The Year Ahead in Web Analytics

Well, it’s 2015 now, and like many of you I’ve been thinking a lot about what the big developments in Web analytics are likely to be in the coming year. Oh, that hasn’t been on your mind? Well, then, worry not: here are my top analytics trends for the coming year!

The Year of Attribution

I’ve been talking with people about attribution modeling for quite a few years now, and it is a constant surprise to me how poorly most organizations understand this critical aspect of their measurement strategy. Even more surprising to me is how little most of the major Web analytics tools offer in terms of attribution modeling.

If you’re not familiar with the term, attribution is basically the process of determining which ad (or ads) should get credit for a conversion. The traditional methods are first click and last click, meaning either the first ad someone clicked through or the last one will get full credit for a conversion.

But, obviously, that’s a very simplistic way to look at things. What about banner ads that someone saw but didn’t click? What if someone had been shopping for something and then at the last minute clicked through a coupon site right before purchase? Complex questions like these deserve complex answers, and there are some very good attribution modeling tools out there to help organizations make sense of this complexity.

However, these tools are by and large standalone products that don’t integrate all that well with either Web analytics tools or marketing automation tools. But some high-profile acquisitions in 2014 by Google, AOL, and now Oracle make me hopeful that we’re going to see a renewed focus on attribution modeling in the year to come.

Ultimately I’d love to see sophisticated attribution settings built in to the major Web analytics tools, but based on who’s been acquiring whom I am not optimistic that 2015 will be the year for that. But if Google were to somehow integrate their new attribution capabilities into Google Analytics (even if only for Premium customers), I think the other Web analytics vendors would feel obligated to come up with better answers for their customers, too.

Swings of the Tag Management Pendulum

Tag management systems (TMSes) are tools that allow you to streamline your tag deployment across your website(s). The “too good to be true” version of the spiel is that you can deploy one tag on your pages, and then write simple rules within the TMS that will automate all of your tracking. Almost no development resources required!

Of course, like most things that sound too good to be true, this is. Some development resources are inevitably required, and whoever’s writing those rules in your TMS will need at least an intermediate understanding of HTML, CSS, and JavaScript. But, on balance, tag management has undoubtedly simplified and rationalized tag deployment for most organizations that use it.

But now that many organizations have been maintaining a tag management infrastructure for a year or more, they’ve started noticing that maintaining a good implementation is not the same challenge as deploying one. Those nice and simple rules built inside the TMS that allowed you to deploy quickly can wind up causing trouble later on, since the site developers often do not know they’re there. So then as the site evolves and the code changes, some of those rules stop working — leading to lost tracking and corrupt data.

For instance, a client of mine had us set up rules in their TMS to track a few dropdown dialogues on their contact form — “what do you need help with today?” and that sort of thing. This worked fine for a few months, but then they changed the form and those rules stopped working. The solution they decided to go with was to shift the responsibility for tracking those fields back onto their website developers.

And often, this is what I’ve started seeing: the same organizations that were initially so excited about how their TMS would absolve their developers of Web analytics duties have started realizing that in the real world, those developers have to be invested in the Web analytics infrastructure or it’s going to break. Sometimes, that might mean going back to more explicit on-page coding; in other cases, maybe training your developers on your TMS would help. But overall, I think the tag management honeymoon is winding down and in 2015 we’ll see a lot more clear-eyed assessments of how to use the technology successfully for the long haul.

Data Layer Standardization

My last prediction might be wishful thinking. But more and more over the past year we’ve started to see traction on the concept of a data layer that would feed data from your website into your analytics tools in a systematic fashion. Right now, there is virtually no consensus about what that data layer should look like; fortunately, though, many of us in the industry are starting to float ideas about how this could be standardized.

One big question, of course, is who gets to set those standards. W3.org has a working group that’s kicking around ideas; some of the other big players are coming up with their own. I’m not too worried about a “VHS vs. Betamax” style fight, because right now even a few competing standards would be better than none! What I’m hoping is that before long we’ll see some of the bigger CMS vendors start outputting standardized data layers automatically, which would allow us to deploy advanced Web analytics on their platforms with very little effort.

Hey, what’s a predictions blog post for if not a little bit of dreaming!

So those are three areas I’m optimistic about seeing real progress in 2015. What are some of the trends you expect to see this year?

Related reading

tencent_emily-ma_featured-image
bounce-370x229
site search hp
ga hp
<