Data Led Reasons Conversion Rates Drop

We published this a long time ago…

Some of the content in this post might be out of date, and some images and links may no longer work.

Discover who we are and how we may be able to help you today:

Analytics, it makes monitoring conversion rates simple right? Log in, click into the conversions section, drill down into the overview and there you have it, get a bit more adventurous and you could even compare year on year (YOY) for trend analysis.

It’s all good, unless that is if things don’t look so good and you suddenly find that year on year conversion rates have dropped. That’s when things get tricky, what do you do? While the go to, obvious answers would involve employing a conversion rate optimisation professional or agency and jumping into a load of A/B split testing. The very first step is to interrogate the data that you already have.

The following tips should hopefully provide some ideas that are possibly missed when investigating conversion rates issues.

#1: You integrate another website

This can be done in a multitude of ways. For example if subdomains are merged into the main www profile or a different website is migrated into a subfolder. The effect of this would be to increase session count for the profile but contribute little to no conversions, reducing the profile conversion rate.

Several ways you can spot this:

Comparing YOY look for points where conversions dip
To look for integrated subdomains add a secondary dimension of Hostname

To look for new subfolders use advanced filters, by clicking edit next to the search box ad use a Regex expression to find when new folders where created.

#2: You write more blog posts

Blog posts generally don’t attribute to conversions directly, viral/link bait blog posts in particular will have an even more adverse affect on conversion rates by driving high volumes of traffic that don’t convert immediately.

To check this:

If the blog content was hosted in a specific subfolder you can segment using the Regex technique above and correlate traffic against conversion rates

#3: You change your goal URLs onsite but not in analytics

This one may sound obvious but if you changed goal URLs on site but left your data guys/gals out of the loop, you won’t be tracking conversions and conversion rates correctly. The alternative is also true in that if more goals are added to the profile that’s being used to track overall goal conversion rates this will also affect consistency of reporting.

To overcome this:

For major site changes ensure you have someone involved who owns the data
Ensure there is a profile that is used to report on core KPIs for the site and/or business and set up extra views for testing and diagnosis

#4: You have a growth in ghost traffic / referral spam

Traffic in Analytics can be polluted by spammers. This is mainly done by using the Measurement Protocol to inject bogus data. The following graphic explains it perfectly – taken from Carlos Escalera’s post on Moz

If you notice the following examples in your traffic report you may want to try and correct the spam. While this spam adds to the extra sessions and pageviews for the site, none of it will ever convert and will reduce your conversion rates.

What to do:

Thankfully our very own Emma Barnes has written about getting rid of referrer spam in Google Analytics before
Similarly Carlos Escalera has also written a good post on the subject of ghost spam (UPDATE: Carlos has updated his post on removing Analytics spam)

#5: More mobile traffic?

Mobile traffic continues to grow year on year and this will have an adverse affect on conversion rates. Having a responsive site is a start but may not be enough to capture those lost conversions.

What you can do:

Understand if this is a problem by going to Audience > Mobile > Overview and compared YOY traffic for devices.
Make sure your site works for top mobile devices. Drill down into Mobile > Devices and add a secondary dimension of Screen Resolution

While resolution reporting is a bit hit and miss it still provides an extra level of detail. The retina display on Apple devices are reported as

#6: Prominent phone number CTA

Ok this doesn’t sound so data led, but bear with me.

CTAs or calls to action highlight actions you want a user to take on site. For a lead generation site for example if you start making a phone number much more prominent than a contact form you will funnel users to call the number rather than fill in the contact form, sounds simple enough? At the same time you’ll probably notice your conversion rate drop in Analytics.

The reason:

Forms and pages are easy to track in Analytics. Phone calls not so easy. If phone tracking or event tracking was never a part of your core measurement strategy for site performance. It’s probably time to reformulate your strategy.
There are third party solutions to integrate call tracking within Analytics, but there are also less expensive ways such as event tracking that can be deployed to understand what’s actually driving conversions.

#7: Double tracking when using tag management systems

The interest in tag management solutions, such as Google Tag Manager, have grown greatly over the last few years as websites evolved to be smarter and faster. But this brings its own challenges when it comes to implementation.

While the deployment of tag management systems is seen as relatively straight forward, it’s most likely a few lines of JavaScript code that’s placed throughout the whole site. But if set up incorrectly you risk triggering duplicate sessions, which again will reduce conversion rates.

What to do:

Seek out advice e.g. edit.co.uk/blog/website-migration-guide-a-google-analytics-tag-manager-checklist/ and https://www.viget.com/articles/10-ways-your-google-tag-manager-setup-might-be-broken
debugging tag management systems probably deserves its own post!

Not all conversion rate issues need to resort to A/B split testing straight away. Your data holds the key and will direct you to where you need to concentrate efforts. To fully understand where and why your conversion rate is changing you need to dissect and segment the data. Having conversion rate issues? You may just need to check that your data isn’t lying.