As a sophomore in high school, I started working at a local running store. I was on the cross country and track teams, an obsessive nerd when it came to running gear, and so it only made sense to find a job that combined both. At the shop, I spent most of my time working one-on-one with customers to help fit them with shoes and occasionally select various apparel.
When I say I was an obsessive nerd on running gear, I truly mean it. I could tell you the stack heights of every shoe, the chemical compounds that made up the soles, and at one point had every color code memorized for 50+ models. I spent most days wearing 2–3 different pairs of shoes while I worked, "product testing" the models so I could speak to them later.
Working in a running store brought a wide variety of customers in. There were plenty of high school and college runners, older adults who had been running for a while, new-to-running persons trying on their first shoes, and plenty of other variations. Given the variety of the customer base, I was involved in a lot of different interactions and as a result, really honed my ability to sell shoes.
The mindset I held when selling shoes was that I never particularly cared if an individual bought shoes. Now, that may sound counterintuitive, but there’s a reason for it. My focus was on finding someone a product that would create a positive experience between them and running. Forcing someone to buy a shoe wasn’t in their best interest, and consequently wasn’t in mine.
So instead of being a "How can I sell this shoe?" question, it was "How can I find the right product for this individual?". This is where things get interesting.
When working with a customer, it’s important to know your audience. If I was working with an experience, competitive runner, the conversation would be very different than if I was working with a first-time shoe buyer. In the first case, I might dive more into some of the technical specifications, explaining the weight differences, material styles, and how they would impact running. In the latter situation, the conversation usually goes a bit differently. Instead of explaining the difference between engineered mesh, EVA foam, and the likes, I would frequently explain things at a higher level, comparing shoes to pillows and springs.
The basic information being communicated was still the same, but the format was different. Knowing who you’re talking to is a huge aspect of sales and being able to put technical language into a comfortable conversation makes it easier to solve the problem at hand.
So how does any of this relate to data analytics?
Well, the problems encountered in data analytics are pretty much the same as in a running shop. In data, you can get as technical (or as non-technical) as you want, building anything from a pie chart to a neural network. However, the solution(s) you end with depend on…
- Who is your audience?
- What do they really need?
In many cases, the information presented only matters if you get buy-in from the audience. If your boss comes from a non-technical background and you throw up slides of t-tests, ANOVA Analysis, and hyperparameter tuning, you’re going to lose them real quick. And if they don’t buy in, then the work you’re doing really doesn’t matter because it never makes it off your desk.
So much like I sold shoes, I now have to sell Data analysis.
While it’s important to be able to do good data analysis and have strong knowledge of the technical skills required, it’s arguably equally (or more) important to be able to sell and communicate that analysis. At the end of the day, if those around you aren’t willing to buy into using data, all the analysis in the world is meaningless.
When working with the data, think about how you would sell it to someone. What benefits does it drive? How does it fit their need(s)? How easy is it to understand?
Working on a problem in data is much like selling shoes. You have to understand your client’s background to determine how you’ll present information to them. You need to be able to translate that technical information into something understandable that makes sense. Sometimes that means statistical deep-dives with scatterplots, autocorrelation graphs, and a table of model parameters. Other times it means a bar graph. And while the prior may seem like it conveys more information, it only does so if your audience understands it and can connect it to the problem.
A mountain of information is no better than a single fact if the client can’t understand it.
It’s easy to get caught up in the allure of data, thinking that the quality of the analyst is derived from the depth of their work. What many (young) analysts often don’t realize is that unlike school, where you’re often required to perform pages of analysis, the real world focuses on impact, and impact is tied to understanding.
With that in mind, how are you going to "sell" your data?
