At the end of 2012, Google introduced some genuinely cool changes to its Analytics products.

These were the kind of changes that get whoops from the audience when announced at the Google event in California (although in the UK the reception was predictably more… British). 

The big thing is a new tracking method: user-centric, multi-sessional, multi-device tracking...

Google is taking Google Analytics (GA) in a new direction, focusing on making analytics more relevant to marketers - the people making decisions about budgets - than to just analysts and webmasters. GA took as its starting point the questions that marketers have been asking of their data for years. Additionally they are making the technology simpler to install and debug. 

The big thing is a new tracking method: user-centric, multi-sessional, multi-device tracking. When I talked about this at Google recently, I used Viv, our client services team leader at Periscopix, to illustrate what the changes mean.

Viv’s favourite website is Etsy. She recently bought a brass lobster from there, apparently this lobster serves no actual practical purpose, but that’s another story. Her journey to that purchase might have looked like this:

  • Viv started by searching for Etsy on her work laptop; she browsed the site, and then registered with Etsy for updates on relevant offers.
  • On her train home, she checked the brass lobster on her iPhone.
  • When she got home, she got out her new iPad to have a final look (everything, even brass lobsters look better on an iPad) and made the purchase.

That’s great for Etsy, which was able to provide all the right browsing experiences for Viv to make her decision. But if I was an analyst, or a marketer at Etsy, Viv is a nightmare; I wouldn’t be able to get a full picture of what she’s doing.

I’d know she started with a Google search for Etsy, browsed the site and registered. And even though she logged in on her iPhone and iPad using her unique registration number (and would then be given a new logged in user ID, that’s not personally identifiable), I wouldn’t be able to connect those actions. I’d see her as three separate users on three separate devices.

That’s not very useful for Etsy. In this instance, they would have learned that iPad users are brilliant and spend lots of money, and that desktop/laptop users don’t spend anything.

The keyword she used originally would have received zero credit and the sale would have been attributed most like to a brand term or direct visit, and yet Viv’s journey started with a search, on a laptop.

It would, in effect, look something like this:

Periscopix - Econsultancy GA post image 1

But there are obviously things that are common across each device that you could connect, in order to see that those three visits are from the same person, and that’s exactly what GA is doing.

It ‘stitches’ the logged in user ID to the original unique user ID, and the purchase to the logged in user ID. The end result is that we can now see one user using three devices and making a purchase.

Now we have something more joined up, like the below:

Periscopix - Econsultancy GA post image 2
This is dynamite for anyone with a transactional website, or a site that requires a registration or login interaction, so you can understand the true user journey.

Google have realised how many practical applications there are for this. For example, it will be possible to stitch online and offline purchase. If you want to buy a sofa, you might research it online, but you’ll probably want to go and sit in it, and bounce up and down to try it out.

Now, you connect the online research with a customer purchase in the shop, if you have (for example) a loyalty card, or customer number, or postcode - anything that can tie that customer to the website activity.

That has implications for attribution, too. For example: Superdrug has a loyalty card. Let’s say Viv looks at nail varnish online, but wants to buy next time she’s in town.

So if she browses online, then buys offline, that offline purchase could be tied to the online activity - such as a search -  that initiated the offline sale. . Good news if you are running a PPC campaign for Superdrug.

In the same launch we learned of so many other exciting new Google Analytics features that should keep us in the world of web analytics very busy for the next 12 months. Think what we can do with free attribution modelling for all, inter-session analysis, sequential intra-session analysis, data imports. The list goes on…

All this is based on a much-simplified tracking system, which makes tracking a lot easier to manage and to spot (and fix) issues, with a reduction in cookies from at least three down to one. Ultimately, this means that you’ll spend less time paying people to fix problems, and more time on proper analysis that will improve the effectiveness of your marketing.

Ashley Burgess

Published 6 February, 2013 by Ashley Burgess

Ben Gott is Head of Web Analytics at Periscopix and a contributor to Econsultancy.

4 more posts from this author

You might be interested in

Comments (18)

Save or Cancel
Tina Judic

Tina Judic, Managing Director at Found

What do you mean by 'logged in user'? Logged into the shop, or to Google?

over 5 years ago


Nick Stamoulis

I can definitely see the value in this. Being able to connect the same visitors across various interactions gives you a much more complete picture. Now if only my "not provided" number stopped going up!

over 5 years ago


Kyle B.

I am also still a little confused as to how this will work. Does the website have to have a login and/or identify users with a unique user ID that is then somehow associated with the GA data? Or how does Google reconcile the information?

over 5 years ago

Alexander Lund

Alexander Lund, Digital Marketing Manager at Beiersdorf

This is great to hear. According to this other article "Store admins can either choose to use the Google-assigned ID or override it with their own ID, which will almost always be a customer login ID"

It will be interesting to see if it takes off and retailers actually implement it. Source:

over 5 years ago



There must be some kind of user login in order for this work. Too bad that is a fraction of user touches.

over 5 years ago



This sounds like a useful feature. Has this been implemented into Google Analytics already?

