Manju: Jim! It’s great to see you.
Jim: Hey, Manju. Good to see you as well.
Manju: How are things? Hope all is well.
Jim: Manju, to be honest, it’s been a tough couple of weeks at work.
Manju: Oh, I am so sorry to hear that. What exactly has happened?
Jim: So, we introduced a few plant-based menu items across all of our locations and for the most part, the locations have increased sales by 7.5% after their introduction, but we have five locations with pretty bad sales numbers. I guess vegetarianism isn’t too popular there.
Manju: Oh. No. What was the reason for introducing plant-based items?
Jim: There are a couple of reasons – (1) our competitors have started offering plant-based options and so it became quite important for us to introduce these items to remain competitive in the market; (2) a lot of people are moving toward vegetarian options due to health concerns and offering these items helps us maintain our credibility of serving healthy options.
Manju: When did you introduce plant-based items on the menu?
Jim: Couple of quarters (six months) ago.
Manju: Do you know what are the sales numbers associated with those five locations?
Jim: On an average, these five locations have sales increase of 3.7%, which is considered bad since we had higher expectations. Now our challenge is to increase sales of these five locations by 7.5%.
Manju: Why only by 7.5%? Why not more?
Jim: I know increase of 7.5% may not be the best but it’s much better than what our underperforming locations have done so far. Once we achieve 7.5% growth at all locations, we can focus on further growth.
Manju: Makes sense. How many locations have sales increase of 7.5%?
Jim: There are twenty locations.
Manju: Do you know why those five locations are not performing well compared to those twenty performing locations?
Jim: Exactly! We have no idea. I mean, we have guesses, but we don’t want to spend money to fix a problem without knowing what the problem is.
Manju: These are new items that are added to existing list of options (including meat-based options), right?
Jim: Correct.
Manju: Do you have any data to look at?
Jim: I knew you would ask that Manju. This is the thing that has given me the most anxiety over the past few weeks. Of course, I can look at the data, but the company wants to go a step further by hiring data experts to run a data science project. These experts would be doing a big part of my job. I don’t know that the company will even need me if they have data scientists performing virtually the same tasks.
Manju: Look, experts would be helpful, but you don’t need experts to run a data science project. You could do it if you wanted.
Jim: I’m not a numbers guy. That’s my problem.
Manju: The biggest misconception about data science is that you have to be a numbers guy to do it. To be honest, a software program or, more precisely, a statistical analysis package does a lot of the math for you.
Jim: Oh, I see. I haven’t heard anything like this before. Sounds so good to my ears.
Manju: Coming back to my earlier question, do you have data to look at?
Jim: What type of data?
Manju: Sales numbers in the last quarter (past three months) for all locations. I want to see how these numbers are distributed.
Jim: I think I can get these numbers easily.
Manju: Great. Jim, remember it is important to have “the right data” for data science projects and that “data should be right.” The second part “data should be right” is exactly what it sounds like: is the data correct?. The “right data” is a bit subtler. Ask yourself “Is this the best data to answer our questions?” If you don’t have the right data on hand, it is important to find or collect new data which pertains to the situation. From your sales data, we can look at the distribution and check if sales increase numbers of all performing locations are in the vicinity of 7.5% or not. I also want to see if sales of all non-performing locations are isolated from those of performing locations.
Jim: Sure, Manju. I can get this data.
Manju: Excellent, Jim.
Jim: Manju, you mentioned two important aspects of the data. I understand “data should be right” part. For “the right data,” how do we ensure that the data is right and who is responsible?
Manju: Great question Jim. In your case, you are responsible for sales data. What I mean is, for sales data, your sales team is responsible. They are the data users. Data users are the best people to know what is the right data to measure overall sales. This is true with all other functions. The regular business people who deal with data in their respective areas will know what data is needed for their work. If the right data is not available, it has to be collected. Having the wrong data presents a huge risk to the organizations. Sometimes, “the right data” part can also be addressed through a good problem statement that describes problem at hand very clearly. Once we know what data is needed and obtain it, then we need to understand impacting factors causing the problem (in your case – impacting factors causing poor sales). After identifying those impacting factors, we need to make sure we have right data on those factors to perform root cause analysis. For these factors, we also need to ensure data quality. One last thing: people sometimes use the phrase “fitness for use.” That just means you have the right data and that data is right (Source: Redman, 2021). The concept of “fitness for use” was introduced by Dr. J.M.Juran, a famous quality expert (Source: Redman, 2021).
Jim: Thanks Manju. This is really getting very interesting. Also, based on what you told earlier, one need not be a numbers guy to run data science projects, right?
Manju: Yes. That’s what I meant.
Jim: So what does a data scientist do?
Manju: I can teach you if you want.
Jim: Now?
Manju: Yes. That’s how simple it is. Data science is more about structure and organizational cohesion than math. One should have some experience in understanding real-world problems. Here, let me recall a quote from a great Indian saint Swami Vivekananda:
Knowledge can only be got in one way, the way of experience; there is no other way to know.
(Source: VivekaVani, 2019)
So, with your experience you can easily solve data science problems.
