Why Do I Have (not provided) in my Analytics Keyword Reports?

You have probably noticed that when you view reports in Google Analytics the term (not provided) often appears near the top of the keywords driving traffic to the site. For many of the larger sites I've looked at, it is often the single highest keyword term.

Why does the keyword report show (not provided)?

This (not provided) keyword will show on any report that includes information for Google natural search visitors.  This is down to a change rolled out in November 2011 by Google. They decided that they would no longer pass on the search term used by anyone logged into any Google account when they made the search.

Unfortunately, due to Google's strength in email (with Gmail) social media (with Google+, YouTube) and other tools, such as Adwords, Analytics and Apps, more and more site visitors complete searches while logged into a Google account. This means that a substantial proportion of your site traffic will have their keywords hidden.

Why did Google hide the keyword search data?

Google stated in their official announcement of the change that the main reason was to make search more secure for their users.There has been a lot of debate as to whether this is actually the true reason but the effects are the same - for a substantial proportion of your data you cannot see what search terms they are using.
Unfortunately, this may have started a trend, as Firefox have recently stated that they are planning to make all Google searches secure by default, which will have the same effect as being logged in.
There is a possibility for websites to use searches from logged in users to identify the specific visitor, something Google Analytics has always prevented you from doing. This change will reduce that, but at the cost of losing critical information on a substantial proportion of your traffic.

So what's the big deal about losing keyword data?

Active website managers are always looking at the keyword data and the pages it is landing on to ensure they are providing the best service for their visitors. If you see a substantial growth in a particular keyword can influence re-writes and new pages being added.
For instance, on this blog I've written a beginners guide to installing Google Analytics.  If I was looking at the search traffic for this page, and noted a substatial number of searches for 'How do I test My Google Analytics installation?, I'd probably write a new post specifically on that topic and link to it from the original. Without this information, I would never know I am getting traffic that is not finding what it is looking for.
Unfortunately, for many on the sites I've reviewed since the change, traffic under (not provided) has higher bounce rates and lower conversion than the site average, but there is no direct way to improve the experience for those visitors.

Is there anything I can do to get round (not provided)?

There is no way of getting that information back Google is blocking it at the source, so even switching to a different analytics package will not get the information back. However, there are some techniques to reduce the impact.

  • Use Adwords traffic data. Traffic from visitors who click on a paid ad still provide the keyword data. This has led to critisism about  double standard, but at least you can use this data to analytse keyword bounce and conversion rates if you are running and adwords campaign.
  • Use segments to hide this data. You can create a custom Advanced Segment to hide the traffic that comes in with a Seatch matching exactly '(not provided)'. This will at least allow better analysis of the information you do have, and improvements to poorly perfoming terms you are aware off will improve the experience for the visitors where you cannot see the terms.
  • Ask visitors directly. There is nothing to stop you triggering an on-site survey to get direct feedback. You may even be able to ask only those where the keywor data is blocked.
  • Use on-site search data. By logging the usage of your on-site search box - the landing pages, keywords used and the outcomes - you may be able to determine what searches are not providing the visitors with what they want on the landing page.
  • Ask Google to change their policy. While there are some good reasons for hiding this data, Google may do more harm than good in the long terms as sites become LESS responsive to visitors needs - the opposite of what all other advice about quality web page should be. If you have any influence with the Google Search team you could ask them to start passing the keyword data gain - but I'm not being optimistic!

Enhanced Keyword Research Tool Help You Target the Best Search Terms

When you are looking at the performance of an existing site, and then deciding where to put your SEO and PPC efforts, one of the first places I tend to go is the Google Adwords Keyword Tool.

It is available with or without a Google account, but you get a bit more data if you are signed in.

The premise is fairly simple. You give it a keyword or an existing site to look at, and it will show the related search terms that people are using, and the volume, as well as the competition. This is hugely important, as it can save a lot of wasted effort creating content, SEO campaigns or Adwords campaigns for terms that no-one is searching for. The competition value gives an indication of how hard it will be to get a first page ranking for that term.

