Head of Google Analytics starts blog

Google Analytics 1 Comment »

Just noticed that Brian Clifton, Head of Web Analytics for EMEA at Google, has started a new blog:

Very interestingly it looks there is a GA book on its way and it appears the blog is setup to support that. Not much content on there yet, Brian?

Omniture buys Touch Clarity

News No Comments »

14th Feb: http://www.omniture.com/touchclarity

It appears Omniture is into some serious acquisitions for 2007. Following on from January’s announcement to purchase the Danish web analytics company Instadia, they now acquire Touch Clarity for $51.5m USD plus stock.

Touch Clarity provide perdictive real time web analytics. Essentially the tool looks at what behaviour a visitor is exhibiting on your web site, then changes the content of the site to suit what ‘it’ thinks the visitor would like to see next.

So for example, Touch Clarity may be able to predict if a visitor is male or female based on the real-time pageview behaviour. If so, it could instruct the back-end CMS to change the contents of a navigation menu, update a ‘view related articles’ list or display different advertisements. All clever stuff.

View the latest market snapshot – vendor timeline.

Google Mini integrates Google Analytics

Google Analytics No Comments »

On the 1st Feb, Google announced its latest release of the Google Mini now integrates directly with Google Analytics.

The Google Mini – and its big brother the Google Search Appliance (GSA), are the “Google in a box” tools that Google sells to businesses wishing to use Google technology to search their own web site or corporate intranet. Previously the Mini had a very basic stats reporting tool but you could customise the XSLT stylesheet and insert your own GA page tracking tag. However that was a manual coding exercise that put many off for fear of breaking the results page.

Now the process is much simpler without the need to manually edit XSLT code.

To set up Analytics reporting in Google Mini/Search Appliance:

When you enable Analytics reporting in your Google Search Appliance, the search appliance automatically adds the Analytics code block to the Front End.

  1. If you do not have an account with Google Analytics, go to:

    https://www.google.com/analytics/ to create a new account. Google Analytics provides you a code block to attach to your web page.

  2. Copy just the account number of the code block. For example, in the following code block, the account number is: UA-123456-1.
  3. Log in to the Admin Console and navigate to the Serving > Front Ends > Output Format page.
  4. Under Page Layout Helper, click the arrow next to Global Attributes .
  5. In the Google Analytics Account Number text box, paste the Analytics account number that you copied.
  6. Click the Save Page Layout Code button.

Thats it. All your generated search results pages will now be tracked within Google Analytics.

Omniture Acquires Instadia

News No Comments »

18th Jan: http://www.omniture.com/instadia

So the acquisitions for 2007 have started in ernest… As we last wrote at the beginning of the year, it looks like 2007 is going to be just as busy as 2006 for mergers and acquisitions. The Instadia acquisition would certainly appear to be an obvious move by Omniture for the following business reasons:

  1. Instadia is a high end product that was pitched directly against Omniture in Europe.
  2. Instadia has a strong market share for enterprise sized clients in the Nordic region and Germany (Omniture is relative weak in those regions).
  3. Instadia has strong market penetration in the Automotive, Manufacturing and Finance verticals

From what we have seen at trade shows and exhibitions, Insatdia also has a strong product offering. Some of its USPs include:

  • The combining of online surveys with web analytics. For example, you can pop a survey in response to behavioural activity or combine survey data with clickstream data and perform segmentation/profiling.
  • Visitor scoring system to measure partial conversions
  • Ad-hoc data reprocessing
  • Real-time reporting

However, I can’t help thinking that Omniture is only interested in Instadia’s client list and staff expertise (instant recruitment for an industry starved of expertise) and that the product itself will slowly disappear.

Will there be more than a handful of WA vendors by the end of 2007?

2006 Web Analytics in Review

Google Analytics, News, Urchin 5 3 Comments »

2006 was a busy time for the web analytics industry in terms of mergers and acquisitions – see vendor timeline (pdf). Probably the most significant event was the announcement by Microsoft in May that it had acquired Deepmetrix – a Canadian based web analytics tool. The similarities with Google’s purchase of Urchin a year before (April 2005) are striking – both in terms of strategy of the parent company and the features of the product they acquired…

Both companies (Microsoft and Google) are major Search Engine players and both have Pay-Per-Click advertising networks. This is a clear signal that online marketing and measurement/accountability are now going to be mainstream. Only a year or so ago, web analytics was considered an optional extra for online marketing campaigns. Now the two will become synonymous. Much of this has already happened with Google Analytics, but it is significant that Microsoft has come to the same conclusion.

