GA is a key tool as the battle to hold on to existing visitor numbers is going to get a lot tougher. Why tougher ? well the biggest driver of traffic, Google natural search, is subject to a whole raft of changes.      1. Keyword information is set to reduce if not to disappear 2. The EU will soon require an explicit opt-in to tracking cookies, for example those on which Google Analytics depends. 3. The Google Panda update still has the potential for a negative effect if not handled properly 4. Competition from other sector providers for Keywords is increasing as more organisations become digital savvy. All this at a time when the really valuable backlinks are no longer so easily available from high rank high authority sites.

This 4 hour training session deals with these issues plus showing how you can get actionable information more easily from the recently upgraded GA. The presentation also looks at how GA can help improve how Google sees your site, be used to measure social media activity plus what KPI`s you should be setting. We will even look at some of the things you need to know that GA cannot tell you and some additional tools that complement GA.

More details on the course here: `Google Analytics for charities`  plus more information and sign up. Note that the first 3 sign ups will receive a 10% discount.

Interestingly this widely uneven distribution of data is quite common. Sometimes know as the power law distribution, the Pareto principle or the 80/20 rule, etc  it is often cited as a `natural ` cause of the 99% /1% wealth distribution we hear so much about these days. Web site data is a classic power law distribution. Site wide averages  such as Average Time on Site or Average Visits per keyword for the whole site give us very little useful information. That is because a huge number of visitors bounce while a small number account for a large percentage of the page views. Rather than site wide averages you want to know more about those who stay, those who come back, etc. So what to do?

1. In Google Analytics use Advanced Segments to isolate groups of visitors who together exhibit interesting behaviour – where they came from, what pages they viewed, etc.

2. Put Goals in for Time on Site at say 1min, 3 min and 5 minutes. This allows you to identify those who liked your content enough to stick around. Those who stay 5 minutes should be of real interest.

The only way to provide data that individual web site contributors can make use of is to break the overall site data into smaller more understandable pieces. So what is the most ( and least) popular section of your site ? How do the visitors who stay the longest find the site ? via search, social media ? But beware any number that is an`average` for the whole site.

UK charity social media sources

increase the overall number of visitors from social media sources to closer to 3%.

In Sept of 2010 I did another analysis of the impact of social media on charity web site visits. This latest analysis uses a slightly wider group of social media sources and charities comparing the period 1st September 2009 to 31st August 2010 with the period 1st September 2010 to 31st August 2011. The data is a simple average across 10 different UK charity sites varying from 10k to 250k visits per month. The data was collected from Google Analytics using an Advanced Segment as described in the previous post. For all the charities the Top 5 social media sources comprised around 80% of all the sources in the Advanced Segment. So the 80/20 rule applies and there is no long tail of social media referrals.

The Social Media Advanced Segment that I used covers many kinds of sources but three distinct groups can be identified. The first might be considered true social media. Here the big two, Facebook and Twitter, are becoming more dominant but for some charities Linkedin brought many visitors. The second group is social bookmarking sites where the two that did well were Stumbleupon and Netvibes. If you are not using them currently they are worth investigating as they have been very successful for one or two charities and as this report shows for national newspapers such as The Guardian and the Daily Mail. The third broad group is Communities and Forums. Examples are in-house Forums plus places such as moneysavingexpert and sector specific forums. As the sites I reviewed used different referral methods the numbers coming from in-house communities and forums may be undercounted. Finally, Google+ sends very few visits directly to the sites that I reviewed and to my surprise YouTube did not feature in any Top 5 list of referring social media sites in the last 2 years.

Google Reading Level - will your readers understand your content ?

You can see from the image where to locate it.  Once selected it will display the reading level for each page in the SERPS results and the overall level for all results. You can check the reading level of your whole site by using the Site Operator. Syntax: [site:www.example.com] and selecting the Reading Level option will bring up the details as shown. Note that you can check the reading level of any section of your site by using [site:mysite.com/folder1] and indeed you can check individual pages.  If a popular landing page or one specifically designed for hard to reach groups has a high bounce rate it is worth checking the reading level to see if things can be improved.

Google explains how the reading levels are arrived at: ” We paid teachers to classify pages for different reading levels, and then took their classifications to build a statistical model. With this model, we can compare the words on any webpage with the words in the model to classify reading levels ” In addition they say ” Roughly speaking, “Basic” is elementary level texts, while “Intermediate” is anything above that level  up to technical and scholarly articles, a la the articles you’d find in Scholar.  “

Google Analytics real time Overview Screen (Click for larger image)

1. Improve response to EMails and Tweets. Testing variations in EMails is a great way to increase response rates. If you do a test to say 5% of your list and watch the results you can immediately see what sections, layouts and links get the best ( and worst) response. If you are keen you can continue to test for optimum results. Similarly you can test Tweet variations containing embedded links.

2. Take advantage of hot news events: If an article is getting a lot of visits say as a result of a news report on a crisis then creating a link on the home page immediately will enable the content to be found easily by more people. You can even benefit if a celeb is in the news and they are an ambassador for your charity.

