Google Analytics Data Accuracy

Like many other web analytics tools, Google Analytics does a good job of outlining trends and helping you to understand how your customers view and navigate through your website. However, like any web analytics tool, reports will never be 100% accurate due to a number of factors: visitors may block cookies; browsers may bounce before code fires and so on and so forth.

Does that matter?

While there should be a reasonable confidence level that everyone is comfortable with, it’s important to remember that analytics should be about identifying trends and providing insight into how to optimise your business and provide a better experience for your users, which in the long run will translate into better profitability.

Knowing that 20% more visitors converted following a marketing campaign is more powerful than knowing that exactly 12 people visited a specific page on your website today. And, in most cases, what web analytics managers need to know is not an exact figure. It’s more about the change from one time period to the next.


Even if the figures shouldn’t be regarded as precise, Google Analytics is still a great way of tracking changes. A fraction of visitors may be invisible or slightly miss-identified, but as long as those fractions and the query you are using on the data stay consistent over time, you’ll get good-quality information about what’s working well on your site and what’s not.



Leave a comment

Filed under Uncategorized

What is the best way to track mobile sites in Google Analytics?

Now that you’ve created your companies mobile website it’s time to think about tracking results. This will help you analyse the effectiveness of your content and ultimately help you optimise your sales.

If you are using Google Analytics on your regular website already, it’s simple to add the same type of tracking to your mobile website.

Once you have installed all the tracking tags, it’s important to set up a separate ‘mobile’ filter within your current Google Analytics account, so that you can differentiate between mobile and main website traffic.

This will allow you to track the performance of your mobile website separately from your regular website and optimise accordingly.

Setting up a separate profile for mobile traffic

At the moment there is not ‘one size fits all’ filter you can use. Some Google Analytics consultants recommend that you use a “include” filter on “visitor operating platform” using the following regex:

 iPhone|iPod|PalmOS|Playstation Portable|Playstation 3|Nintendo Wii|SymbianOS|Danger Hiptop

Others suggest creating a custom filter which singles out visitors by using the height and width of the browser screen as a reference. The following filter matches any traffic coming from sources with a resolution of 320×480 or less:


While these filters might work for the majority of users, the difficulty I’ve had however is that our client’s mobile website is for high end devices only. For example, I cannot visit the site via my BlackBerry. This makes both of the above filters pretty useless.

A solution

To overcome the limitations of the above filters, we implemented a “_setVar” cookie within the mobile website page tags, which allowed us to identify visits to the mobile site.  The page code looks something like the following:

 var _gaq = _gaq || [];

_gaq.push([‘_setAccount’, ‘UA-XXXXXXXXX-1’]);

_gaq.push([‘_setDomainName’, ‘’]);

_gaq.push([‘_setVar’, ‘mobile’]);



To separate out our mobile and main website traffic we used the following custom filters. We applied the first filter to our regular website profile to exclude visits to our mobile website.

 Name: Exclude visits from our mobile website

Filter type: Custom filter > Exclude

Filter field: User defined

Filed pattern: Mobile

Case sensitive: No

We then created a separate filter specifically for our mobile website.

Name: Mobile website only

Filter type: Custom filter > Include

Filter field: User defined

Filed pattern: Mobile

Case sensitive: No

Now you can use all the same, great actionable insight you get from your regular website Google Analytics profile for your mobile website.

Leave a comment

Filed under Uncategorized

Monetising your Web Analytics Recommendations

Imagine the situation….. you’ve done your weekly reporting, you’ve got your actionable insight and you’ve made some solid, prioritised recommendations – all of which form the basis of a sound web analytics framework. However getting the commitment, budget and resource to implement them is a bit more of a struggle than you anticipated… Sound familiar?

Sometimes you might find that your recommendations are not met with as much enthusiasm as you would have hoped – more often than not, clients and key stakeholders are typically overwhelmed with projects and have limited resources to complete them, meaning that more often than not your recommendations can be put to the bottom of the “to do list.”

So how do you get the ball rolling?

For many businesses, the only “action” that matters is driving revenue into the company and profit to the bottom line. So without sounding to cliché – all you really need to do is “show them the money!!!!!”

Giving weight to your recommendations  

The process of assigning values to site behaviours is known as monetisation, and from experience, it’s one of the most important tools a web analyst has. While it might sound quite complicated, it’s really quite easy to understand at a basic level and it can be very powerful in motivating businesses to embrace the idea of identifying opportunities and improving site performance.

An example

Once of our clients has a really pants checkout process. While their conversion to basket is healthy, around 25%, drop-out from their checkout is currently running at 77%. In monetary terms they loose on average £60k of revenue each week!

How did we make them sit-up and take notice?

By producing a forecast which looked at how much incremental revenue they could expect each year by reducing the drop-out rate (click for a better image).

 web analytics forecasting the impact of your recommendations

 Such charts are especially useful when trying to convince clients and other stakeholders to act on the opportunity. It’s also highly useful for justifying the overheads of developing new designs and running split tests – as you can use them to look at payback and ROI.

