Guest Post By Pere Rovira, Elisa Group
Bigger than data
Whenever I kick off a client conversation about data-driven decision making, I ask the client to give me three concrete examples where analytics significantly helped increase revenue, reduce costs or improve customer satisfaction, i.e. the three most fundamental KPIs to any business. Some clients use Google Analytics while others spend a fortune on Omniture or Coremetrics, analytics solution providers that perform advanced statistical analysis with complex, expensive business intelligence software packages. However, all of them have something in common: when you ask about concrete examples with numerical results attached to them, they have a hard time providing you with three examples.
The problem is not big data: the problem is bigger than data. Don't get me wrong. I do believe in data and I am convinced that data-driven insights lead to a better world; and particularly, to a better website, app or marketing campaign. But data does not lead to insights automatically. There are two main blockers to transforming data into insights: data visualisation and resistance to change. The rest of the article will deal with them.
Analytics software is terrible at visualization. Endless tables of numbers hardly lead to any insights, yet, still, most of analytics is about reporting with tables. The more expensive the analytics package, the more dimensions and metrics you can add to the table. But the human brain is not particularly good at dealing with large amounts of data in the form of tables. Furthermore, the increasing complexity of the digital environment means tables are increasingly larger, and insights therefore harder to reveal.
By applying some basic data visualization techniques to your data, not only will you find insights faster, but you'll be able to communicate them effectively. Managers will understand you. Action points will be defined. Changes will be made to the website/campaign/app/etc., some of them will yield positive results and, hence, data will be appreciated. You'll have a positive story to share with me when I kick off a conversation with you about data-driven decision making.
There's no better way to communicate insights than to show a client a scroll-reach heatmap or a click heatmap. It's the fastest possible way to understand that a particular page (landing page, payment form, etc.) is performing poorly. Everybody gets it. Your boss gets it. You do not have to talk about bounce rate, click through, event tracking or conversion rate optimization to convince them to change that button. If you just "show" them the button is not clicked, or not seen, everything's that much easier to understand and, most importantly, the need to take action urgently is understood perfectly. By showing the client what happens in such a visual and transparent way, you produce that "eureka moment" for your boss or your client. They see the insight and they're happy because they're not expecting such straightforward data . They will take action, for sure.
There are many visualization techniques. Try using radar charts and scatter plots for a change (google analytics motion charts are great at it, for instance). They are great at presenting the big picture, because they can deal with many dimensions at a time. They are great at revealing the unknown patterns, the outliers, the sources of interest. Don't be afraid to make your data look sexy and appealing. Spend enough time on it, it won't be wasted. I would even dedicate more time to present results than to gather them. Just like good usability might mean the difference between a website success or failure, good data visualization might mean the difference between being heard or being ignored.
Resistance to change
In physics, we can't even make definitive predictions about the movement of a system of three particles. How do we expect to predict the "movement" of thousands, even millions, of website visitors? It's impossible. For all the data we may gather, we can't know the truth because there is none. There are only better conversion rates. The good news is that the objective of a business is not to find out the truth, but to increase its conversion rate.
Many companies approach data as the answer to their problems, as the way to point out where to go, as the solution to a failed business model or a lack of marketing creativity. Big mistake. Data can inspire ideas, sure, but what it does best is to test ideas. Therefore, it means that the most effective use of data involves testing and change. Change, however, involves people, and people do not like to be involved many times.
Why should the IT department be bothered about an AB test to increase the lead-generation conversion rate? What's in it for them? Some organizations are so big, that its departments behave like isolated entities inside the company, with their own objectives hardly ever aligned with the overall company objective of making profits.
In order to overcome this resistance to change, shared objectives for the entire the organization are a must. This may sound obvious, but it's probably the hardest thing to achieve. Other solutions involve giving more power to the analytics function, for instance via the creation of a CAO (chief analytics officer) position in the organization. Finally, a more inclusive approach proposes a different way to understand traditional company functions. For instance, substituting the traditional CIO (chief information officer) and CMO (chief marketing officer) roles for a more relevant CMT (chief marketing technologist) position, that combines the strength of the others to adapt better to the digital world. Scott Brinker, one of my favorite bloggers, has written extensively about the necessity and advent of the CMT.
In summary, in this article I have argued that both resistance to change and poor data visualization are the two main blockers to effectively implementing a big data strategy that drives more revenues, reduces cost and improves customer satisfaction. A combination of new technologies and organisational structures is needed in order to overcome these blockers. The problem is not big data, the problem is bigger than data.
What's the experience in your company or with your clients? What techniques or strategies do you use to make data useful? Now it's your turn. What do you think? Does what I have written hold true?
About the Author
Pere Rovira is the Country Manager of Elisa Group in Spain, specializing in online business optimization. He has worked as a "formador" for businesses including Prisa Group, Softonic, La Caixa o Atrapalo. Rovira is also a regular conference lecturer at SMX in London, eMetrics, Practitioner Web Analytics, ESADE and at the University of California. Rovira holds an M.Sc. in Physics from the University of Barcelona, and has a post graduate degree in Digital Economics from UCLA Berkeley, as well as a M.Sc. in Information Systems from the London School of Economics. He is currently a professor at the Autónoma University in Barcelona.