When we consider that 90% of all the data in the world has been generated over the last two years and the volume of business data worldwide is expected to double every 1.2 years, it’s no wonder why so many of us often feel lost in what can only be described as a data jungle.
This dense forest of information can be hard to navigate at times and many businesses don’t know where to even begin, but there’s really just one way out of the jungle and that is by turning data into insights.
Based on some key learnings I took from the Google Analytics 2014 Conference in Sydney, hosted by Loves Data, I want to share 6 things you can do to to make your data analytics feel less like a tangled, wild jungle and more like an open savannah.
1. Begin with the end in mind
Tim Wilson, Senior Partner of Analytics Demystified, made a great analogy of data being like sand. Every grain is beautiful, unique and interesting. Every grain tells a story. But it doesn’t tell us the bigger picture. What we really want to know is what is happening in the desert. What shape are the dunes and what does that tell us? When analysing data, we often get lost in the beauty of the grain of sand instead of what the dunes tell us about the direction of the wind.
The same point can be made with regards to planning and measuring your online activity. Jim Sterne made the point that if you don’t invest this time in the beginning and if you’re not sure of the reliability and accuracy of your data, then everything you do next is essentially meaningless. You need to understand your data, the analysis tools you’re using and the question(s) you are seeking to answer.
2. Remember the bottom line
Whatever question you’re trying to answer should link back to your organisation’s bottom line, whether that be to sell coffee, sell loans or drive interest in your local region. Whatever your business objective, everything you do online, all the data you analyse, and the questions you seek to answer should ultimately help you to achieve your core goals. If they don’t then you’re probably wasting your time.
3. Tell stories
Let your data and your analysis tell stories. Don’t worry about the details of the story, just pick the key headline, choose a nice emotive image to illustrate it and ensure the story relates to your bottom line. For example, “we saw a 25% increase in revenue from Adwords this month” tells a much more compelling story than “30% of sessions came from PPC; 25% from organic, etc.”
Also when illustrating your insights, avoid creating a jungle by using complicated charts (like pie charts). Instead, always think of what will create the clearest picture. A key image that stayed in my mind was a bar chart typically titled”% Mobile Sessions by Device” with the headline “62% of Mobile Sessions came from Apple Devices”. This is a great example of how the headline tells the story and I don’t have to decipher the chart and think about it.
4. Use segmentation to get more clarity
To start untangling all that vegetation in the data jungle, you’ll need to think outside of the GA box (aggregate metrics and session-based metrics) and start adding your own segments to really get clarity.
Aggregate metrics tell us very little about our audience and how they behave. While there are some good segments inside the GA box (new versus returning users, traffic source, device type, geolocation etc.) also consider adding your own segments, such as converted users versus non-converted users, readers versus non-readers.
Event tracking can also give you a greater depth of information on the kinds of interactions people have with your website. You can track pretty much anything that you think contributes to (or hinders) your primary goals, such as video engagement, scroll tracking and form abandonment.
5. Encourage contribution & testing from all levels of your organisation
Get true optimisation happening by getting buy-in from all levels of the organisation. This can be easier said than done, particularly if your current activity is working well. But as we know, it can always perform better, right? Allow everyone from your client to CEO to contribute to your hypotheses and educate them on the importance of testing over simply continuing with the tried and tested method (as reliable as that may be).
Test everything you can but understand your limits and consider the effects on offline events, for example, are you getting more phone calls due to an online test? Develop a testing road map to ensure you are covering all angles, and be sure to triple-check your results.
6. Develop frameworks
Frameworks provide a consistent process that you can use to communicate with your stakeholders. There are a number of frameworks that you can apply to analytics. At Traffika we tend to apply a methodology similar to the DRIVE Optimisation framework discussed by Carey Wilkins. This consists of Discover (identify needs), Reveal (produce measurement plan), Implement (design and implement analytics configuration), Validate (assure quality of data collection and configuration), Examine (produce reports, visualise and analyse data) and Optimise (make recommendations and take action).
Other frameworks offered included the LIFE cycle (Learn, Investigate, Favour, Engage) framework, which is a digital version of the original AIDA marketing funnel; and another based upon Maslow’s hierarchy of needs.
Still feeling a bit tangled up in the data jungle? Leave me a question or comment, I’d love to help!