Also, thank you for such a clear explanation of how this works.

over 5 years ago


Nick Hilditch, Town Clerk at Hythe Town Council

Fascinating article, and looks very promising.

I'd love Econsultancy to update the July 2012 best practise guide to show a step by step gude of how to practically make use of these features.

over 5 years ago

Paul North

Paul North, Head of Content and Strategy at Mediarun

@Tina Logged in to the site.

Ben, are you saying there are more parts to the Attribution/Multi-Channel Funnels area coming? They are useful but limited so far.

over 5 years ago


Rob Willox

It all sounds and is very exciting and something that has long been needed; and partly addressed by Multi-channel funnels recently introduced in GA to help with attribution.

But, in the example used it is predicated on Viv logging in or creating a new account on Etsy on initial visit so that future visits (when logged in) can be tracked.

Unfortunately, not all visitors will be motivated to create an account until they decide to purchase and it is commonly agreed that forcing visitors to register has a measurable negative effect on conversion.

So, until there is a good way to link each and every visit made with some form of individual identifier it will be useful but limited.

over 5 years ago

Nichola Finan

Nichola Finan, digital marketing at Media-input

Sounds amazing, yes, ditto Tina Judic's question on definition of 'logged in', and how could postcode be woven into a positive ID, and link to a cookie dropped following an online search?
Sounds like I need to do some training....

over 5 years ago

Ashley Burgess

Ashley Burgess, Head of Web Analytics at Periscopix

Hi all, thanks for your comments.

The system (as far as we know so far) will rely on a login to the site in question. As Rob mentions above this will limit the usefulness for some sites. Others such as daily deals sites and subscription sites will find this very useful. For the rest it will likely be a case of 'I know x percent login so I will analyse that portion and extrapolate if the number is significant'.

With regards to postcodes etc, you will in the future be able to import data into GA and also fire hits from non-website entities. So you can have your CRM hooked up to your in-store tills and then fire a GA hit when a user makes a purchase in store which will be tied to their online behaviour via their user ID. This user ID should be anonymous in the sense that it can only be linked to user information with access to the CRM. Some more insights into that below:

@Paul - yes we are starting to see multi touch models appearing in MCF along with longer attribution windows (up to 90 days). when mixed in with the above innovations they are going to get more and more powerful.

over 5 years ago


Sangita Joshi

I think this is great - its exactly the backbone of sCRM. The question that will hinder adoption though - is that before attribution, you need the CAPTURE. So, most forms when u buy something offline shd capture SOME unique number - that can then be match on online avatars...

over 5 years ago


Matt Stannard

Really interesting article and discussion.

On the point about being signed in, most people are becoming more accustomed to this, Google have encouraged people to sign in and get a more personalised search experience. Remeber too the shift last year on "sharing" data between Google Apps? Analytics is a Google App. One of the key features of UA is that more work is done server side, so is being logged in that barrier it perhaps once was?

People are also becoming more used to using Twitter / Facebook to connect, in fact 51% of visitors to our site are logged in to a social platform so perhaps these offer an alternative for user identification.

In my view, if the user experience is good then people will sign in / connect if indeed Google can't already identify you ;)


over 5 years ago


Justin Sedgmond

Great article Ben. Do you know if this development enables you to view (and attribute) all sites visited along the path of purchase too? Using your example, lets say Viv started her search on Google via her desktop, but later switched device type and started her search from a comparison shopping engine.

over 5 years ago

Ashley Burgess

Ashley Burgess, Head of Web Analytics at Periscopix

Hi Justin, yes indeed it will. It should make the already good Multi Channel Funnels much more useful.

over 5 years ago

Sophie Hoult

Sophie Hoult, Online & Social Media Executive at DFDS Seaways

The world of data across multiple devices is definitely developing, however we are still at the point of relying on users to consistently log in across all the devices that they are using. For me, there are a couple of areas to overcome. The first is transactional sites that do not require logins until hitting the booking funnel - how can we analyse onsite user behaviour across all the devices? And how about Guest logins - what trends can we learn? Multi device should be as much about the conversion trends & the effect multiple devices have on them, but also, consideration of external factors influencing users' behaviours on each device - for example, the commute home, or weekend browsing.

over 4 years ago


kumpul Drivers, dosen at unsyiah

the information is very helpful friends,
I enjoyed reading it Ashley

over 2 years ago


Phil Hart, Owner at Windows Bulletin

6 years later and Google still can't get it right. Not impressed with Google Attribution so far..


5 months ago

Save or Cancel

Enjoying this article?

Get more just like this, delivered to your inbox.

Keep up to date with the latest analysis, inspiration and learning from the Econsultancy blog with our free Digital Pulse newsletter. You will receive a hand-picked digest of the latest and greatest articles, as well as snippets of new market data, best practice guides and trends research.