Whether the future is merely a function of the present may still be a matter of debate among theoretical physicists, but all agree that your ability to predict the future is limited by your knowledge of your present.
When considering an investment, liquid funds are a fact that you probably want to take into account. Your bank summarizes all your past transactions and provides you with a simple figure, your account balance. While it is safe to assume that your bank gets the numbers right, pitfalls lie in the misinterpretation of what this number actually means. You are probably very aware of this when it comes to your account balance. For example, you would factor in upcoming obligations. But not all cases that seem obvious actually are.
Let’s take a look at a real world example I encountered when working for a German wholesaler. One key indicator that is of great interest throughout this company is the delivery time of items ordered from suppliers. The calculation seems straightforward:
Delivery Time = Date of Delivery - Date of Order
Complaints about the accuracy of this metric from several departments gave rise to investigations. Contrary to expectations, there was no error in the calculation or the underlying data, but rather a major misunderstanding in the interpretation of the results by various departments. See, “being delivered” actually has different meanings, depending on where you are working. For the receiving department goods are delivered when they are offloaded the truck. For the customer service, on the other hand, goods are only considered delivered when they have been stored and can be dispatched to the customer. People who work in between have similarly divergent definitions. Of course, this also applies to the order date, not to mention which items are even taken into account in an aggregation, such as the average delivery time per supplier.
Does this sound familiar to you? One big step is already to acknowledge that metric definitions require more than catchy names. Also, subtle differences hold significant implications for daily operations. Embrace those variants and regard them as first-class metrics under the ownership of those who utilize them. A better example of the seemingly straightforward metric above would be the following:
Metric: Dispatch to Delivery Point Duration
Owner: Receiving Department
Application: Assignment of intermediate storage capacity and personnel
Definition: The time between the dispatch of an ordered consignment and its arrival at the delivery point. A consignment is considered dispatched 1 hour after the transmission of a dispatch advice. A consignment is deemed to have been delivered with the first incoming goods scan of an item belonging to it.
Defining metrics for your business cannot be outsourced. At PCG we offer guidance and tools for the establishment of well-defined metrics, but it's essential that the individuals working with those metrics take on the primary responsibility for their management.
That being said, there are tools that can help in the development of meaningful metrics. One of those is dbt. Especially its concept of a “semantic layer” which allows data engineers to abstract away purely technical details and provide an unobscured view on your business metrics and their meaning. If you want to know more about it, we recommend this blog on how to Build, centralize, and deliver consistent metrics with the dbt Semantic Layer
.
Well-defined metrics unlock a deep understanding of your business, offering a clear picture of both past and present scenarios. And once you have that, you are ready to predict the future.