Where To Find Cash to Run More Analytics
If you are short on cash for headcount or other analytics projects or tools, then consider cutting back on the number one waste of cash . REPORTING. Now I have gone back and forth about whether to push this post because it can be a controversial topic. It is not the report itself that bleeds cash but the way your team's time is used to produce something with very little value. You can determine if any report has value by asking two simple questions: Was it read? If so, then what actions were taken from it? Now, on with the post.
Most people think of reporting and its interpretation as analysis, but this is incorrect. Unless your report is actually built on a data model where test ideas and actions flow from beautiful visualizations and subsequently generate business impact, you really just have a big dump of data. And that is definitely not analysis. To be clear, I am not saying ALL reporting has no value, but I am saying MOST reporting suffers from two problems that make it low value and high cost. First, it's labor intensive, even when automated. Second, it lacks impact.
Reporting is labor-intensive
In almost every case, recurring reports are used as an extensive data dump to monitor what happened. How many people visited the site? How many likes did we get? How many emails were sent? This data is so boring and useless that most people would rather clean the dust out of their keyboard than suffer through it.
To make matters worse, when insights are not forthcoming from this mind-numbing data, then more data is added which further inflates the reports and simultaneously increases production costs. Labor costs come in the form of the time to combine data, verify data, interpret the data, track down anomalies, and then package and send detailed reports.
So let's assume for a moment that you have two analysts spending 80% of their time running reports. A fully loaded headcount (salary, benefits, overhead) is about $200K so you're actually spending $320K a year on these reports just to get them out the door.
Consider this quick internal test. Stop or delay all reports and see who you hear from. Do you hear crickets? Who notices? By all means, keep those reports that are useful. Then reverse the analyst work ratio to 80% analysis and 20% reporting and redeploy the people you already have. Pour this time and labor savings into analysis.
Reporting lacks business impact
Second, stakeholders who receive these expensive reports barely look at them. I hear this all the time. It's true. And while the boring nature of the data dump report does not help, in my experience, the root cause is really a lack of clear guidance from senior executives about how to marry digital data to the overall business. For many big companies, the digital arm is the fastest growing, but still less than 20% of the overall sales volume. Yet, digital data is an excellent leading measure for most customer activity. When properly unpacked it contains nuggets of insight about product innovation, loyalty, marketing effectiveness, product marketing positioning and much more.
If executive guidance is vague, then consider letting your analysts suggest a more precise direction by diving into the data. Give your bright analysts access to as much data as possible. This includes not only sales databases, but also market research data, focus groups, survey data, and of course, web and mobile app data. Guide their discovery with what is most important to your business in the next 6 months? Cost savings? New revenue sources? Understanding customer behavior? Increasing loyalty? Do not skip this step and send them to find a needle in a data haystack.
With a goal in mind, what will come back are correlations in digital data, which are simply educated guesses about what looks interesting with respect to this business goal. Your analysts will then need to branch out within the organization to business stakeholders to explore theories about why these correlations exist and what they might mean. They may quickly find an answer that explains the correlations - or they may need to run tests to check assumptions.
The time they spend mining and testing will generate actual nuggets of business insight - even if the answer is to say there is no connection. Knowing what's not impactful is as helpful as knowing what is. And in some cases, a serious competitive advantage might be found. eBay Motors, for example, was actually developed this way a decade ago and quickly became the number one revenue category for eBay.
Analyze data first, Generate reports second
Mining the data first takes advantage of in-house knowledge which is often spread across departments. It has the effect of bringing smart people together to jointly solve a more meaningful problem. When you start with exploratory analysis, your analysts will be happier and the reports you eventually build will be smaller, better understood by stakeholders and have more business impact, locking in savings and productivity for years to come.