Bits and bobs and Google Analytics

A few of us in the Production team recently shuffled our way through a Google Analytics short course. It was a great way to brush up on our knowledge as it had been a couple of years since I'd needed to dive into analytics, and a lot has changed in that time. Here are some useful snippets of knowledge that I thought were worth mentioning. Gone are the days where Google Analytics was just about hits and page loads, now I can't imagine a successful web project without it!

Predefined Views

This was an interesting session as we explored some very easy ways to use views and filters to make the statistical data more acurate. What you're actually doing is limiting or modifying the traffic that is included in your reports.

For example, did you know that you can setup a view for excluding visitors from a particular IP Address? This would be a great idea for any site, but especially if you have a large client (like me) where multiple staff daily are visiting the website, as well as a dedicated development team who are always on the site working, checking and testing things. If we wanted to we could add the Wiliam office IP address to the exclusion list, as well as the client's many offices around the world, then we would be presented with a more meaningful set of reports on visitation.

For a large client, exclusions could be a drop in the ocean, even if it is a significant amount. But this could be equally important for a startup or a low traffic site - you don't want to be skewing your statistics unintentionally. (Oh look, we had 200 visitors today! A pity that 180 of them were boss checking the work every 5 minutes).

There are other pre-defined views than just IP exclusions. Just as useful could be an inclusion - create a report that ONLY shows you traffic from one particular source. You can also exclude/include traffic to particular URL (e.g. /blog or perhaps /help/faq) or even to a particular hostname (blog.wiliam.com.au).

Importantly, as I discovered, when you setup a new view, you do't get to keep historic data so your stats for this particular view will start from the point you create the view. Not ideal but that's how it is, you can still use the default reports and then benefit from the more segmented views as the months and years roll by.

If you want to read up on Google Views a bit further, there's a page dedicated to this over at Google.

Annotate your timeline

Did you know that you can leave comments on your timeline? Quite a useful tool, which means next time you send out an EDM campaign, you can explain to everybody viewing your analytics why there is a huge spike in traffic on a particular day. Or was it scheduled maintenance that was responsible for a drop in site visits between 6am - 7am on a particular date? Annotate your timeline so that next time somebody logs in and generates a report within that timeframe, they can be more informed about anomalies which are easily explainable but not always obvious.

This is particularly useful if you are trying to compare previous periods. Generally speaking, traffic should trend upwards year after year - unless you're not doing it right! - so annotations can help explain improvements and declines when you are viewing one or more periods together. Re-designed your whole website? Then you should note the date you launched it. What about a new site feature or landing page that's driving traffic? Note the dates that you ran specific campaigns, or even more generic things like school holidays or industrial strikes, which might explain changes in user behaviour.

Bounce Rates - Is it your coupon?

We reviewed a scenario on a shopping cart where there was a high bounce rate at a particular step in the checkout process - the point of entering a coupon code. Turns out that some A/B split testing had revealed the high bounce rate was due to people leaving the site to go and look for a coupon code. Makes sense when you think about it; when was the last time you ordered Pizza Hut or Dominos without a coupon? That's one industry where almost nobody buys if they have to pay full price. The behaviour seen online in this particular example tended to prove that the meer presence of a 'coupon code' input field was putting shoppers off because they percieved that they weren't getting the best value without a coupon. So they naturally went off to Google and did a search for coupons for that site.

the meer presence of a 'coupon code' input field was putting shoppers off

Whether or not this is the case for every site remains to be seen but if you have a coupon input field, or are thinking of implementing one, there are a couple of things you should consider. The obvious one is to carry out your own A/B testing and find out how this affects your own bounce rate. You don't necessarily even have to have an active coupon to do this. Further discussion also revealed some ideas about using coupons only on specific landing pages where you are driving the coupon traffic to, which will prevent the need for an off-putting coupon code on your regular checkout page. Another option was to only enable the coupon field when there was an active coupon, and hide it when it didn't need to be there. Lastly, what about giving a coupon to every customer - advertise it on the site somewhere - that way people won't need to leave your site to go looking for one. It could be as simple as giving them 'express shipping' instead of regular shipping, or 5% off the order total, or free shipping. Throw in a free promotional product like a hat or stubby holder... Be creative!

Establishing Conversion Goals

Who knew there were different types of goals in GA? There have been plenty of articles written about conversion funnels on the Wiliam blog, and I always just assumed that they all used the same method for tracking against URLs. Turns out I was ill-informed and that there are a number of cool things you can do here. There are four different types of goals and I will quickly touch on each below.

