and for both new and returning visitors, the change is probably not the result of a fluctuation in your traffic mix. The culprit could be something on the site itself.

3. Begin this part of the investigation with a fallout report. Beginning with the homepage, pull the percentage of visitors that click through to another page. Then look at the most clicked-on page and see what percentage of visitors click past it to another page. This should drive you through the pages in your primary conversion path. Now compare the current click-through rates with those in a previous period. Looking at both should tell you whether a specific page has experienced a major drop in conversion. If so, drill down into the changes recently made to the page to see what might be causing the drop. The before and after click-through rates should also show if any new pages have been added that are diverting traffic away from converting.

4. If all of these leads are cold, it’s possible something in your analytics platform has gone awry. This is significantly less common, but definitely a possibility. Page changes can result in broken data capture and inaccurate metrics.

This step-by-step, heuristic analysis will help analytics managers figure out why the conversion rate took a dive. But they’ll still need to make changes that ultimately improve the site’s performance. Fortunately with the right data, it’s easy to know how to focus the solutions.