Forensic Google Analytics: Tools and Tips for UX fixes using data

At the start of every web project, we have the client add us to their Google Analytics account. We dive in looking for some of the basics such as uniques, time on site, bounce, new vs. returning, best-performing pages, and desktop vs. mobile breakdowns to give us (and the client) a high level view of their site’s performance and to give us a baseline to work from for the rest of the IA and UX portions of the design process.

Google Analytics is a powerful tool whose utility often gets overlooked in this effort to give the client some easily digestible, understandable metrics. This isn’t to say the above isn’t important to know, it’s just that some of the most useful data is buried just beneath the surface. To unearth some of the diamonds here, we use a couple of tools.

A few months back we worked on a site where we put these tools and techniques to good use. For context, they are a startup whose product is an enterprise-level piece of software. It’s a complex, but powerful tool that’s generally only discussed in broad terms on the site and is only offered through working with a sales agent because of the complexity and cost of implementation.

 

Tool 1: Behavior Flow

The first of the tools is the Behavior Flow tab. This allows you to see the paths users take while moving through the site. From each page, you can highlight the next page they move to, and the next and, well, you get it. There’s a ton of information here and it can be completely overwhelming. However, if you slow down a bit, think like a user and look at the data, there are some really amazing insights.

Case 1: Sussing out Cross-Pollination

We were interested in how a “Markets” section of a page was performing. Looking at the numbers, it wasn’t unsurprising: the first few (ordered by importance on the site) had the best traffic. However, once you got into the page for these Markets, the only way to view another was to go back and select again. Looking at the numbers, this was exactly what we saw. However, after looking at the second Market, the numbers dropped off significantly with only a single-digit percentage going beyond viewing two Markets.

Looking further at the specific markets, we noticed that users were going to Market pages that weren’t necessarily related (for instance, going from Travel to Finance to Food Delivery.) We surmised that users were looking more for information about how the product was able to meet the needs of a wide variety of use cases and to get a bit more information into how the product itself actually worked (again, because the site only discussed the software in very high-level terms.) Looking further, we saw users frequently going from Markets to Customers and the Case Studies and back, gathering bits of information here and there. But since each of these sections was on different pages and required a lot of going back in the navigation, there was a rather high drop off.

So what actions can we take from these findings? First, we designed Market pages with links to other Markets on them, eliminating the need for a user to go back to get to where they want. We also surfaced related Markets to add more context to their exploration. Additionally, each Market page now contains a Case Study band again to facilitate the user’s journey between these sections.

 

Tool 2: Page Analytics Extension

The second tool is the Page Analytics extension for Chrome. When logged into Analytics and when visiting a site that you have credentials for, this tool will show an overlay of how often links are clicked on a given page. This is hugely helpful in providing context into how parts of a specific page are performing. It’s similar to the Behavior Flow tab, but in the context of the site itself, you can more organically browse the site like a user, opening the door to interesting usage discoveries.

Case 2: The Bad CTA

Our client had a site with a pervasive CTA that stated simply “VIEW DEMO.” This primary action was on nearly every page and was in the main navigation. Clicking this CTA would lead the user to a CRM intake form. The user would fill out the form and land on a “Thank You” page with copy saying they would hear from a salesperson. No demo anywhere in sight. In fact, the viewing of the demo only happened while you were on the sales call with the rep.

Our first thought here was, “I bet you’ve got a lot of let-down prospects.” And using the Page Analytics, we were able to determine this was the case by looking at one point: the percentage of users who again clicked “View Demo” upon landing on the Thank You page. A full 20% of the users went RIGHT BACK to the exact same CTA they’d just chosen. The best performing link on that page was the link they’d just clicked.

To remedy this, our obvious suggestion was to change the name of the CTA since you don’t get to actually get to “view” anything. We opted for “Request a Demo” since it’s more faithful to the interaction and doesn’t get the user’s hopes up. But this still doesn’t get to the heart of the matter: users want to see the software in action. To help streamline their lead funnel, we suggested a landing page where users could see a couple of sample videos of the software in action and from there decide if it was worth it for them to reach out for more information and a formal demonstration.

 

Conclusion

It’s easy to get stuck in the sea of data that Google Analytics provides. But rather than looking at it as data, think about it was users. In a way, it’s like user testing on steroids. If you focus on thinking like a user and use a few specific tools with a pointed purpose, it’s actually quite easy to pull out some really big takeaways that might not have even been possible with traditional user testing.