Harry Potter and Ron Weasley sit in their very first divination class. Today, Professor Trelawney is instructing the class on the mysterious practice of reading tea leaves. At the request of the professor, Ron makes a poor attempt to read meaning from the dregs at the bottom of Harry’s teacup.
“Well,” Ron mutters, “Harry’s got a sort of a… uh, wonky cross and that’s trials and suffering, and that there… that could be the sun, and that’s happiness. So, you’re gonna suffer… but, you’re gonna be happy about it?”
“Give me the cup,” demands Professor Trelawney. She quickly takes hold of the cup, glances at the bottom, immediately gasps and drops the cup.
“My b-boy,” she ominously stutters. “My dear… you have the Grim.”
And that, my friends, sounds like a web analytics project to me.
Two people…same set of data wildly different interpretations. The novice peers at the cup and sees a series of interconnected, but contradictory, patterns, thereby delivering a nonsensical prediction of future activities. The expert gazes into the dregs and reads an entirely different (and more importantly, a far deeper) meaning, with greater implications on how one should move through the future.
First, a caveat:
I am in no way saying that web analytics are magic. In fact, when done properly (for example, by our amazing analytics crew at Magnani), it’s a highly scientific process requiring hypotheses and experiments to measure the veracity of those hypotheses against the data we collect.
Web analytics are a wonderful thing. They allow us to see how our marketing programs and web properties are performing in the real world, with actual customers. But by their very nature, the data we collect from our web properties are measures of the past.
How can we use the past to make better predictions against the future? Better yet, how can we structure our dashboards to provider a deeper read of our customer’s digital activities… and how can we tie those activities to the measures that directly affect the bottom line of our business?
Hop on your broomsticks, and follow me.
Begin with the macro
The real failure in most analytics reporting isn’t that the reporting or the cursory evaluation of the metrics is wrong. Yes, it’s important to know how many people visited your website last week, what they looked at and what they did. The larger issue lies in not linking the micro metrics (visitors, time on page, etc.) to macro metrics (business goals). If, for example, you haven’t identified the relationship between the number of unique sessions on your landing page and the overall sales goals for your company for the year you’re A) wasting your time in putting the report together and B) not supplying meaningful data to the people who need to make decisions based on those metrics.
While developing a dashboard (and the activities that are represented by that dashboard), start with what you know. In any organization, most employees should know the business goals for the year. Is your organization targeting increased revenue? Increased profit? New users? Lower call center volume? What’s that magic number?
Now, evaluate the program you’re thinking of implementing. Does that program help any of those numbers? How? What specifically will your program do to move that magic number in a direction your C-suite would approve?
Do you have the answer to that question, yet?
Yes? Super. That’s the metric your whole dashboard (heck, your whole program) is centered around.
Next, focus on the micro
Once the macro-metrics are identified, examine the micro-metrics available to you. These are going to be the standard analytics often seen populating dashboards. Some examples may include: sessions, users, pageviews, bounce rate, etc. Think of these numbers as the raw materials of our dashboard. They have a direct correlation to our customers’ activities in the real world and most directly reflect the success or failure of a particular campaign or initiative. Identify which micro-metrics (or what combinations) marry best to the activities you’re wishing to encourage during your campaign.
But how can we infer that activities measured by our micro-metrics correlate with the large macro-metrics? This, my friends, is where experience and creativity begin to play greater influence.
Jump the gap
In most organizations, the keeper of the macro-metrics sits apart from the keeper of the micro-metrics (i.e. the CFO may not even know the name of the analytics manager). This is both common and unfortunate. Unfortunate in that the CFO might be pleasantly surprised by a spike in sales at the same time the analytics manager might be pleasantly surprised by a spike in web traffic that comes as no surprise to the marketing manager who has been busting her hump for the past two months working on a well-crafted, well-designed, action-oriented paid media campaign. But our marketing manager has no idea that her actions had a direct reflection in the sales numbers because that information is kept behind a wall.
So how do we change this? Shocker… it’s going to require communication. First identify the keeper of the information you need to inform the success or failure of your campaigns. Then become best friends with that person. What’s his or her favorite coffee drink? How many pets does he/she have? And the names of those pets? Set a standing meeting. Have a phone call once a week to discuss the high-level macro-metrics. You need to jump that gap between what you know (the micro-metrics) and what you may not be privy to (the macro-metrics).
Now take the information you know exists and structure your dashboard around it. Roughly, your dashboard should start with the broadest information on the first page and get increasingly granular as you move through it. Additionally, you may consider creating custom metrics that help demonstrate a helpful correlation between micro- and macro-metrics. Custom metrics add value by providing insight into the overall trends in your data, instead of merely the point-in-time perspective that many dashboards stop at.
Read the tea leaves
The real expertise in evaluating metrics lies not in viewing the numbers… it’s in understanding how the numbers reflect actions real people took in the real world, and how that activity helped or hurt your campaign. Attempt to see the unseen. And this requires two skills: the empathy to understand the how and why of your customers and the ability to draw a story from their actions as expressed in the metrics.
Developing those skills require repetition and reflection. It’s something we’re constantly working on here at Magnani, which is why we’re a great partner in helping our customers craft, execute and analyze their campaigns.
In closing, the more we attempt to look through the numbers and relate them to the larger stories of the businesses and clients we work for, the greater value we’ll inject into our dashboards, to provide real, actionable information our teams can use. In telling those stories we’ll be a little less Ron Weasley, and a little more Professor Trelawney.