Google Analytics: Avoiding Premature Speculation by @brettdixon
The numbers don’t lie, but they are often misinterpreted.
Most of us use Google Analytics – it’s free, it’s accurate and it’s brilliant. One of my challenges as an SEO is to regularly educate clients on how to actually interpret Google Analytics and understand what it all means.
With Google recently encrypting keyword referral data and no longer providing us with the terms users used to find our sites, it’s more important than ever to understand key performance indicators and avoid reacting from data that doesn’t paint an accurate picture of the situation.
Many webmasters overreact to individual metrics without considering the bigger picture leading to premature speculation as to whether or not SEO is producing results. So calm down, sit back and look at the situation in detail. Never overreact to partial data, it could literally kill your website.
Don’t prematurely react to Analytics data, take a look at the following examples and look at the bigger picture:
Make sure Google Analytics is set up correctly in the first place.
It’s surprisingly easy to have issues with your data that could prevent complete and accurate data appearing in Google Analytics. You are going to take business critical decisions from this data so it’s absolutely crucial you make sure Google Analytics is giving you correct data, otherwise you’re going to potentially have problems later. Incorrectly configured conversion tracking or pages of your website that are not tracked at all can lead to huge errors when you interpret your data. Get a specialist to take a look at your code implementation to ensure your data is correct.
Connect Google Analytics to Google Webmaster Tools
In the world of 100% keyword (not provided) this is even more important. By connecting Google Analytics and Google Webmaster Tools you can pull data into Google Analytics and understand your impressions in search, average search position and click-through rate for example.
The data isn’t perfect (and it isn’t available for the most recent 2 calendar days) but leveraged with other metrics it does help provide you with a clear view of the bigger picture.
Don’t focus on individual metrics.
It’s very easy to misinterpret individual metrics, in short: don’t do it. Here’s a few examples of how jumping to conclusions could send you in the wrong direction:
If visits are down then it might not mean your SEO is failing. Compare it year to year to see how you performed previously. Are you effected by seasonality? Have you changed your website? Have you put your Analytics code into the new website? (I’ve seen this so many times!)
Bounce Rate up?
If bounce rate is up are you blogging more? You may find your blogs attracting more traffic but many visitors tend to read your blog and then leave, add calls to action to direct visitors to relevant parts of your website to try to encourage them to stick round. Perhaps your traffic isn’t as relevant as it has been previously? Have a look at specific pages and cross-reference them with Google Webmaster Tools to learn what terms are producing specific pages.
Conversions are down since last month?
Are you affected by seasonality? Most online businesses have some form of seasonality. You might feel like you’re having the worst month ever but is it always your worst month? Look back at historical data and see if there is a trend. Have you changed your product offerings? Are you out of stock on your biggest sellers?
Understanding the difference between ‘Visits” and ‘Visitors’
It’s important to understand the difference between ‘visits’ and ‘visitors’. A visitor can visit multiple times so don’t consider visits as the number of unique visitors to your
website: that would be listed as unique visitors in Google Analytics.
I recently had a client complain that the number of visits to their e-commerce website were down and thus the SEO was not effective. What the client didn’t realise (perhaps unbelievably) was that their conversions year on year were up 40% and unique purchases were up meaning their e-commerce conversion rate had increased.
Why? The clients original content was waffle: their blog didn’t really relate to their business and their product / description text wasn’t very relevant to the actual products they were selling: they were over doing it with text in attempt to rank for as many terms as possible: regardless of relevance.
Obviously content is crucial but it has to be relevant, by improving the relevancy of the content and rewriting it so it made more sense to humans led to more relevant and targeted traffic and a higher conversion rate.
Separating Branded and Non Branded Terms
Branded search terms are important and often overlooked by SEO’s and web masters alike. We’re not just talking about your company name here we’re talking about searches that include any of your brand names be them company names or product names.
Given that Google will soon stop passing keyword data completely, this task is about to become more complex – but it is still possible.
Google Webmaster Tools still provides some information on search terms sending traffic to your website but combined with Google AdWords for example you can segment branded from non branded traffic (in a fashion) by bidding on your branded terms. By using impression data you can develop an understanding of the volume of searches that are branded. Using this data (and how it evolves) will help you understand how your SEO is effecting your branded searches which answers questions such as: “We’re seeing a surge in organic traffic but what proportion of it is branded?”
Update: At the time of writing Google Webmaster Tools has stopped providing search query data from the 23rd September – Hopefully this is just a temporary problem and not related to Google no longer providing keyword referral data – fingers crossed!
Landing pages count.
Take a look at this screenshot showing landing pages and the number of visits / pages per visit and average time on site. It helps paint a picture of how these landing pages are contributing to the engagement of your visitors.
Landing pages are the pages of your website a user lands on from a search engine. Often overlooked, regular improvements to landing pages can really make a big difference to your bottom line. Consider if a landing page is giving a searcher a “good experience” would changing text or adding images make it easier to read? Would a stronger call to action improve conversions? Experiment and improve your landing pages regularly and do a better job of converting traffic you already get whilst increasing the terms that result in the page being ranked in search.
Month to Month or Year to Year?
Comparing performance over time is crucial to understand if things are progressing well but you need to make sure you’re comparing ‘Apples to Apples’
Most businesses are affected by seasonality be it Christmas or summer for example. Many E-Commerce websites will find Christmas is their busiest time so comparing July performance to December performance is not giving you an accurate picture of progress.
Equally, simply comparing month to month won’t always give you a good indication of progress. It’s unlikely you will see continuous month-to-month improvements so again, look at the data over the same period a year previously (or even further back).
It is dangerous to misinterpret analytics data: the last thing you want to do is incorrectly assume your SEO is working or not working. You need to react to facts and to do that you need to develop a clear view of the big picture. It’s getting even harder now to accurately interpret data, especially thanks to keyword (not provided) but by taking time to really understand the different metrics and methods available, you can come out on top and avoid premature speculation.