Microsoft in October said publicly that they will start to role out their version of Deepmetrix in late Spring 2007. What will be interesting is whether it will be free – as is the case for Google Analytics. I would say almost certainly so, but may only be to adCentre advertisers in the initial phase, as opposed to all users – as was the case for Google Analytics. That would enable them to scale more efficiently than Google did when they had to introduce their throttled invitation system one week after launch (it lasted 9 months!).

Similarities of Urchin v Deepmetrix*

  • Both companies were of a similar size around 30-40 staff.
  • Both companies had a similar sized and loyal customer base.
  • Both companies have a hosted (ASP) service and software solution.
  • Both companies targeted the market at similar price points (mid and high tier clients).
  • Both companies use essentially the same page tagging technology (javascript) to collect data for their hosted service.
  • Both companies have a very close feature set – site overlay, geo-overlay, campaign breakdown, x-segmentation etc.

*Since the launch of Google Analytics its feature set has moved on significantly. Also ,since the notice of the Microsoft acquisition, Deepmetrix is no longer available to trial. So it will be interesting to see what additional features Microsoft have added in the past 12 months – or maybe that time is required for the integration with MS’s adCenter…

Our web analytics market predictions for 2007…

  • MS releases “DeepSoft” or whatever they rename Deepmetrix to. Likely to be free.
  • 100% advertiser adoption – continued rapid growth in adoption of web analytics tools to the point where whenever you setup an online advertising campaign, analytics will be a part of it by default.
  • The number of mergers and acquisitions will continue to increase. With both “DeepSoft” and Google Analytics, are there opportunities for other tools other than large well established vendors that provide heavy customisation i.e. WebSideStory, Omniture, Webtrends. Other tools, excellent though some are (e.g. Indextools, Moniforce, Nedstat, Xiti, Instadia, Redeye) I feel will have to merge or partner to survive – either with other vendors or with agencies that build their marketing services around such products (similar to Webtraffiq and Clicktracks in 2006).
  • As has already started to happen in 2006, web analytics terminology will simplify and become more business orientated and less technical. Expect to see buzz words such as “accessibility” and “discoverability” in the near future. At some point even web analytics as an industry term may disappear as the technique merges with business intelligence (that’s probably 2 years away, but it will happen!).

Happy new year to our subscribers!

Are Your Website Response Times Costing You Sales?

Google Analytics No Comments »

It’s often thought, now that broadband internet access is widespread in the UK, that page loading times are no longer an issue. In fact, exactly the opposite is true. A recent report from Akamai and Jupiter Research highlighted that online users are now so accustomed to speed, that the maximum length of time an average online shopper will wait for a page to load has come down to 4 seconds. Just a few years ago this was 8 seconds.

The problem with widespread broadband internet access is that users no longer blame their internet connections if a page takes a long time to load and so a load time in excess of 4 seconds not only drives customers away, it damages your brand. The key findings from the report:

  • One-third of shoppers with a poor experience abandoned the site
  • 75% were unlikely to shop on that site again
  • 30% of dissatisfied customers will develop a negative perception of the company or tell their friends and family about the experience
  • 65% indicated they are likely to return to a site that is easy to navigate, particularly during the registration, log-in and checkout processes

So to make sure your server response times are fast enough and your buying process is streamlined, you need the right tools. Through our sister company Omega Digital Media we have a Remote Server Monitoring Solution that can test response times for any website, send out email alerts and notifications on server failure and provide weekly and monthly reports, viewable online. We can also provide a more detailed assessment of your server response times as part of a wider website audit.

Of course, Google Analytics has a role to play as well. There are several reports under Content Optimisation -> Web Design Parameters that can tell you what technologies and connection speeds your users have, thereby allowing your web designers to build the site that is best for your market. Configuring tracking on your transaction process and adding this to your reports as a funnel can also highlight any bottlenecks in your purchasing process (Content Optimisation -> Goals & Funnel Process) that are driving customers away.

