Quite often you will receive an email once a month accompanied by automatically produced reporting. This is fine as long as you have the time to trawl through the mass of data and charts provided, in order to firstly work out what it all means and, secondly work out what you should do about it. All too often this challenge is very daunting, and many people settle for checking if visitor numbers are on the increase or not, and do little more.

There is a real need to interpret this data if your marketing campaigns are to be properly managed and optimised for your business. So don’t assume that being sent analytics reports will provide the answers, check to see if the company has analytics savvy staff who can interpret the data for you, and give you actionable recommendations to help improve your marketing efforts. The investment will pay for itself many times over by ensuring your marketing spend is driving real returns for your business.

In order that I could assess my chances of meeting this challenge I asked for access to both the AdWords and Analytics accounts and set to work analysing the current performance. My first port of call was to use use a custome report I created a little while ago. This report simply uses the Metric ‘Total Goal Completions’ and the Dimension ‘Medium’. The report generates a view of all of the conversions resulting from organic, cpc, referral and direct traffic.

This was the beginning and end of my initial investigation as it showed that less that 5% of all conversions arose from cpc traffic! The prospective client had been convinced that his PPC campaign had been driving conversions, so was a little taken aback with my findings. Still the cost of the campaign is now being saved whilst next steps are being considered and a new approach to meeting the objectives is agreed.

Occasionally there are no changes in trends, or all volumetric data remains relatively static from one month to the next. On these occasions it can be difficult to provide the client with any compelling insight that adds value. However convinced you may be that nothing has changed, the simplest expedient of looking at different charting options or segmenting the information in different ways, will almost certainly reveal a new dimension worthy of exploration.

I recently found myself struggling to meet my own objective of providing a meaningful recommendation in every report, when I discovered an anomaly in bounce rates for referral traffic. The overall bounce rate for this source was just over 20%, however the rate for one site was over 60%. Further investigation revealed that this inbound link was from a site that was not a very good fit because it was targeting a different demographic. My client was uncomfortable with the association and an exchange of emails resulted in the link being removed. My client was delighted, and the overall quality of referral links was improved.

So what looked to be a dull and largely repetitive report turned out to provide some useful intelligence for the client.

If you have tripped over anything obscure that provided a useful insight, please comment and share your findings.

Before you start looking at all of the data and charts, it is as well to put some structure on your approach. Just collecting the data without any context for it, is going to be rather meaningless. So what might you be looking for? It is important to set a few Key Performance Indications (KPI’s) so you can determine how your website is performing over time.

If we take visits as an example, what do you want to know? is it total visits, new visitors or returning visitors, or perhaps it is unique visitors? What about bounce rate? This can be very valuable in determining how well key pages perform. Then, of course, there is the question of which pages are most popular, and which ones are the main entry or landing pages to the site.

You will most likely improve the value of this information if you segment it. The first port of call ought to be the source of the visit, was it via a search engine, a referral site or maybe a paid search campaign? You are probably going to want a handle on all of this information on one form or another. In the coming weeks I will delve more deeply into this and start offering guidance on a structured approach to managing all of this data.

New Visitors

A high proportion of new visitors suggests that your marketing is working well to get traffic to the site. From an analytics perspective however we need to dig deeper to understand what happens when these visitors engage with your site. By first creating a new visits segment in you data you can then explore further segments such as:

1. paid and non-paid traffic
2.  keyword usage in each of the above segments
3. top landing pages
4. bounce rates ….

…. and so on. This provides a rich source of information on the effectiveness of marketing campaigns and the reaction of these first time visitors to your website. In my view this is an extremely important area to examine, as the behaviour of new visitors will provide much more information about the effectiveness of your website from a conversion perspective. If these visitors are flowing through your site efficiently, picking up on your calls to action, and engaging immediately via contact forms, or newsletter subscriptions, you can be confident that your site is well designed from a usability perspective.

If on the other hand they spend very little time on site, and your bounce rates in this segment are high, then you really need to review your proposition and perhaps the marketing strategy.

Returning Visitors

A high proportion of returning visitors implies that your site content is engaging and that visitors enjoy coming back. In my experience analysing returning visitors is more difficult. Clearly there is a need to create a returning visitors segment, the further segmentation needs to be different to above however, and I tend to focus mainly on the following:

1. top content
2. goal conversions

I try to identify the most popular pages to determine what is particularly engaging and then to apply this to other pages in the site. If you have tagged downloads, then it can be revealing top see if visitors are returning for the quality of your downloadable content – perhaps this should be monetised in some way. Finally if it is an ecommerce site then repeat purchase information is the most telling statistic to analyse.

Ratios

The discussion inevitably led to the question of what are the best proportions of new to returning visitors. I don’t think there are any correct answers here, however a quick look at my most successful sites suggests that a 75% new to 25% returning ratio is fairly common. In the case of ecommerce sites where there is a strong prospect of repeat business the ratios tend to be much closer to a 50 / 50 split between new and returning.

Any other statistics you might have would be appreciated ……….

Most companies access their own websites for a wide variety of reasons. Each time they do they add to the traffic that is reported in the server logfiles or by the analytics tools that are used for performance measurement. This can considerably distort subsequent analysis, and therefore it is best to ensure this is excluded.

The specific means by which this traffic can be filtered out varies depending upon the analytics tool that is used. The process you need to follow, however is always broadly the same and requires that you identify the external IP address range that is used in your organisation.

The best way to find out your external IP address is to check with your IT Team, or you can use an IP lookup site such as http://whatismyipaddress.com/ or http://whatsmyip.org/.

Once the IP address range is identified you will need to determine exactly how to set up an exclude filter for your analytics toolset.  This is best achieved by reference to the setup instructions for your particular tool. I also ensure that my own external IP address range is excluded from the traffic data for any websites that I provide services for.

Once the filter(s) have been setup you can be confident that all of the traffic is from prospective customers!

I thought it might be worth sharing the approach I used in the hope that it saves someone else a little time.

In this case I am assuming that Google Analytics is being used. The first step is to check which version of the tracking code is in use. If it is the legacy code (urchin.js), then this will need to be replaced with the new code (ga.js), if this approach is to work.

Google Analytics does not track PDF files by default, however this can be overcome by changing the download link on your web pages so that it is seen as a virtual pageview. To do this scan through the source code to find the link for the download, it will look something like this:

<a href=”mybrochure.pdf”>Download my Company Brochure PDF</a>

Replace this link with the following new link which will track the file in the /downloads directory:

<a onclick=”pageTracker._trackPageview(’/downloads/mybrochure.pdf’);” href=”downloads/mybrochure.pdf”>Download my Company Brochure PDF</a>

You will now be able to track the downloads in the Content section of Google Analytics.

Your comments and tips on any similar techniques are very welcome.