The One Thing You Need to Make Sales Scientific

Written by Kally Pan on Apr 12, 2016

One of my first tasks starting out at AppInsights was to help companies set up their sales dashboards. Coming fresh from a PhD program where I spent 7 years on quantitative biology, I was looking forward to using my analytics skills in a new field. So like a good little scientist, I began with a literature search (a fancy word for Googling). Within seconds I’d found dozens of articles with titles such as:

  • “Science and the Length of Your Sales Cycle”
  • “The New Science of Sales Force Productivity”
  • “Bringing Science to the Art of Sales”.

Wow, I thought in relief, looks like they’ve pretty much got it figured out; understanding and setting up sales analytics is going to be easy.

data-science-for-sales-blog-1.png#asset:

Years later, I look back and realize that most of these articles talking about “scientific sales process” had no clue what the word even means. Despite good intentions, I found most of the “scientific examples” in these articles to be… well, unscientific.

Process is not science

Don’t get me wrong, many of these articles were informative, thoughtfully written, and had loads of useful tips for organizing sales data. But the idea that a methodical process is science is absolutely false. Here is one particularly cringe inducing example;

“By the way, did you notice the word – Process. Right away this screams science [emphasis mine]. It’s a documented, repeatable way to generally get the same results.”

By this definition, running a factory, sewing a garment, or even painting are all sciences. They all involve “documented, repeatable processes that will produce the same results”. But these along do not qualify science. Science requires is the formation of a hypothesis about how things work and controlled experiments that test that hypothesis. Not to say that there is no science involved in factories, garments, or paintings, but if you are not testing, then there is no science.

Looking at numbers and developing methods doesn’t make you a scientist, any more than taking of drugs would make you a doctor. You can try to optimize sales processes through repeated trial and error, but it is not science if you’re just pulling levers on a black box. More importantly, it will not work to produce long term improvements that you can build on.

How do you test?

What makes for a Scientific sales process? Though numbers and methods dominate the scientific world today, they are by no means a requirement. What is required is experimentation, testing! It’s not enough to just see a bump in numbers and guess that two things are related, you must experiment to understand the nature of that relationship. In other words: “Increasing x causes an increase in y” is NOT science. “Increasing x causes an increase in y because…” is science.

Sadly, the first scenario is where most sales processes stop. I mean, once you know a lever works you just want to press it again and again right? Well, sometimes it's ok. Many relationships are either very obvious (sales for Christmas cards are up because it’s Christmas), or impossible to verify (changing the color of the catalog decreased sales... somehow).

However, understanding the underlying principle that drove a particular sale can help leapfrog incremental improvements and lead to a big breakthrough. In sales, this frequently happens intuitively. Like when I finally understood why asking for a budget or a time line is an important step in the sales qualification process; my efficiency skyrocketed.

But when sales processes are more complicated, real experiments are necessary to reveal the relationship. Case in point, a client was only concerned about increasing traffic to their site. But when pressed, they tested to see if that had any impact on their business and found that revenues actually decreased. Only by instituting dashboards that monitor sales, support, and traffic together and testing more hypothesis did they actually understand what was really happening. In their eagerness to increase traffic, they used a number of services which brought in low quality visitors who never had any intention to buy, and diluted the experience of the visitors they organically generated.

The turning point was when the client tested their assumptions. This may seem trivial, but poor assumptions are almost always the reason why bad decisions are made. And assumptions are made. All. The. Time. It's almost unavoidable. 

Ask questions, test!

Sir Ian Fleming's discovery of penicillin was accidental; moldy oranges in his lab produced a chemical that contaminated his cultures and killed off the bacteria that he was growing. But if he didn't pause to ask why his bacteria didn't grow around the mold, he would have just thrown it all away since most people would have assumed the contaminated samples were useless.

Penicillin producing mold isn’t quite the same as sales in the post internet world. However, science has a lot to offer the world of sales and business in general. Experimentation is hard but they work much better than just trial and error. The payoff is so much greater, and the results are something you can build on.