3. Convince content creators: I think that this may be a big side benefit. One of the biggest hurdles at many charities is convincing those who can create great content that doing so is a worthwhile and effective use of their time. While they know that people read web posts seeing it in real time is quite powerful. Perhaps the same as telling a donor that you will do good things with their money and showing an effect in real time – they believe you but seeing makes it more real. I think this R-T feature has the potential to engage and convince contributors. Particularly if you Tweet or Email about their content and they watch how people responded to their efforts. (More details on GA R-T from Jason.)

Social Media spreads the word

1. Direct Referrals: You can be sure that visitors have read and reacted to a post if they click a link in it.  There are many ( too many ?) link shortners which allow you to track how many have clicked on a link in your post. Twitter`s proprietary shortner, t.co, is now fully operational  and automatically works on any links that are included in a Tweet. Bit.ly remains the most popular choice as it can be used to log how many people clicked on links to sites and pages other than your own. You can see visits that have come to your site via shortners in your Google Analytics account. GA >Traffic Sources> Sources > Referrals

2. Better search engine rankings: Social media activity is now clearly affecting search engine rankings. Postings and reaction to them are becoming a more important part of the collection of `signals` that search engines use to determine what content to display for a given search query. Ian Laurie has noticed that Google is now putting more emphasis on Twitter and Facebook account activity. So a Google search for a brand or charity plus a top social media site such as Twitter or Facebook eg `rnib Twitter`, or `breast cancer care Facebook` now produces results which include sitelinks. Normally Google only shows site links for Organic and/or AdWords results for highly regarded sites. Ian`s view is that sitelinks are shown only for those SM accounts that have some unique content. A quick check that I did on a number of charity accounts seems to bear that out. Over at SEOMoz they did some tests to check if social signals drive search engine traffic and they found that they did. However they do suggest that Google +1, which has yet to be widely used by visitors to charity sites, is a big part of it.

3. More Backlinks to your content from authority sites is still the most powerful signal to Google that your content is worth showing. Without backlinks it is almost impossible to rank in Google. To date charities, who normally have well written content, could rely on media and government site to mention and link to their content. But the web is daily becoming a more competitive place so social media is a great way to identify key influencers and get them to link permanently to your content. If you can use SM to build a relationship that leads to a link from an important site or blogger then you will get visits and kudos with Google. So, long after your Tweet or FB post has been forgotten the backlink will still be sending visitors and convincing Google that your content deserves to be shown. Backlinks: the gift that keeps on giving.

Image by inanpw at Flickr http://bit.ly/q9KqTC

The table shows the Twitter Influence of each account as measured by 4 well know tools. In addition, for the 11 charities, I have included  data on the number of referrals to the web site from Twitter and as a further measure of influence the final column shows the the number of referrals per follower. Here low is good.

Twitter `Influence` and performance for 11 charity sites

The 11 accounts shown have individual web sites and Twitter accounts. Each Twitter account has between 1k and 10k Followers and has sent between 1k and 5k Tweets since they started which is mostly between late 2008 and early 2009. The accounts are divided into two types. 1 to 6 are information and advice oriented – think debt info, pregnancy advice, etc. Accounts 7 to 11 deal with conditions -think conditions related to heart, lung, MS, etc. For all accounts I have listed 4 measures of Twitter influence as calculated by – Klout, Twitalyzer, Retweet Rank and Peer Index.

Also shown is the Follower Ratio which is the number of Followers divided by the number Following. GA:refer/Twitter is the number of visitors to the web site taken from Google Analytics for the period Jan to June 2011. ( Traffic Sources> Sources> Referrals>Twitter). The final column is the ratio of Followers/Referrals ie the number of Followers divided by the number of referrals to the web site by Twitter. So if you had 10 Followers and 5 referrals your score would be 2.0, 10 Followers and 10 referrals your score would be 1.0. So lower numbers show that a higher percentage of your followers are interested enough in your Tweets to click on a link and visit your site.

The Klout and Twitalyzer ratings for both groups are quite similar. Retweet Rank which is “the number of times users been retweeted recently” varies quite a bit probably because of the variation in recent activity on each account .  PeerIndex which is “a single measure of status ” Overall the ratings for both groups are quite similar although those concerned with conditions do slightly better. Overall these numbers are good indicators of where you should expect your Twitter account to be.  With their lower scores the `communication` charities are, as might be expected, somewhat better at getting Twitter followers to visit their web site.

You can check how your account compares by using the free web checkers from Klout, Twitalyzer, Retweet Rank and Peer Index. In addition you can see how effective you are at driving traffic to your web site by checking your Twitter referrals in GA for the same period.

Content rich sites such as nfp, ngo, charity, news, etc may have much better Bounce Rates and hence user appeal than the data suggests.     The content on these sites is carefully crafted and checked, often indeed meeting approved standards for clarity and accuracy. Yet often these pages have the highest Bounce Rate which for those using Google Analytics is seen as a negative KPI.

