Very informative post from Vicky Brock recently where she writes about the growing importance of the mobile device. In a survey of 22 UK ecommerce sites, Vicky reveals that up to 7% (with a median of 3.5%) of the traffic to these sites is now coming from mobile devices. And what is surprising is that these are figures for general websites – not mobile specific sites or apps.

Mobile Device Type

Apple devices accounted for over 80% of mobile traffic on the surveyed sites. However other devices appear to have a proportionately higher share of conversions (which Vicky suggests may be due to the profile of particular mobile brands matching the profile of high value customers for particular sites).

Usage Patterns

One critical finding was that mobile users are very responsive to out-of-hours email campaigns, for example between 6pm and 11pm, and many sites are generating high proportions of revenue at these times from mobile users. This highlights a very significant opportunity for marketers to target an engaging audience segment.

Mobile Web Analytics in Ireland

Many of the findings above are corroborated in a review of Google Analytics data I undertook recently for a client. During October my client was a guest on a late-evening television show on RTE. During the show there was a significant boost in traffic to his website.

Reviewing the Google Analytics data for the evening in question revealed some interesting findings. Data for the four week period prior to the day of the tv program indicated that mobile devices typically made up 2.3% of site visits. However visit numbers for the evening of the RTE appearance (specifically for the time period 9:00 pm until midnight) showed that mobile visits as a proportion of overall visits were significantly higher at 14.5%.

Graphic showing mobile visits for October. Mobile visits jumped considerably on the evening of the tv appearance.

The report below shows a breakdown by hour of mobile visits to the site on the evening in October:

A section of a Google Analytics report showing visits from mobile visitors broken down into hourly units.

As regards mobile devices used that evening:

• Apple devices accounted for exactly 80% of traffic (which matches Vicky’s survey)
• The iPhone was the most popular device with 67% of mobile visits
• The iPad consisted of 7% of mobile visits. Interestingly the iPad user viewed on average twice as many pages as other mobile device users (6 pages as against 3 pages).

The mobile devices used to access the website on the evening in question.

Summary

It is common knowledge that the mobile phone’s usage as a browsing device is growing considerably. What is interesting however is that the unique pattern of usage and responsiveness of mobile users (particularly during after-office hours) make the mobile user an engaging and potentially lucrative market-segment to target.

Major Contributors: a new intelligence feature in Google Analytics

Google have announced a very useful enhancement to Custom Alerts which provides very valuable information and saves time mining for the data!

If you have a Custom Alert that flags changes in visitor traffic, then Google Analytics will display up to 5 segments that contributed to that traffic change. Google refers to these as Major Contributors.

Google give an example of a Custom Alert that flags a 10% increase in visits. In their example (see graphic above) the alert highlights the following Major Contributors which had an impact on, or were impacted by, the  increased traffic:

• a specific landing page (contributed 88% of the increased traffic)
• a specific website (referred 43% of the increased traffic)
• the specific page on the referring site (referred 40% of the increased traffic)
• the exit page (34% of this traffic left via this page)
• the country (53% of  these visitors originated here)

This enhancement has significantly increased the value of Custom Alerts, and their ease-of-setup should make this a quick win for any website.

$ Index Value is the key to quickly evaluating every page on your website.

Web analytics feature in the current edition of the Irish Marketing Journal. In his article “Web Analytics –  the New Kid on the Market Research Block“, John D Byrne opens with a great quote from Hugh Hefner, who when asked about the success of the Playboy Magazine website said: “Well we had over one million hits in the first couple of days. Problem is, we don’t know if that was a million individuals visiting our site once – or one, visiting us a million times!”

In his piece John states that approximately 70% of the budget of a website is spent on the homepage alone, yet only 10% of visit time on the site is spent on this page. He says that devoting internet expenditure to places where it is likely to be most effective should be a no-brainer for any marketer worthy of the name.

So that got me thinking… what is a fast yet effective method for a marketer to evaluate the effectiveness of every page on a website – in particular those pages which don’t themselves directly deliver a conversion, for example pages such as About Us, Client List, Mission, or Testimonials?

The key to identifying a page’s true effectiveness is to use Google Analytics’ $ Index Value feature, but first we have to look at Goals.

Goals

Typical website goals (non-transactional) are brochure downloads, mailing list sign-ups, and query form submissions. Google Analytics can track when these Goal events occur.

However to get real value out of the Goal feature, we must assign a monetary value to each goal in Google Analytics. So you might assign a value of 10 Euro for a PDF download, 20 Euro for a mailing list subscription, and 30 Euro for a query form submission.

So how is this Goal value used to measure the effectiveness of other pages?

$ Index Value

When a visitor completes a goal (say a download), Google Analytics assigns a value to all pages that were visited up to and including the goal page.

So to give an example, if we have a goal value of 30 Euro for a visit to page F, and a visitor views pages A, D, E, and finally goal page F – these 4 pages will have this 30 Euro value used in the calculation of their value, while all other pages receive no value. (This is simplified – for those interested in the more detailed calculation see the note at the end of the post).

Over time, as visitors continue to convert, $ Index values accumulate for contributing pages. These values can be viewed and compared in the Top Content report, where you can evaluate whether individual pages are contributing or not. There;s no guesswork –  it’s all based on your most valuable visitors – those who convert!

$ Index Value (highlighted in red) as it appears in the Top Content Report. The report can be sorted on $ Index Value to make comparison easier.

Note: I have simplified the calculation of $ Index Value in order to get the essence across. For example ecommerce websites will have transaction value taken into account in calculating $ Index Value. For more information see Google’s post on $ Index.

