Benchmarking in Google Analytics: Using Relative vs Absolute Comparisons of Data

Reading Time: 6 Minutes


I love Google Analytics. It's one of the first tools I ever used when starting out my career in digital analytics and it’s the one tool I continue to use today. It’s beautifully designed, intuitive, customizable and easy to traverse layers of complexity in your data. The product updates they’ve rolled out over the last few years have made a huge impact, from the demographic and interest profiling features and real time monitoring to the more recent Google tag assistant recordings.

But, for all the wonderful things Google does with Analytics, they occasionally get a few things wrong. I was recently playing around with the new(ish) benchmarking report, which rolled out late last year. This, for me, is an essential feature as you need benchmarks for any common metric you report on in order for your data to have meaning. For example, if you opened up GA and found that 100,000 people visited your site last month your first question should be, so what? Was this above or below a short term (e.g. month over month) or long term (e.g. 6, 12 month) average? Benchmarks offer both meaning and context for data that is essential for you to be able to make a decision with it. The finance world does this really well. Just look at any stock chart and you’ll be inundated with benchmarks.

Take a look at the chart below. Beautiful isn’t it. This is the stock chart (for Tableau, #DATA) using Google Finance. Although different stock and finance portals display the data somewhat differently, the fundamentals (or more specifically, technicals) are always there; 1 day to 10 year price comparisons, current price, price range, price over earnings ratio (P/E), and the list goes on. These are all critical benchmarks that investors use to make decisions every day. 

Google Finance Stock Chart

As a side note, I do a fair amount of dashboard design at work and always turn to the finance industry for inspiration on how to present data, especially when it comes to benchmarks. So if you find yourself designing reports in Excel or platforms like Tableau and want some inspiration, check out the stock market!

Ok, so back to Google Analytics (GA). As I mentioned about a year ago Google rolled out a new benchmarking report which you can find by going to the Audience section. 

Google Analytics Benchmarking

When you run this report there are some really interesting things happening. Take a look at the filters at the top of the page.

Google Analytics Benchmarking - Filtering

So GA looks for a relevant industry vertical, in this case, Beauty & Fitness. Then, based on data provided by other sites tracked using Google Analytics you can compare your site against potential competitors or the industry. They even tell you how many web properties contributed to the benchmark. Great! 

You can also make the benchmarks more relevant by selecting a country and site size based on daily traffic volume. The latter is especially important as it ensures you are comparing against other sites that are pulling in roughly the same volume of traffic. Traffic volume isn’t everything, but this is useful as it ensures you compare apples to apples and it can also help you see what kind of performance you can expect as you gain more visitors over time. Worth noting that by default Google Analytics will decide which settings are ideal for you to benchmark against for both the vertical and site size.

This is really helpful data to have as I’m often asked by clients for relevant benchmarks. As a quick side note, there are typically 3 types of benchmarks you should be thinking about:

Historical

Benchmarks based on your own historical performance. Examples include month over month, year over year change, etc

Competitor

Benchmarks based on your competitors. This usually requires you to focus on publicly available data so you can compare apples to apples.

Industry/Category

Benchmarks based on an industry vertical or category. There are lots of ways to approach this, which include getting 3rd party data from vendors (e.g. Alexa, SimilarWeb, etc), or taking the average performance from you plus your competitors to create a custom category benchmark.

I always tell my clients to focus on historical benchmarks first as this data is usually the most accessible and relevant. I’m also generally skeptical about using industry benchmarks provided by 3rd parties as they run the risk of not being relevant enough to your vertical, market or relative company/website size. That said, Google Analytics benchmarking completely addresses my concerns about building/using industry benchmarks as it’s based on real data that you can adjust for vertical, market and size for greater relevance. So this is good news for web analysts.

That said, it’s not all unicorns and rainbows as the feature has a few limitations. One issue I have with the AS-IS benchmarking report in GA is that it relies too heavily on absolute values. For example, I ran the device usage benchmarking report and was presented with the following:

Benchmarking Absolute Comparison

So based on the above GA is telling me that my site is below the industry benchmark for all devices. Which is a bit strange as you’d think you would need to be equal to or above at least one of the categories.  What’s happening here is that GA has compared my site to an absolute standard instead of a relative one, and this can lead to some issues when interpreting the data. All the above table really tells me is that I need to get the volume of traffic up for each device to meet or exceed the benchmark, which isn’t particularly helpful. Let’s look at the data a different way.

Below, I’ve calculated the % share of total traffic each device represents for both my data as well as the benchmark.

Benchmarking Relative Comparison

 

Ah hah, now we’re on to something. By looking at a relative share I now know that I’m above the industry benchmark for tablet and desktop but below for mobile. This, for me, is much more interesting because now I can see how close I am to a trend for device preference. This has lots of applications. For example, I’ve had a few cases where a client was trying to decide whether to invest in building an improved mobile experience on their site but their own historical data didn’t justify the cost as traffic from mobile was historically low. So we turn to the benchmarking report in GA to see if other sites in our vertical are experiencing similar trends in device preference. With the way GA currently presents the data one might look at this and say, well, we’re below benchmark for all devices so let’s get our traffic up across the board (not very helpful). However, by looking at relative comparisons you can see that mobile is the problem area (or opportunity) for you based on the industry.

This isn’t to say that GA’s current approach to representing the benchmarking data is wrong, rather, it doesn’t lend itself to meaningful interpretation that enables you to make a clear decision. Benchmarking is still a relatively new feature in GA and I’m sure the folks at Google are working hard to develop it further. My suggestion, or request, for a future update would be to allow analysts to swap out absolute values with weighted figures (such as % share of total) for volume based metrics like visits, pageviews, etc.

Anyways, the benchmarking report in GA is a great new feature and something you should start experimenting with now if you haven’t already. Especially if your client or C-suite has ever said, “ok, but how do we compare to the industry?”

Have you ever used GA benchmarking before, and if so, did you find it useful? Let me know in the comments below.