Web Analytics and Data Overload
"If a measurement matters at all, it is because it must have some conceivable effect on decisions and behaviour. If we can't identify a decision that could be affected by a proposed measurement and how it could change those decisions, then the measurement simply has no value"
- Douglas W. Hubbard
The Value in Measurement and Analytics
"Not everything that counts can be counted, and not everything that can be counted counts" - Albert Einstein
When deciding how to approach measurement, we have always found that it can be very hard to remain focused on what really matters. So many new and shiny tools have existed and are dreamed up every day that we often see an approach that looks more like, "well, let's get the report and see what it tells us." That, unfortunately is the beginning of how data begins to lead us by the nose. More data does not mean more intelligence or insight.
The real value in measurement is when you create specific goals, determine where you want or need to be, create a strategy, and try different tactics to achieve the goal. Only then do metrics that track your progress serve a meaningful purpose. And only when they demonstrate the key performance indicators that apply directly to achieving (or not) the end goal. This is where data overload creeps in and often muddies the water.
Can Data Lead Us Astray?
Sadly, it often does. It is human nature to look for data that supports a theory, often a theory of "general progress." And with enough data on hand, it is usually quite easy to find something in the dataset that has positive value. But that indicates obfuscation of the original intent - a willingness to only see the forest for the trees. Change is hard. Measuring progress is actually fairly easy. Especially when you can select the data from many options.
What Makes Data Tell Us The Truth?
Sadly, no data set is usually the full picture. But the biggest obstacle we normally encounter is that there is often a lack of understanding of what exactly is being measured - the base indicator - and why it is important. For most businesses, the most important indicator is sales and revenue. When revenue is increasing, businesses are normally growing and thriving. Profit is normally the next most important metric. How does web traffic relate to these numbers? That's where it gets tough to do a meaningful correlation. Is web traffic a leading or following indicator (if we take time into consideration as well)? Is there any correlation at all? These questions clearly depend on the specifics of your business. Online retail is clearly nearly a direct correlation to web traffic. Most services businesses have very little direct correlation. But does that mean web traffic is unimportant to a services business? Not at all. It is equally important in our estimation, but for different reasons and on a different timeline. Web traffic for a services business is often a very early predictive indicator (as a broad generalization).
What Can Data Tell Us?
On one project, an automotive online model configuration tool, there was assumed to be no real direct measurements applicable for the broader business. But upon post-deployment data analysis, we found that the color and model selections predicted (in advance) the general distribution of cars that the manufacturer would be selling. When the assembly line planning function began to consider the configuration tool in its process, overall predictive product sales conversions dramatically increased and excess inventory decreases. Customers were sending a message by using the tool. The business just needed to know to listen and use the information intelligently. This was a case where the analysis created a strategic insight - the tail wagged the dog - but its a great case that allows our clients to think a bit more creatively about the real-world data that they may already have and how to harness it for greater efficiency.
If you offer twenty services on your website and you see online traffic increase for one of those services far beyond the others, where would you think to add more staff? Clearly, business indicators are often slightly indirect and nebulous, but it is generally far better than having your business get wagged by things like internal politics and personalities, isn't it?
How Much Data Is Too Much?
Our opinion is any reports that are not clear in what they are showing or what it means are not useful. Data can certainly be used for investigation and discovery purposes, but efforts like this need to be very separate from day-to-day operations. And it should be clear that they are exploratory. Like our balanced marketing approach, data and reporting should also be approached with a risk-mitigation strategy where investigation and exploratory work should be only a percentage of the overall effort. The basics should be in place and well established before more speculative data gathering activities commence.
Should you want help getting your basic reporting and metrics established, or if you need to take the next step into exploratory data sets and correlation analysis, give Envigna a call. We're always happy to share our knowledge and help make sure your direction is headed directly toward success. We love geeky data more than we care to admit. And we treat your data with the privacy and respect it deserves. Give us a call.