It’s important to reinforce the fact that the client / company may not see these exact improvements. This is a method to estimate the potential impact of different initiatives and forecast their financial impact in terms that are easy to understand.

As always, if you have any questions about the above post I’d be glad to answer a few so please use the contact details at the top of the blog or email me at

Leave a comment

Filed under Uncategorized

How to use Google Analytics “Assisted Conversions” Report

 The traditional ‘Ecommerce’ report within the ‘Traffic’ report utilises the “last-click wins” method of attribution. It’s simply a measurement model giving “credit” to the last keyword and/or referrer to your site before a conversion occurs. In other words, no matter how many different channels / media a consumer has previously been exposed to, the last click before purchase is attributed 100% of the revenue generated.

However, it’s pretty rare that people buy in one step, so I am sure you can understand 1) why this model is completely flawed and 2) that some media buys can appear to be very poor performers on a last-click basis (such as display and early funnel keywords) but are in fact very effective for creating demand that is subsequently satisfied through other channels.

How Can The “Assisted Conversions” Report Help?

If campaigns were optimised solely with the last click before purchase rule, there is a danger of under optimisation of display buys and early funnel keywords. For example, on the report below, my Generic Keywords are only showing 62 “last-click” sales (remeber to click for a larger image!).

Generic Keywords Assisted Conversion Report

However, when I switch to the “assisted conversion” report I can see that my Generic Keywords actually “assisted” in generating over 329 sales and £300k worth of revenue.

Based on the “last-click” wins model, it looks like my keywords are underperforming and I might have made the decision to disable those underperforming terms (and/or ads) and focus on what’s working – eeek what a mistake that would have been (as I would have lost a load of sales down the line).

How can you use this information?

If you could go beyond typical conversion data and see how often a campaign or specific keyword provided an “assist” prior to a conversion—you could make the smartest decisions about which keywords are driving return on ad spend, and adjust your campaigns accordingly.

For example, rather than calculating Return On Advertising Spend (ROAS) on a “last-click basis”, you could look to re-calculate this using the ‘assisted conversion value’ metric to create a sort of Assisted Value to Spend Ratio.

Worth mentioning

Keep in mind that “assisted conversions” only have a 30 day window, which means as long as your visitor converts within 30 days of their first visit, Google Analytics will stitch together all the different traffic sources that they clicked on prior to conversion.

Also – Multi-Channel Funnels data collection lags by up to two days, so the conversion count in Multi-Channel Funnels will not match what is shown in other Google Analytics reports.

Leave a comment

Filed under Uncategorized

What’s the difference between “Visits” and “Unique” pageviews?

A common question among some of our clients is whether they should be using the number of “Visits” to a page or the number of “Unique Pageviews” to assess your pages performance.

From my understanding “Visits” are not valid when we talk about pages-level analysis. That is, when you look at the “Visits” metric for a page, what you are actually seeing is the count of “Entrances”, because Google Analytics logs a single “Visit” count on the first page of the visitor’s session.

Each subsequent page that the Visitor goes onto view gets a “Pageview” count. A “Pageview” is counted every time someone hits on the page, so if the same user hits a page 10 times, it’ll be tracked as 10 “Pageviews”.

On the other hand, “Unique Pageview” would record the instance as “1 Unique Pageview” because it is all from the same user.

Therefore, when analysing data from the Content report, it’s best to use “Unique Pageviews” as a measure of the number of “Visits” to a page.

As always, if you have any questions about the above post I’d be glad to answer a few so please use the contact details at the top of the blog or email me at

Leave a comment

Filed under Uncategorized

Google Analytics Merchandising Report Part 2

If you are like many who sell products or services online, you’ll probabally be interested in knowing which ones on your site are converting better than others.  In my previous post I showed you how to tag your website so that you can track the conversion of your product through various different stages:

  • Visits => Product
  • Product => Basket
  • Basket =>  Checkout
  • Checkout => Order

In this post I am going to show you how to add extra insight to the merchandising report. The information from the report can be used for merchandising, designing store layouts and other marketing efforts around products.

Getting started with the data

Before you get started you need to export the data from your Google Analytics account to Excel. As Google Analytics only allows you to export 500 rows at one time you need to add a hack to the URL. To do this, simply add ‘&limit=1000’ query parameter to the end of the URL for the report you want to export. Now export the data using the CSV file and this will increase the number of rows returned on export. Easy!!

Creating your Excel document

I generally export all the data from the ‘views’, ‘basket’ and ‘sale’ categories. Once I have done this I add all the data into different tabs within the same spreadsheet and rename the tabs accordingly.

On the view tab, I then create a new column called ‘vlookup’ and using the vlookup function within Excel find those products which have been viewed and also added to a basket.

I repeat the process to find those products which has also resulted in a sale. The end result should look something like the following (click on report for a larger image).