The latest update is a great little improvement. If you enter an existing website address in the box, you can see how that website is performing in terms of the number of times it appears on the first page of search.

The new tool seems to be in beta at the moment as I am only seeing data intermittently.

How Do I Use this New Keyword Tool Data?

This is going to be primarily useful for sites that already have quite good visibility on Google. Use the tool to work out your approximate visibility for the searches for each of the terms you are targeting. More than 80% is going to be a good rate for first page appearances, due to differences in location and personal based targeting.

You can also use your Analytics data to determine natural search or Adwords click through rates for those exact terms, but you will have to do this manually.  Typical click rates for top of natural search are 2% to 4%, but there is a lot of variations, especially for brand name searches.

Now you have identified the key terms where you are getting lower visibility or click through rate, you can look at the page and work on improving both visibility and clicks. First, have a look at where you are ranking for this term. If you are showing below first place position, there is obviously an opportunity to get more traffic by increasing your search position for that search term, so use your favourite SEO techniques to boost that search engine rank.

You may decide that you want a new page to more specifically target this term, which could be a new blog post, product or service information page, which you can then promote both from your existing pages and off-site.

Lastly, look at the landing page itself. First port of call is the 'snippet' that appears beneath a search result. Make sure this is as enticing as possible for the searchers you are targeting. This is usually controlled by the Meta Keywords attribute on the page. A Good snippet will boost the click through rate no matter what the position you appear in.

Google Consumer Surveys - Get Your Questions Answered

The latest development out of the Google ideas factory is a bit of a surprise - I had not even heard rumours about this one!

They have developed a system similar to Adwords, but where sites can be paid for running single question consumer surveys on their sites. Sites can use this system as a form of 'paywall', only allowing access to premium content after you have completed a survey. It's a great way of monetising traffic to a site as it is likely to have a much higher payment conversion rate than standard on screen ads.

For brands or any company looking to gain valuable intelligence, this offers a really cheap way to get valuable consumer insight. It looks like you are limited to a single question per visitor, and from the layout it will have to be fairly simple, although pictures are allowed.

The big advantage is that it will allow you to ask questions of a wide cross-section of the population, and looks like it will be easier to set-up and administer than alternatives like paper or email surveys.

You can see more information on the Official Google Consumer Surveys guide.

Sample Survey Question

What is 'Bounce Rate' in Google Analytics

Bounce rate is a term you see a lot in Google Analytics reports. It's attached to visitors, sources, landing pages as well as the overall site summaries.

Bounce rate is one of the metrics that does cause confusion. What exactly does it mean? Is having a high bounce rate bad, and a low bounce rate good? What should my bounce rate be?

Bounce rate is a measure of a visit that only looked at a single page on the site.

All the bounce rate can tell you is that they only loaded a single page. It is used as a measure of engagement - how interested the visitors was in the page - but has major limitations due to the way it is calculated.

So is a high bounce rate bad?

The answer is...usually. We know that a bounce rate is someone leaving the site, but on it's own it says nothing of the users intent.  Here's two examples that would show as a 'bounce'.

Visitor one searches for 'blue widgets', and lands on the blue widgets product page, and then leaves. This is a negative bounce - assuming this is an e-commerce site, you want them to add a product to the basket and complete a purchase. A high bounce rate is bad.

Visitor two searches for 'the widget company phone number' and lands on the 'Contact Us' page, and then leaves. In this case, we assume they probably made a call. The visitor found exactly what they were looking for, so there was no need to search around the site - a high bounce rate is good!

E-commerce sites are generally simpler. For visitors looking for products, bounce rates should be low, and for visitors searching for help or customer services, it should be high.  Lead generation sites, on the other hand, may have a harder time determining what is a good bounce rate, so will have to segment their visitors further.

What can I do to get accurate bounce rates?

There are some changes you can make to ensure bounce rates are as accurate as possible.

