Tuesday, March 11, 2008 | 2:00 PM
This is the first of a series of practitioner posts from our team on best practices we've seen through developing and working with customers on Google Analytics.
A Planned Approach
There can be few scarier moments in the humble Google Analytics practitioner's life than realising that a simple error or omission in the past has amplified over time, to play havoc with your data in the present day.
It's important to set up your filters, goal settings, and custom segments from the outset according to your well defined strategic plan. If you don't implement Google Analytics with an eye towards what you'll want to know in the future and how you'll want to monitor your campaigns, you may feel reluctant to make the necessary changes down the road. Even if you do make changes, you won't be able to apply those changes to historical data.
So, how do you produce data with integrity and immediate business application from day one? It pays to have a plan. Long before you think about pasting that little tag on your pages, you should consider a few important and underrated questions.
1) What data do I actually need?
Here we have a question so blinding in its obviousness, that it actually prevents people from asking it. But it is critical to define what it is you actually need to know, and do it rigorously.
Which metrics and measurements are actually going to inform the decisions you make day to day? Any web analytics product you use can produce a torrent of data and that data may seem like an end in itself. Be ruthless about each piece of data you ask for. Are you going to use it, or pat yourself on the back simply because you and your tool are smart enough to collect it?
2) Is the data I need available by default?
This is where the hard work really starts! Once you've been alone with your thoughts (and hopefully the thoughts of a few other interested parties) and decided what it is you want to know, you'll need to figure out whether this information is available by default or if you'll need to apply custom filters, user-defined segments, or goals.
Google Analytics has many reports available by default. They cover everything from the popularity of your pages of content to the keywords driving traffic to your site, to the screen resolutions of your users. However, you aren't the average user are you? Your needs are specific and important, and may require planning up front.
Do you want to be able to easily see the purchasing trends of your returning visitors as a group? Do you need to know the numbers of unique visitors to segments of your site? Do you need to separate out the visits of your registered users from casual visitors? All of these are possible, but all require customization to capture. Make these customizations at the outset so that you'll have a consistent historical set of data.
3) Do I know where I am going?
A degree of honest self-appraisal is required here. At first, it may be hard to imagine that you'd need more than what is offered by default without any custom filters or segments. Alternatively you may already know that you'll want to explore as much as you can.
Implementing any Analytics tool requires rigorous thinking, even when that tool is free. Prevailingly, the insights you can glean are tied to the boundaries set by the scope of your initial set-up. Sudden alterations of direction are tricky maneuvers to pull off, both in terms of technical costs, and the danger of creating a breach in the comparability between the two sets of data.
Be bold in scoping the ambitions of your Google Analytics implementation. Expect that your first taste of successfully applied analysis will create a hunger for more.
It's easy, when setting up web analytics on a site, to regard the numbers themselves as the ultimate goal. You may rush to simply get the code on your site and see reports populated with data. It's easy to perceive the challenge as a technical one, a question of how to apply a tag to a page, rather than a theoretical and organizational one. This is an understandable, but sometimes misguided focus.
By taking an insights-based approach to your initial implementation and configuration, you will be able to implement efficiently, rely on your reports, and future-proof your need for data.
Hopefully you are up for the task yourself, but if not and you want some help with these functions, we have many Authorized Consultants who are always happy to help. You can take either path, we wish you all the very best in your analytical journey.