Web analytics confusion

Web analytics is still a missing art in many businesses, not just retail. Analytics is the last station on the investment train ride and is often compromised to pump more money into direct revenue generating digital marketing like PPC.

But why would any sane person put more money into something they don’t fully understand and for which KPIs may not be optimised? It seems a strange decision.

My gut feeling is that there are too few optimisation specialists Client-side who really get web analytics 2.0. Dashboards are created and reports circulated to tick the analysis box yet limited insight is provided.

If conversion for referral traffic has dropped off the cliff, is that good or bad? I don’t know. Even your data doesn’t know but hidden within are nuggets of insight, you just need the focus and perseverance to find them.

This blog looks at a few examples of how data can be turned into insight to drive commercial decisions.

Conversion is not the be all and end all

Fact: not everyone who visits your site will convert, no matter how amazing the experience and how targeted the user journey. Some people just like to research and browse. So obsessing over conversion metrics as the sole indicator of success is in itself a failure. 

Social Media is a case in point. Most brands’ social media traffic adversely affects KPIs like average order value, revenue and conversion. Why? Social media is better associated with building relationships and increasing content engagement, less so with direct selling. Of course the channel can convert, just look at Dell, but the point is I’ve seen as much as 95% social media referral traffic contribute de nada to the revenue pot if you simply look at direct sales.

What can you do? Well, it’s not a great leap of faith to assume that social traffic is more likely to use social bookmarking to share content. This is easily monitored via web analytics tools including Google Analytics. However, what is the impact of that shared content on your website? Well, by embedding a neat bit of js code into the bookmarks you can monitor people who visit your website from the content that others shared. Now create them as a custom segment and you can evaluate the impact on KPIs like time-on-site, bounce rate etc.

If social media visitors are increasing viral effect, you can start to build a contribution model. With the increased insight, do you know perceive a greater value to your social media visitors?

Avinash Kaushik lists a few neat social media reporting tools in his Occam’s Razor blog. My favourite is Tynt which measures content after it has left your site.

The value of custom segments and data filters

I’ve just done a piece of analysis for a retailer, digesting their Google Analytics data to provide insight into the impact of site search. The top level stats (22.7% new visits use site search with a conversion rate three times that of those who don’t) don’t give me anything juicy to sink my teeth into. I wouldn’t pay me simply to say visitors who use site search convert better, old news my friend. 

However, if I start adding filters to show the top 20 products (defined by total revenue for last 12 months) matched against Google SERPs and then run against the live site search, the light flickers. I can see that 17 of the top 20 don’t even return search results yet on Google exact matches appear in the first 3 pages. Why? A deep dive reveals the problem: issues with the search algorithm and poor data matching.

Taking the core web analytics data, questioning what it tells us, then filtering data and using complimentary analysis has given me insight to take to my client.

The next step is to add engagement metrics from Google Analytics. By adding % search exits, % search refinements and conversion rate by search term, I now build a picture about the impact of using search refinements driven by the merchandising tool.

Relating this back to revenue and conversion sheds light on the relationship between search refinements and conversion; the correlation proved really high, telling me that here is a big opportunity for testing to drive commercial benefit.

(click image for a larger version)

James Gurd1

Creating a custom segment for visitors using search refinements also enables the ongoing analysis of this feature to understand trends. I can recommend reading this blog on Occam’s Razor waxing lyrical about using data to drive actions

In short…

Now i’m not denigrating the importance of conversion optimisation, just saying that web analytics is a much bigger picture. 

The commercial benefits of web analytics are dependent on the quality of planning and level of effort spent asking the right questions and then interpreting data and information. The information above only scratches the surface but hopefully it illustrates the point i wanted to make.

Web analytics gives you the what, voice-of-customer data overlays the why. By blending intelligent data analysis with continuous customer feedback, website owners can piece together an often puzzling yet highly fascinating engagement jigsaw.

So what do you think? Do you get insight from your web analytics or do you struggle to know where to start?

Follow-up reading

There’s a nice post by Avinash on heuristic evaluations of website performance in relation to answering the “Why” question.