  • Create thank-you pages for online submission forms, web chat or similar functions. As another page is loaded after they have completed them, it will not show as a bounce.
  • Consider using phone tracking software that simulates a page load when they call, such as Splice Insight, as visitors who make a phone call instead of taking another action will not show as a bounce.
  • You can add a dummy page load in Analytics after a user clicks on a mailto: email link
  • Consider using events and goals to track users who stay on a page for more than a certain time as an alternative to bounce rate for measuring engagement.

What if I use a single page web site?

Your bounce rate is going to be 100%, unless you have a 'thank you' after completing a web form. Instead of bounce rate, you will have to look at other ways of measuring engagement. Google Analytics will not give accurate measures of time on page, but some other Analytics packages do.

What should my bounce rate be?

This is really open to interpretation, but from my experience a 'good' bounce rate would be:

  • Below 30% for an e-commerce site
  • 30 to 40% for a lead generation site, but lower if you are tracking phone calls
  • Up to 50% for a blog site, but this will vary depending on the nature of the blog. Blogs with tightly focused topics will tend to get lower bounce rate, those with very disparate themes will be higher.
Above these numbers, the page is generally not providing the visitor with what they want. Have a look at the page content, and how it compares to what you know of the visitor. Are the terms most visitors search for strongly highlighted on the page using headings, bold text or bullet points? Is the next step the visitors has to take clear enough? Does the page generate trust by looking professional, especially for an e-commerce site? Are you using split testing or other methods to test different page looks and content to identify the most effective versions?

Any comments on this? What is your bounce rate - and what steps have you taken to improve it?

Automatically Email Google Analytics Reports

Just added last week was the ability to schedule a Google Analytics report  to run and then be emailed either to yourself or anyone else. It's a great method for checking critical figures, as well as reporting to anyone else involved in the site, like your boss or marketing department.

First, choose your standard or custom report from the reports list, and set an additional options, such as segmentation, previous date compare, the metric to display in the graph and the graph time period.

The reporting has not been around long enough to test ho the data selection works, but I would suggest leaving the data range on Yesterday, Last Week or Last Month to ensure you always get a useful date range.

Once you are fully happy with the report, open the report emailer by clicking on the 'Email' button just beneath the report name - it's currently labelled 'Beta'.

You will be presented with a new pop-up window with a bunch of options:

  • From address is fixed to your login an cannot be changed, except by changing your main Google account.
  • To is one or more email addresses that this report is to be sent to. Don't forget to include yourself if it is for multiple recipients. You need to separate the email addresses with a comma.
  • Subject will be pre-filled based on the report, but depending on the number of reports you configure you may want to change this to be more descriptive.
  • Attachments is the type of file to be sent. If you want to see the report in all it's graphical glory, select PDF. CSV creates a comma separated text file suitable for importing into a spreadsheet, database or other analysis system. Although they are just text, CSV files are not really suitable for reading in a text editor.
    TSV is similar to CSV, but the fields are separated with tabs instead of commas. This is mainly used by newer versions of Microsoft Excel, although most other spreadsheet and database systems can handle the files. The option is TSV for Excel makes a few minor changes to make the file look better in Excel.
  • Frequency sets how often to trigger the automatic email. Set to once will turn off all automation, and just send the email one. Daily will send every day, weekly will send once per week and prompt you for the day to send it. The most common option will be Monday where the date range is set to 'Previous Week', but it depends on your needs.
    Monthly triggers once per month, and allows you to set the day of the month. Be careful using and day of month above 28th - it's not yet clear how this will work for shorter months.
    Quarterly emails once per quarter - there are no options for this, as the quarters are fixed.
  • Advanced Options only appears where the report is schedules (not for 'Once') and allows you to set how long the report will be repeated for. The default is 6 months.
  • The white space if for the email message. You must complete at least some text in the message box before you hit send.

If you have more than one report that is triggered on the same schedule, you can add them together as attachments to a single email for convenience. Instead of hitting the send button, click on the 'Add to Existing mail' link in the bottom right. If you cannot see it, you probably have not yet created any automated reports.

If you select this a new page will open with the various frequencies listed, select one, and tick the box on the email you want to add this new report to, and click save. I quite like using this for two purposes: grouping together various different weekly reports, and for adding a CSV file to a PDF report so you have both the pretty graphs and the raw data.

