Historical snapshots… Can’t predict the future (or help grow sales)
“Big Data” and “AI” are a monolithic concepts these days. Everyone’s doing it, so jump in!
Right = Focus on a few key questions. Use the right tools to get the answers.
(Preview: this means statistical modeling and predictive analytics, not historical charts and graphs.)
I observed first hand in various sales, sales management and sales operations roles, plus something like a thousand (crazy but true) meetings with sales teams over the last 10 years, that it comes down to these four questions:
- Will we make the number?
- Are we focused on the right customers and prospects?
- How can we increase customer retention?
- What is the next best offer to drive growth?
What most analytical projects lack is the ability to actually answer these questions.
Here’s what I learned the hard way…
Historical charting and graphing won’t hunt.
Traditional charting and graphing focuses on one or two attributes at a time on the X and Y axes. This is a needle-in-haystack scenario because there are potentially hundreds of attributes combined in a complex system through time that contribute to business outcomes. What tends to happen is exactly the opposite of what’s intended.
Try this… On a long conference room table lay out printed charts of historical sales by 10 or more different slices, such as industry, deal size, profitability, product line… by month, quarter or year. Invite 5 – 10 different managers and executives into the room one by one to view the charts. Ask what insights they glean from the charts.
Charts Confirm Existing Beliefs
Here’s what will happen, and what happened to me, and why I ditched the historical charts in favor of predictive modeling and statistical analysis…
Each person comes up with a different point of view on what is important.
There’s no way to tell who is right, or if any of the charts on the table provide insights into what drives sales growth. What’s really going on is that each person looks for justification of their pre-existing ideas. It’s just human nature. We all have our beliefs, and when we see them reflected in data… BINGO. That must be the answer, especially because data never lies, right?
Rather than having what is supposed to be an analytical process help teams arrive at insights (things that we didn’t know before the exercise), charts tend to confirm firmly held beliefs, which aren’t really insights. On top of that, due to politics, the person with the highest rank in the room tends to be “right”.
My training at GE taught me that you risk getting fired if you raise problems without offering a solution, so…
Use historical charting and graphing for what it is meant for, to record history and understand segments of the business – but don’t expect it to uncover meaningful insights in a complex system.
Statistics to the Rescue
If you want to answer complex questions, use the right tools to analyze the data, which requires statistical modeling, and strap in for investing in a process that should be with you for the rest of your career in the Big Data era – predictive modeling.
Frame of reference: In 2015, searching for “Data Scientist” on LinkedIn yielded almost 18,500 hits whereas “Vice President of Sales” yielded 15,250. In 2018, those terms yield 51,900 and 19,500, respectively.
The train has left the station.