Online customers are a fickle and impatient breed. Unless you give them the best possible shopping experience, not only will they not buy from your site, they’ll tell their friends not to as well.

How to Assess Search & Content Network Advertising using Google Analytics

Google Analytics 21 Comments »

As a follow up to our recent post on getting detailed PPC keyword data from Google Analytics, we’ve been looking at more ways to improve the granularity of the referral data from PPC campaigns. For example, if your Google AdWords campaign is being served out over the Google content network, your ads may be showing on many different sites, but these won’t show in the GA reports by default – your referrer in these cases will be ‘Google’. This makes it difficult to assess the profitability of search and content network advertising on all the major PPC engines.

Our Show Referral Data filter displays this information by taking the referral data from the URL and adding it to the PPC referrer reports. With this filter in place, your PPC referrals will look like these examples:

google, www.tripadvisor.com[cpc]
ysm, www2.nuseek.com[ppc]
msuk, search.msn.co.uk[ppc]

Once you have this data, you can exclude selected sites from the Google Content Network (instructions on how to do this for Google are here) if they are not providing a good ROI. Unfortunately, you can’t exclude individual sites from the Google Search Network or from YSM’s Content Match system, but this data can still allow you identify whether or not these network are profitable. If not, you can exclude your ads from the whole of the Google Search Network or YSM Content match.

To implement this filter:

1) Create a new profile that is a copy of your existing one.
IMPORTANT: This filter should be ringfenced in its own profile, as you only want this filter applied to PPC referrals.

2) Add a filter to show only PPC referrals in this profile. This screenshot shows how.

3) Add the filter that shows the Search & Content Network referrer, as per this screenshot.

3) Wait for the data. All referral reports will now show the exact PPC referrer.

Let us know your comments or feedback.

How to Get Detailed PPC Keyword Data from Google Analytics

Google Analytics, Hacks and Tools 80 Comments »

At GA-Experts, we also provide PPC Management services through our parent company Omega Digital Media, so we are continually looking at ways to improve the data that Google Analytics provides about a PPC campaign. With this in mind we have developed a new filter that shows exactly the Search Term (i.e. keywords that are searched on) rather than just the Bid Term (i.e. the keywords that triggered a PPC advert).

For example, if a visitor searches for “sony digital camera” and you have an ad configured for the keywords “digital camera” as a Broad Match in Google AdWords, then this search will show your ad. But in your GA reports, the keywords for this search will be reported as “digital camera“. Whilst this gives you good data on your Bid Term, it doesn’t help you refine your keyword triggers because the reports don’t show the actual Search Term.

If you use this Override Bid Term filter then the same search will show “digital camera, (sony+digital+camera)” in any Google Analytics report that shows keywords (e.g. Marketing Optimisation -> Search Engine Marketing -> Overall Keyword Conversion) or any time you cross-segment against keywords.

This data now allows you to fine-tune your PPC campaigns by getting greater detail from your GA reports. In the example above, if the ‘sony+digital+camera‘ search is a frequent one, you might consider creating an Exact Match advert that took visitors to a dedicated Sony landing page. Or replacing the Broad Match ‘digital camera‘ term with a more precise ‘sony digital camera‘ Exact Match to increase your clickthrough rate.

There’s lots that can be done to optimise a campaign using this filter, so please see below for configuration instructions:

1) Create two filters as per the images here:

The Field A regular expression is: (\?|&)(q|p)=([^&]*)

In English, this means:

Look for character immediately following a ? or & (i.e. a URL variable), that is named q or p, and extract the value of this variable.


2) Create a new profile for your existing site. This keeps the detailed keyword data in a separate profile. You can apply this filter to your main profile(s) if you prefer, but a separate profile allows for continuity and top level reporting in your main profile with keyword detail available as required.

3) Apply the two filters to the new profile, in the order Override Bid Term 1 *then* Override Bid Term 2

You can now view detailed Search Term data from any GA report that shows keyword information. Please let us know any feedback via the blog comments.

For further examples of using filters, get a copy of the book Advanced Web Metrics with Google Analytics published by Wiley & Sons, March 2008 (ISBN: 978-0-470-25312-0).


