A Data Science Central Community
Good Morning and Welcome to this addition of the Morning Analytic Coffee Blog.
Sorry for the recent delays, between holidays (Happy Canada Day and 4th of July, everyone!) and some personal items I’ve attended to.
For a while, I’ve been generalizing about Analytics. Today, I am getting into what I do, and that’s transportation analysis. Specifically, I find out the quantitative and qualitative reasons goods and services are transported, then I analyze the process, suggest improvements and map the dynamic inflows and outflows. It’s a niche, but it’s what I do best. It covers anything with an origin, a destination(s), something(s) carried and a purpose or purposes. All four matter. Otherwise, it’s just a random trip.
To my firm, analysis is just that: we look at all aspects of moving objects and what they carry, people or cargo. We analyze origins and destinations, flows, etc. These are all good things, must-have information in this world. However, I think qualitative analysis is often given the short shrift. It provides a broader of picture of why goods and services are transported, and the reasons companies make micro-level decisions to export, other than “for the money”. Ultimately, everything moves “for the money”. The reasons that lead to the decision to move goods is fascinating, both quantitative and qualitative.
Here, I think, is where the possibility of poor decision making lies. A given product may succeed, fail or muddle through its lifecycle based on demand, but the process of deciding how to move it and where to move it to logistically, which customers gain trust, and what type of business model a company engages in is all based on qualitative process. How do you model the qualitative as well as the quantitative?
My company focuses on both. A quantitative model begins with the questions: What and Who? Then Where, When and How? Once you determine the what (product) and to whom (customer), the next you have to agree on a place, time and method of delivery, for example, to the customer’s warehouse, by next Tuesday, and by plane then truck. Over the course of more than 100 shipments, a customer model begins to develop: the customers who make you the most money generally use your product to create a given profitable result. A rule of thumb often floated around is that your “best 20% of your customers produce 80% of your revenue.” But how do you know that’s true? My preferred method for carrying this out is to gather the data from the firm and then see what results those five questions bring, and then create seasonal / locational variance maps and deliver what sort of goods move where and when. This establishes patterns of strength in delivery, and allows firms to consult with logistics companies that can help meet their demands. Knowing your supply chain and customer base gives a firm the advantage and helps the logistics provider understand delivery better. This can, in turn, create a better partnership for the medium to small business.
Qualitative analysis is different, however. Let’s say that my Toronto-area firm does business in Buffalo, NY, Montreal, and Vancouver, primarily. I know that my A type product does well in Vancouver, but my new B product has taken over in Montreal and Buffalo. Sure, I can ask my buyers “why?”, and get some understanding. However, that’s not qualitative analysis. That’s confirming what you probably already know. Qualitative analysis takes quantitative analysis above and then asks: Why does ABC Company in Montreal, and DEF Company in Buffalo by primarily during winter? Obviously, with something road salt, the answer is obvious. However, what about auto parts that have nothing to do with seasonal variance? Does the customer see benefit in adding more brake pads in June than in September, as an example? From these types of question, called Why or Purpose questions, a better bond is formed, and in addition, you may gain better understanding of your customers’ secondary concerns.
In a recent Globe and Mail article, businesses expressed concerns that skyrocketing electricity costs are a major business threat. If your customer comes to you and says we are looking at your product because it helps us cool while drawing less electricity, then you know that not only is the product working, but that macro-level conditions are impacting their business, which can lead to research and development emphasis on creating a product that lessens dependence on electricity. A map of qualitative analysis would then give reasons for choices based on location, and allow a firm to identify new opportunities early.
In passenger modelling, some customers choose on economic necessity (speed or distance for air flight, public transit vs automotive ownership) but some choose productivity (driving itself is not productive time) and others choose for environmental or other non-quantifiable reasons. These choice models can also be identified by location, such as census tract or postal code, and qualitative data assigned to the quantitative model. There are many approaches, but only in understanding both sides of the equation, quantitative and qualitative, is the data analysis holistic and most useful.
In concert, a fully-fledged customer model appears, answering what, who, where, when and how, and why. This, in turns, provides deeper customer knowledge, and gives you the reasons for customers not just buying your product, but choosing your company. Thank you for reading.