One thing missing on the current Beta version is the ability to easily edit the scheduled reports for timing or recipients. Once this function is added (or I can find where it is hidden!) I'll update this post or write a new one.

There is more help on Google Analytics reporting at the official Google Reporting Help section.

Google Analytics (Finally) Gets PDF Export, Auto Emails

It's a small upgrade, but anyone who has to compile reports for otehrs from Google Analytics will be delighted that they have now added a PDF export functionality to the new version of Analytics.

Previously, if we wanted to get the pretty graphs, a screenshot was about the only way to go, but you can now create a nice PDF of whatever report you are looking at, including custom reports.

To get a PDF of the current report, it's as simple as selecting the PDF from the menu:

Even better, you can now also use the email button next to the report to email the report in PDF to recipients of your choice. Set this up for weekly or monthly reporting and you can cut down a lot of manual copying and pasting.

Next post we'll look at the automatic report emailing options in more detail.

Critical Google Ranking Change Leaked

At a panel style question and answer discussion a few days ago Matt Cutts - regarded as the public face of Google Search - let slip that Google are working on an update that may have a significant effect on thier search rankings.

In response to a question on how Google handles searches where there are some very well optimised pages that are outranking everything else for a particular search, Mr Cutts said:

"Most search engines have made a lot of progress in being able to crawl through all that richer content... what about all the people optimizing really hard and doing a lot of SEO? Normally we don’t normally pre-announce changes but there is something we have been working in the last few months and hopefully in the next couple of months or so we hope to release it. The idea is we are trying to level the playing field a bit. All those people doing, for lack of a better word, over optimization or overly doing their SEO, compared to those making great content and great site.

We are trying to make GoogleBot smarter, make our relevance better, and we are also looking for those who abuse it, like too many keywords on a page, or exchange way too many links or go well beyond what you normally expect. We have several engineers on my team working on this right now."

So what does this mean for your web site?

The good news is that if you have been doing what Google have always asked from the start - writing for people, not search engines - you may find your rankings improving as other pages lose position.

However, if you do have a lot of pages that are 'over optimised' you may lose some rankings. Since Mr Cutts did not give any details, we can only base it on earlier comments about what search engines look for.
Write unique content that is written for human readers
Do not repeat the keyword excessively in the page content. You would expect it in the headers and a few times in the content.
Look at the links pointing to the page - if all the links are using the same single keyword or key phrase, this may look suspect. Natural links tend to have varied link text.

Of course, we will only really see the effects once the changes are made live, and Google never reveals it's complete ranking algorithm, so actively monitoring your site will be vital.

There is some more information and a copy of the audio at SearchEngineLand.

How Do I Install Google Analytics?

Google Analytics is pretty easy to install. You will need to have some way to edit either the entire code for the page or to edit your main site template files.

If you have FTP access to the site files, you can achieve this with a simple text editor. If you are using a template driven site builder, you will need to find the main page templates, and switch to HTML view to edit them.

Once you have access, you can begin. Go to http://www.google.com/analytics/. You will need to sign in with an existing Google account, or create one.If you choose to create a new account, it can be for your existing email address or you can even create a new Gmail address as well.

Once you have the account created, you have to add the first Analytics account. The process is fairly straight forward. If you are not already in the New Account screen, click the new account button now.

Top Tip: Always create a new website as a new Account, NOT a new Property inside another account! If you wish to give another person access to your full Analytics account in the future, they will have access to all Properties inside the account. By splitting them up you can make life easier in the future.

You will need to give it the Account name. I usually use the URL, but it's for your display only, so ffel free to shorten it. You will also need to give the URL (full website address), the time zone and the country. At the moment. you selects whether the site uses http://, https:// or something else in the first box, and the rest of the URL in the second. If you site uses www. make sure you include it.

Next you will be presented with a page with the tracking code. Leave the selection at Standard and Single Domain. These will be fine unless you have more complex needs outside the scope of this blog.  Copy the tracking code - select all of the text in the box and do right-click>Copy or CTRL-C on your keyboard.