Updated UK & International Search Engine JavaScript for Google Analytics

Google Analytics, Hacks and Tools 11 Comments »

The Google Analytics reports include data about referrals from organic search results. That is, how many people visited your site from search engines like Google and MSN by clicking on one of the non-paid advertising links that make up the bulk of the search results. Google Analytics gets this data from a list of search engines in the urchin.js file, however this list is not necessarily suitable for all geographic regions because it focuses on .com domains and US search engines such as CNN.com.

Update 08-Jan-2009: Please note, this script is now being maintained at:

Because GA Experts is based in the UK and serves a European client base, we thought it would be a good idea to try and update this list to distinguish between UK and international search engines, as well as to enable GA to recognise some country-specific SE’s, for example the BBC’s search engine, or Voila in France. So we have developed two separate JavaScript files that can overwrite the default list of SE’s and replace them with a more up to date and relevant list. One is designed to update and replace the existing list with a range of international SE’s, the other is designed specifically for the UK market.

Once this file is in place, GA will recognise and report on a wider range of search engines, including keyword data and campaign source/referrer data.

If you just want to update your current list of SE’s, then use the International file, if you are targeting a UK or European base then we recommend using the UK file. Both files recognise the same international search engines, however the UK list differentiates between UK and international domains, e.g. ‘google.co.uk’ vs. ‘google.com’ and also adds a few UK specific SE’s, such as the Google web search on the Orange homepage. The full list of search engines and installation instructions follow:


Simply add one of the following lines to your Google Analytics Tracking Code (GATC), *after* the call to the urchin.js file and we will update the file regulalrly on your behalf:
– Many Thanks to Matt Trimmer of iVantage for contributing to this file

<script src=”http://www.omegadm.co.uk/uk_se.js” type=”text/javascript”></script>

<script src=”http://www.omegadm.co.uk/international_se.js” type=”text/javascript”></script>

For example:

<script src=”http://www.google-analytics.com/urchin.js” type=”text/javascript”>
<script src=”http://www.omegadm.co.uk/uk_se.js” type=”text/javascript”>

<script type=”text/javascript”>
_uacct = “UA-xxxx-x”;



The following list of search engines is taken from online research and our own data on the most common UK & International search engines. If you think we have missed any or would like to see any other amendments, please let us know – post a comment below.

International List:
Ezilon.com (European)
SearchEurope.com (European)
Voila (France)
El Mundo (Spain)
Terra (Spain)
Web.de (Germany)
Funnel (South Africa)
Ananzi (South Africa)
Anzwers (Australia & New Zealand)
Rambler (Russia)
Yandex (Russia)
Baidu.com (China)
Indiatimes.com (India)
Araby.com (Arabic)
UK List:
Yahoo UK
altavista UK
Lycos UK
HotBot UK
Excite UK
Ask UK
Orange UK
myway UK
Ezilon.com (European)
SearchEurope.com (European)
Voila (France)
El Mundo (Spain)
Terra (Spain)
Web.de (Germany)
Funnel (South Africa)
Ananzi (South Africa)
Anzwers (Australia & New Zealand)
Rambler (Russia)
Yandex (Russia)
Baidu.com (China)
Indiatimes.com (India)
Araby.com (Arabic)

* AOL.com encrypts keyword data so you will see some strange character strings amongst your other keywords.

Please Note: Use of this JavaScript code is entirely at your own risk. GA Experts have tested this code extensively, however we can take no responsibility for errors, loss of data or any other complications arising from the use of this code

Google’s Commitment to Privacy

Google Analytics No Comments »

We recently blogged on Google Analytics and data privacy, with the intention of correcting some of the public misconceptions about Google and the way it handles analytics data. One of the issues that prompted this debate was the fact that it was possible (with some clever filtering) to show your visitors IP address in the GA reports.

However we have been testing this feature and it is apparent that Google have recently disabled it. Previously, by stripping out the visitor IP address and adding it to the User Defined variable, the IP address would show up in reports. However, now this report (Marketing Optimisation -> Visitor Segment Performance -> User-defined) now shows a value of ‘(not set)’ for all visitors, as it would if the User Defined variable had never been used.

This is another indication of how seriously Google takes data privacy. It is no longer possible to identify individual users by their IP address in Google Analytics. Not only was this disallowed in the Terms of Service, but in the most recent update to the GA backend, Google have been proactive in disabling this feature which had the potential to identify individual users.