Econsultancy have published an interview with web analytics guru Eric Peterson. Here are just a few quotes from Eric:

The only company really serious about web analytics anymore is Google with their Google Analytics offering.Google has locked up that sector (web analytics) and continues to innovate, forcing everyone else to look elsewhere for core revenue and driving evolution in the sector.

…senior leadership still hasn’t seen the value that digital measurement and analysis can provide to the broader business.

Without (multivariate, A/B) testing, web analytics falls critically short and fails to provide anywhere near the benefit possible.

The companies that are having success are the ones with clear hypotheses and solid questions they are trying to answer.

The issue was caused by Chrome version 6.0.472.55 incorrectly executing Google Analytics JavaScript leading to artificially high visit numbers for each visit.The issue with the browser was corrected by Google on September 22nd.

To determine if you may have been affected, Google advise users to compare Chrome visit data between September 7th-22nd with a previous date range.

They suggest the following advanced segment to filter out traffic from affected Chrome browsers:

Google suggest using this Advanced Segment to filter out affected Chrome visits

The regular expression in the graphic reads:

6\.0\.472\.(55|56|57|58|59|60|61|62)

This week I got a call from Peter, a website owner, asking if I could take a look at his Google Analytics setup. He had originally set up and configured his own Google Analytics account, and was very familiar with the reports. However he was confused by unusual data that had recently began to appear on his reports.

The Problem

Some weeks previously Peter’s Google Analytics reports had indicated a sudden and sustained increase in the number of visits to his site. “Great” he said and took no further notice.

But the following week while reviewing the reports, he noticed the Keywords report contained a significant number of search terms that were entirely unrelated to his business. That’s when he decided he needed help.

The Cause

I had seen similar data characteristics previously and had an inclination what might be the cause of the problem. The Top Content report revealed another clue: this report contained references to many pages that didn’t exist on Peter’s website!

What was happening was Peter’s Google Analytics profiles were collating data not just from not his own website, but from a third party site also. In this instance the third-party site owner had accidently used Peter’s tracking code ID which resulted in this problem.

How to Prevent This?

To prevent this reoccurring I set up a filter on Peter’s profiles that would include traffic solely from his domain only. He now could be sure that all data being collected was from his site and no other, and that third parties, either accidently or maliciously, could not corrupt his data.

The graphic below shows a filter that ensures that only data from my own domain “analytics.ie” is collected in my Google Analytics account. Be sure and implement something similar in your own account.

Filter including traffic to domain analytics.ie only

• 87% of companies surveyed are using Google Analytics for online measurement
• 38% are use Google Analytics exclusively

The majority of companies are using online measurement to drive actionable recommendations which can make a difference to performance:

• 89% of companies say that web analytics definitely (31%) or sometimes (58%) drive actionable recommendations, an increase on last year’s report.
• 88% of agency respondents think that insights gathered produced useful recommendations.

More details are available on the Econsultancy website.

Faster web page with the new Google Analytics tracking code

The Google Analytics asynchronous tracking code is now the default snippet when setting up new accounts.

For anyone using the old tracking code there is no requirement to upgrade. However Google have identified the following benefits for using the new snippet:

• Faster tracking code load times for web pages due to improved browser execution
• Enhanced data collection and accuracy
• Elimination of tracking errors from dependencies when the JavaScript hasn’t fully loaded

The fact that Google are now considering page load time as a determinant in search engine ranking might also hasten a move to the new tracking code.

Google have provided extensive documentation for those wishing to upgrade their snippets

Identifying Problem Pages with Google Analytics

Cottages4you weren’t happy with the overall conversion rate for their online booking facility. Using Google Analytics they were able to discover that 40% of visitors to the first checkout page were dropping out of the checkout process. This finding was probably identified using the Funnel feature in Google Analytics. This feature allows you examine a visitors path through a series of pages on your site (eg. checkout) and identify where along the path visitors tend to abandon the process.

Funnels can indicate where pages in a process are leaking!

Using Website Optimizer to Improve Pages

Having identified the problem page Cottages4you then reviewed potential problems with the page e.g. excessive text, elements below the fold etc. They produced a series of page designs which varied in terms of content and layout. Website Optimizer was used to present these alternatives to site visitors. Following a 3 week testing period a winning design was evident with a 40% improvement in click-through rate on the original page.

Some Benefits of Website Optimiser

What’s great about Website Optimiser is it uses real visitors to your site to determine what works and what doesn’t work with a page. You and your web designers can hypothesise for forever-and-a-day but Website Optimiser allows the people that matter – your site visitors – to make these decisions for you.

Another great thing about Website Optimiser is it’s free!

Using Google Analytics and Website Optimizer

At the end of their case study Google offer some advice:

1. Use Google Analytics to identify areas of your website that could benefit from testing
2. Decide on some hypotheses you’d like to test and then create some alternative pages to the original
3. Use Website Optimizer to run new page variations concurrently until one begins to perform better than the rest
4. Implement the best performing page and continue testing on other areas of your website

More to Come

We’ll follow up this post with more indepth information on using Website Optimizer for your own testing.

Others have commented that people who want to opt-out can already do so by disabling JavaScript or running browser extensions to block scripts.

The notion of a web analytics opt-out facility is not unique to Google Analytics. Leading web analytics companies including Omniture and Webtrends already offer an opt-out facility.

Forrester suggest some measures which can be taken by businesses to determine how the opt-out might affect their websites. They suggest benchmarking traffic and known visitors before and after the release of the plug-in. They also recommend that website owners should be ready to accommodate the bias that the plug-in may incur on site metrics.

I will comment further as further details are released in the coming weeks.