Posted on 25th September 2015 by Kaye Symington in Analytics

Back in the day, we used Search Engine Optimisation techniques to drive what we referred to as ‘traffic’ to websites. This was mainly an exercise in quantity rather than quality coupled with a distinct lack of knowledge on the true value of a visitor to the site and therefore the return on SEO investment. With the birth of free browser tag-based analytics tools and in particular, Google Analytics in 2005, we were given a cost effective opportunity to track which pages this ‘traffic’ viewed and whether any visits ultimately ended up with a conversion such as a sale or lead. Utter genius! Now we knew which keywords delivered traffic that grew business and increased revenue and we could rest on our laurels and leave the work rest of the work to the link building team. Mwah, ha, ha, ha…

That was the age of keyword optimisation.


The more discerning of us quickly realised that some websites were analogous to a leaky bucket. Our clients’ sites were at number one positions in Google SERPs. It was therefore plain as day that we were delivering potential custom through fully optimised keyword campaigns to perfectly Google-y crafted landing pages, yet this traffic just would not play the game. This traffic kept leaving our clients sites without converting!

“My goodness, they searched for blue widgets, we sell blue widgets – why won’t they buy our blue widgets!?”

By becoming more adept at using web analytics packages, we learnt that it was now the content of our websites that we could see hindering conversion potential. Conversion Rate Optimisation tools and consultancies were popping up everywhere to help us understand what made this traffic reject our clients sites whether it be their business proposition, browser incompatibility, web copy or website design. We couldn’t rest on our laurels though this time as we knew that Conversion Rate Optimisation was a ‘kaizen’ process.


That was the age of website optimisation.


At this point, the world was a beautiful place. Clients were happy and so we were happy and then IT happened. Google was ‘…evil’ and…

They stole our bloody keywords!

Once we had dusted ourselves down and wiped away the snot and tears, we looked to clever SEO folks like Dan Barker, who blogged about ways in which we could still gain knowledge around the kind of keywords our traffic were using by viewing and analysing the content and meta data of the landing page. We could still get an idea of what users wanted — we started talking about user intent and we stopped talking about ‘sending traffic’ to our clients’ sites. Instead we started ‘acquiring users’ for our clients. We no longer reviewed ‘content’ reports, but instead analysed the ‘behaviour’ of our users. We stopped talking about ‘traffic’ and started talking about ‘audiences’. We went beyond new vs. returning visitors to segment and label our site users into cohorts. We were able to know if a user was a member of our site, and whether they were logged in or logged out. We were able to tell a retained customer from new customer. We could even tell you the life time value of that customer and by Jove, how many different channels had been assists to that sale. We were going to conquer the World!! …


That was the age of Audience Optimisation.


Then we remembered device attribution…

And that pissed on our bonfire.

One session from a Safari browser on an iPhone could not be connected with the same user in a different session using a Chrome browser on her laptop. We couldn’t really tell where the first click to a site really came from. We couldn’t really tell the order of anymore after that either. The only click we were guaranteed to clearly attribute to a conversion was the last click, which was where we started back in the Keyword Optimisation era. Furthermore, on some devices our users were viewed as one cohort with a particular intent and on another device they would belong to an entirely different cohort. We needed to connect distinct devices.

Google Analytics moved towards solving this conundrum by allowing us to assign user IDs to anonymous users so they can be tracked across browsers and sessions. This solution however, only works if a user has signed up to your website and is logged in (although if a user logs in during a session then the earlier hits can be knitted together) and so can only track users who are in a particular state based on their intentions.
Google have also given us Tag Manager (GTM) to house the plethora of tags needed for all the different tracking packages and code customisations we need to optimise ‘audiences’. Oh yeah, and apparently so we don’t have to speak to developers anymore.

The affordability of cloud services and Business Intelligence solutions has given us the ability to pull together data as information from multiple and distinct sources.


This is the age of Analytics Optimisation.


We are clearer than ever about what we need – a single user ID. How this happens will be a combination of technological innovation coupled with an epoch where we as humans will expect devices to recognise us as individuals. Companies will have to partner to share technologies to enable user authentication for the greater good of the bottom line.

This will be the age of Connection Optimisation.

This article by was posted on 25th September 2015

Sign up to the mailing list for more info about the event
How we use the data in this form.