Destination

This is the obvious one, where you have a multi-step form funnel (such as purchase, checkout, reload, enquire, etc.) and you place the tracking code / setup the goal so that it tracks when the user hits particular URLs, in order to ascertain whether the conversion was successful. To do this you don't want to get too tricky with clever (so you thought) AJAX transitions or slides between steps. Each step ideally needs to be on its own URL like /step1, /step2, etc. You can track the whole funnel (all steps) if you want to see data about where people may be dropping out of the process, or if you have a simple one page form or just want to know how many people completed the process, just track the end result like /thanks.html.

Duration

Now we start getting into new territory. Track the number of people who stay on your site for a pre-determined length of time. Or get more specific and limit this to time spent on a particular page.

Pages

Similar to duration but this type of goal doesn't track time on pages, but will allow you to nominate X number of pages as the total number of pages viewed in a single session, to achieve the goal. So if I was to setup this up and nominate 5 pages, then users would need to navigate to at least 5 different pages on my site in order for this goal to be reached. Why would you want to do this? Perhaps it's ideal for customer support sites, or perhaps your site only contains 5 pages so you want to know how many people are reading your entire site? 

Importantly, this goal type is linked to the session time which is configurable elsewhere in GA but defaults to 30min. Check this value if you are using page goals, as what this would mean is that the page count would reset after 30min and you may decide that you want this to be longer or shorter.

Event

The last goal type is quite a powerful one. Track things like how many people downloaded a specific PDF, or clicked on a banner, or exited to a specific external URL or through social media links. Pretty much any event you can trigger with JavaScript. This is a slightly more difficult one to setup compared to the other goal types because you need to tag the code from your site with events tags (like onclick). I used this on a site once because we couldn't use the destination type of goal because (as I mentioned above) we were stupidly using AJAX transitions on a multi-part form which sat on a single URL. So the only way to properly track the drop outs and successful conversions was to use event goals which fired on the 'Next' buttons and somehow the developers mangled it so that there were fake URLs in the web code so that it looked like separate URLs for each step. It was strange, convoluted and messy and the client didn't really get it in the end. We since redesigned the form responsively and used this opportunity to completely remove the AJAX component. Ladies and gentlemen, we have a winner.

Event goals are very useful and probably one of the more versatile types of goals. Think of all the places (events) you could fire off a tracking script, almost limitless! Want to know how many people watched a video? Fire the code when they click the play button. Then set a duration goal to see how long they spent on that page. Wouldn't it be great to know that people were actually spending five minutes on the page watching a video that is five minutes long? Or setup a script to fire on a link every time somebody downloads a document like a brochure or menu or order form.

Endless possibilities

In conclusion, I think the thing that I really enjoyed the most from the GA short course, was the realisation that nearly anything is possible. Think up a scenario and chances are that Google will enable you to track and report on it. I asked the lecturer if it was possible to setup a report for a client of mine who sold multi-currency travel cards. I thought it would be interesting to see a report about how many people from each country are logging into their account (like internet banking) but from different countries - e.g. to load funds, view transactions, etc. This would be a more valuable report (segment) than just a general view of people who are visiting the entire site. I suspect that a high percentage of customers who are logging in from places outside Australia are there purely for account management purposes. After all it's a travel card and these people are probably already using it - i.e. we don't need to sell the product to these people, we just need to support them and incentivise them to reload. It's highly likely that the majority of people purchasing the card in the first place are still in Australia, so this is where the marketing dollars should be spent.

How's that for identifying potential new markets!

With great success, I was able to learn exactly how to do this with a combination of new views which segmented the data and also through setting up a number of new goals. Now that the stats are rolling in, soon I'll be able to tell which cities and states in Australia are the biggest markets, and where the marketing dollars would best be spent. Do we have more people buying cards in Sydney than in Melbourne? What demographic of people purchases the most Euro currency? Should we start targeting campaigns to sell particular currencies to specific areas and not others, based on knowledge of previous sales history? I could even go some distance to answering another question I had, about whether people who were travelling to countries that trade in different currencies to those supported by this card (it only has eleven), were still in the market for one. The product can be used anywhere even if you can't pre-purchase every currency in the world, so the fact that I was in Sweden spending in Krona didn't matter, it would just convert money from, say, my US dollars purse, at the rate of the day. Wouldn't it be interesting to know if there was still a huge demand for this card for people travelling to Sweden, even if Swedish Krona wasn't formally supported through pre-pay? How's that for identifying potential new markets!

I think I'm going to have a lovely lunchtime chat with my client soon, and show them all the fantastic ideas spawned simply through the knowledge of what we can now track and analyse, thanks to the forever improving and changing world of Google Analytics!

P.S.> The course we attended was run through Sydney University, I highly recommend it. Book a spot here.