Product tracking example

From here, you can then calculate ‘Product Success Rate’. This is defined as the number of visits in which that product was added to basket divided by the number of visits in which that product was viewed.

You can also look at ‘Sale Conversion’ – the number of product sales divided by the number of visits in which that product was viewed. I’ve provided an example below (click on report for a larger image).

product page sucess rate

So what to look out for within this merchandising report? 

Key products to identify are both popular products (high visits) with a low ‘Product Page Success Rate’ and less popular products (low visits) with a high Product Page Success Rate. For a recent client report, I broke products down into four categories:

  • High views high sale conversion: the most popular products based on the number times the product was viewed and the number of units sold.
  • High views and basket conversion only: for some reason, people just weren’t convinced on these products – they were viewing it, adding to basket but only a low proportion ended up buying it.
  • High views no conversion: popular products that aren’t being added to basket. These products need to be reviewed urgently.
  • Low views high conversion: products that not many people are finding but they commonly add to basket when they do view it.

Gaining insight and recommendations

The whole point of creating such a report is so you can feedback to your merchandisers, buyers and website managers about what is and isn’t working. Insights you may make on the back of this analysis are:

  • High views high sale conversion: given the high conversion of these products, you could look to  you increase the visibility of these products by deep linking to them from the homepage, prioritising them in the search results listings, increasing their paid search budget etc.
  • High views and basket conversion only: there is a need to reduce buying friction on these pages, therefore you could consider adding live chat, customer reviews, creating an attractive returns policy etc.
  • High views no conversion: there could be some quick changes you can make on these pages which leads to incremental sales for these products. For example, you could re-look at your pricing strategy, enhance your image quality, add image variants etc.
  • Low views high conversion: it might be these products only appeal to a small niche market but for some, if they were exposed to a wider market via promotion on the site or a paid search ad, their sales could be significantly boosted.


Now you can begin to see some information that is actionable (that is some products are converting better than others) and start addressing these new key findings. As always, if you have any questions about the above post I’d be glad to answer a few so please use the contact details at the top of the blog or email me at

1 Comment

Filed under Uncategorized

Google Analytics Merchandising Report Part 1

To create an effective merchandising report you need to look beyond looking at what products customers actually purchased, and start collecting information on how well your individual products are converting, based on: how many times individual products are viewed, added to a shopping cart and ultimately purchased.

Out of the box Google Analytics doesn’t have any Merchandising Reports, however, you can use the Event Tracking feature in Google Analytics in order to generate this insight.

The report allows you to answer – which products are visitors most interested in – what do they think is hot right now – which products do visitors need some extra reassurance for before they commit to purchase?

The tracking tags

In order to get these answers you need to add some custom code to your existing Google Analytics script.

Product page code

Use the following code on each of your product pages. The ‘product_name’ element of the tag should be inserted dynamically using your Content Management Solution (CMS).

_gaq.push([‘_trackEvent’, ‘product’, ‘view’, ‘product_name’]);

Basket page code

Add the following line(s) on your basket page – you’ll need to make this call for each product in the basket. For example, if there are two items in the basket, then this line should be called twice – one for each item.

_gaq.push([‘_trackEvent’, ‘product’, ‘cart’, ‘product_name’]);

Checkout page code

Add the following code to the first page of your checkout process. Again, you’ll need to make this call for each product in the cart.

_gaq.push([‘_trackEvent’, ‘product’, ‘checkout’, ‘product_name’]);

Confirmation page code

Finally, on the order confirmation page, you should add the following code. Once again, you’ll need to make this call for each product in the basket.

_gaq.push([‘_trackEvent’, ‘product’, ‘order, ‘product_name’]);

Where does the code sit?

The code needs to sit in the main Google code before the ‘_trackPageview’ element, as shown below.

script type=”text/javascript”>var _gaq = _gaq || [];
_gaq.push([‘_setAccount’, ‘UA-XXXXXXXXX-1’]);
_gaq.push([‘_setDomainName’, ‘’]);
(function() {

Viewing Reports

The reports will be available within the Event Tracking section inside Google Analytics. If you want to see the overall progress at different stages, you can start with the “Categories” report shown below.

product tracking google analytics

From there, you can drill down into each of the different elements. For example, when I click on ‘View’ this drills down into all the products that visitors have viewed within the specified time period.

product tracking google analytics

If you want a different view, you can go o to the “Labels” report which provides a list of individual products. Select a specific item within the report and see how many times an individual item was viewed, added to cart, checked out and purchased.

Making sence of the reports

Look out for part two of my post where I show you how you can use this data to create a merchandising report within Excel.

Quick note

One thing to mention isthat Event Tracking generates extra views in your account – which impacts your pageviews and bounce rate data.  For example, if a visitor hits a product page and bounces, because you’re using Event Tracking to track the page view event, you won’t be able to see the bounce event take place.


Filed under Uncategorized