However a high Bounce Rate for a page may not be telling the whole story. Google Analytics in its standard form cannot distinguish between a visitor to a page who leaves it within say 10 seconds and a visitor who stays 4/5 minutes or longer and then leaves. Both are bounces, defined by Google as “..the percentage of single-page visits or visits in which the person left your site from the entrance (landing) page…” The former is clearly a real bounce as the visitor did not find what they were looking for on the page nor did they see a reason to go elsewhere on the site. However the latter is clearly someone who finds the content engaging and useful and should not be thought of in the same way as the sub 10 or 20 second visitor. Brian Cray gives more technical details.

Bounce rate KPIs are mostly derived from experience on commercial sites where you are looking for a visitor to land, move through a funnel and complete a purchase. Content oriented sites have lots of destination pages. These are pages that provide visitors with detailed specific information. Many visitors will take the time to read all of the page content and then leave. Given this these pages and sites will have a higher than usual Bounce Rate.

Real effect of lower Bounce Rate

Google Analytics can be easily adapted to distinguish between real and `fake` bounce visits by adding a small piece of code to the GA tracking code that is already on your pages. This is a task for the web dev team but it is not difficult to do and once implemented will give you a much clearer picture of how your site is performing. The necessary code with a good illustration of how the data will appear in your GA account can be found here. A more comprehensive and elegant solution is available from Brian Clifton a UK based GA guru. His code allows you to define a bounce as anyone who stays less than 10/20/30 seconds (you decide) on a page. There is an added side benefit that it provides easy tracking for all events even on very large sites. With it you will see in GA the number of PDF, Word, etc downloads and the number who clicked on outbound links such as `mail to` or to an external site (very useful for showing commercial partners how many visitors you are sending to their site). Installing this kind of code on your pages could give you a very different view on how successful your site is.  And of course help justify the costs involved in creating and maintaining it.

I recently entered searches in Google such as `affected xyz`,` causes of xyz`,` treatment of xyz`, where `xyz` was replaced by a range of well know conditions. The top 45/ results that I got were often not the top UK charities in the field but what are known as Content Farms. These are sites that give answers and information but make money from the ads that appear beside the information. While there is nothing wrong in this practice it does mean that people who are probably in the early stage of searching about a condition or illness will not find you and may well end up on a competitor site. You have lost the opportunity to build a long term relationship with that person so you will have to incur the cost of advertising to reach them when a little upfront work could have made a page from your site the top result for many quite common queries.

These `Content Farms` have become a major irritant on the web. They find search queries, such as ` causes of xyz`, that have 1. plenty of search traffic and 2. where the resulting displayed web pages are easy to push off the top position in the search results. The information these sites provide is usually very basic but is sufficient to keep visitors long enough on their pages to see and use the ads. Google has now attempted to clamp down on them. Both the Content Farms themselves and Google`s attempt to stamp on them could have serious consequences for your web site traffic.

First, as we have seen above you could already be losing traffic for search queries that are popular but second you could also be (accidentally) penalised by Google as it attempts to reduce the frequency with which these sites appear in the search results.

To deal with the first problem you need to find out what some of your important but less popular search queries are. To do that you need to go to your Webmaster Tools account ( sign up and get data for your site. It is completely free). Now check under `Your site on the web` > `Search Queries`. Here Google shows you all the queries that people use to find your pages (wow – did you know that so many people used *that* search term?) If you see a search query that you think is important, see what your top page for that term is. Now you need to move it up the Google results until it is at or near the top. The quickest way to do this is to put the search query or a variation of it in the Page Title and for good measure add an internal link ie a link from somewhere else on your site, back to this page with the search query as the anchor text. Over the following weeks you should see a rise in visitors who have searched using that keyword/phrase as their search query.

The second problem is becoming more important. The reason for this is that Google is now starting to penalise those sites that have `bad housekeeping` For example sites with many low quality pages, poor quality content and poor user experience. In my experience most UK Charity sites have many, many more pages listed in Google than they really have on their site. I have seen sites with 30/40/50,000 pages indexed in Google where the number of real pages is in 100`s or 1,000`s.  You can check this for your site very simply by typing into the Google Search Box: `site:www.mysite.co.uk` (Note there are no spaces) You can narrow down problem areas by checking a section of your site eg `site:mysite.co.uk/folder`

Google Site Operator

Causes of this `page bloat` can range from Calendar’s that create a separate page for every day for 50 years(!), separate pages for every single log-in and log-out of your site that users do, the use of search query parameters and many others. You can identify the pages causing the problem by going to Content > Top Content in your Google Analytics account and reviewing the url`s of pages with just 1 or two view`s.

Generally Google`s algorithms recognise the quality of UK charity sites and indeed seems to treat them leniently with regard to these kind of problems. However that will not last forever so be warned.

If you consider your web site as a supermarket would you prefer to `shop` in a supermarket where the aisles were badly signposted, goods were available in 10 different places, and the shop was ten times too big or would you go to a supermarket where you could find what you wanted easily and quickly ?

Yup. Me too.   Google searchers will increasingly do the same.