Next, on your website editor, you need to locate the tag that says </head>. It is usually near the top, and there should only be one in the page or document. You can often use Find (CTRL-F) to locate it. Go to the line above this, enter a new line and paste the code in.

Save the web page or template, and upload it by FTP if you need to.

If you have multiple pages, or more than one template, repeat the process on every page or template. It must be on all the pages on your website for it to work properly.

Top Tip: Adding the tracking code can be easily missed on some pages. Check especially for 'Thank You' pages that load after a form is submitted, 'Static' or 'Fixed' pages for blogs that often use a different template from the main blog posts and the home page, which again can sometimes use a different template.

Some website and blog creation systems have Google Analytics built in, and all they need is the reference number. On you Analytics account, have a look at the top of the page. It should say:

Tracking ID: UA-30109067-1

That code, starting with UA is the tracking code. Paste it into the correct space and save - you generally don't need to do anything else.

Testing Your Google Analytics Tacking Code

Now it is installed, you need to test it. The fastest way is the new Real Time Tracking. It is currently in Beta (testing). Click the Home button on your Analytics account, then select 'Real Time' on the far left menu and select 'Overview' once the menu opens.In a separate browser window, go visit your new site, and you should see one visitor appear under Right Now in the main page section. Success!

Problems? If you do not see your visit after a few seconds, check that the newly added tracking code is appearing on the page. View the page source in your browser (generally right-click on a blank bit and then 'view source'). Look just before the </head> tag. If it is not there, try refreshing the page, and if it is still not there go back to your editor and make sure you saved it.

Lastly, it's a good idea to check all the pages have the tracking code added. Use either the Real Time viewer or the 'view source' method to check all the major pages or page types. This can be list views, individual item views, forms/contact pages, the home page. If the code is missing from one page it can create some very odd numbers later in Analytics, so it is a good idea to check now.

That's it - well done. Give yourself a pat on the back.  Come back in a few days to see how the site is performing. Now all you have to do is get people on it.

Top Tip: The main Analytics reports are NOT real time, and can take several hours before data from site visits is available. For safety, generally only look at data up to the day before when reviewing your site.

That's all for this one. If you have any specific questions about installing Analytics, please ask in the comments below.

What is Google Analytics and Why do I Need It?

Google Analytics is a system for measuring activity on your web site, blog or page. It records information about people who view the site, giving indications on what they liked and how they got there.

While Google's product is only one of a number of web analytics packages, it has one huge advantage: it's free. There are plans to start charging the very largest sites, but for the vast majority it remains completely free to use all the functions.

So what is it good for?

  1. Basic tracking of visitors, so you can tell how well many people see your site. This is vital for most sites - both commercial and non-commercial need to know they are being read, but for very different reasons.
  2. To determine where visitors are coming from. Most sites get visitors from a variety of sources - search engines, social media, other web sites - and it's really useful to see which is providing the most traffic. This is especially important if you are paying for links to your site, as you can then establish roughly how much each visitor costs.
  3. To find problems with your site or pages. Analytics tracks a huge number of variables, and these can be useful for tracking down site problems. One report I use regularly is the summary by browser type. If I see that visitors using a particular browser always leave after one page, it's a warning that something is wrong and needs tested. It could be the page layout is broken, or my menu is not working. This is especially important for mobile browsers such as iPads and iPhones which have some limitations as to the types of content they can display.
    The latest version on Analytics also allows you to check page load speeds. A fast-loading page gives a better experience for your visitors, and you can often improve the experience by taking simple steps like reducing the size of images or adding videos in a different way.
  4. To perform calculations on your return on investment. Even if your site is not commercial, its useful to know what your efforts achieve. Analytics can track views of specific pages, completed webforms and even (with some outside help) phone calls. This allows you to determine which visitor sources, topic or pages provide the most return for your time marketing funds.

There are loads of other uses for the data that Google Analytics can provide. Our next post will be on how to get Analytics added to your website.