Ken Goldberg, Managing Director, Modus Management
When it comes to managing a fleet of vehicles with some plant and equipment thrown in, large amounts of data are generated. With all the zeal around analytics and visualisation tools, you’d be right in thinking that a treasure-trove of information lies hidden within all this data. Uncover it and you’ll be able to gain insights into key business areas like asset utilisation, financial, behavioural and environmental concerns to name a few. However, attempting to analyse data and make key strategic decisions without having a core diagnostic process means fleet management strategy may largely be inadequate. Simply put, data does not equal strategy!
Organisations should have fleet strategy processes in place or work towards improving these processes. Willie Pietersen, Professor of the Practice of Management at the Columbia University Graduate School of Business, says that strategy processes should be geared towards leveraging key insights to make critical choices about an organisation’s limited resources in order to outperform competitive alternatives. In relation to fleet, many organisations do a reasonable job using data to tactically manage their fleet. However, the reality is that stores of data can simply be overwhelming leading to minimal to no strategic insights and actions.
So, what’s the answer? The key is to identify and understand the problem first and only then do you get to work on the data.
This notion that data will uncover hidden gems is prevalent but as the world-renowned neuroscientist Beau Lotto states, data is meaningless without context because it doesn’t tell you what to do. Apparently, our brains are tuned for understanding relationships rather than processing absolutes. In fleet, this means that it is not about getting loads of data for example from your financial system, fuel provider or FMO. Rather, it’s more about considering relationships. Some examples that spring to mind include the relationship of refinancing and actual asset value used, fuel usage and driver behaviour, or servicing performance and technical skills training.
Once the problems have been identified then it is time to find or collect the right data to help answer the question(s). This step can be tricky because data not only needs to be collected from the appropriate data source but then needs to be properly prepared so no values are omitted or are incorrect. Otherwise, results may not be accurately interpreted or trusted by decision-makers. So, the right data and data quality are of major importance to properly analyse data and harness insights.
Once the data is prepared then it is time to do the analysis. Of course, most fleet-centric organisations already use tools like Excel or Google Sheets. Also, many use older more traditional or legacy “purpose-built” ERP-type fleet applications to help manage operations. But, how many fleet managers check the efficacy of data from these older applications? The NSW Government published an article on how systems can quickly become overloaded with vast volumes of problematic data leading to ineffectiveness. This is especially significant when data becomes orphaned or duplicated through data migrations, faulty code releases or a plethora of other database issues. These types of issues can severely affect analysis so data needs to be periodically and systematically cleansed or purged.
Besides legacy systems, there are a range of modern-day tools that may be more suited for data analysis. A few examples include Tableau, Plotly, Highcharts, and NodeBox. These types of tools can be used with relational databases or for datasets that enter the realm of Big Data (more on that in a moment). That said, there are no faultless tools or applications - just solutions for people with certain goals and mindsets. It should also be considered that every tool and application will compel an organisation and/or a fleet manager down a pathway due to underlying design, functional and servicing limitations.
A quick word on Big Data. The term Big Data has become ubiquitous in our daily lives and is now overused as a catch-all phrase for data management. So much so, that it seems to have lost the original meaning that was first coined by numerous individuals in Computer/Data Science and Econometrics back in the late 1990’s and early 2000’s.
Big Data is a description of the quantity and even quality of a dataset that goes beyond the capacity of traditional relational databases and associated computer server infrastructure. Francis Diebold from University of Pennsylvania suggested that Big Data solutions tend to exceed file sizes of 200 GB. Of course, this may vary by organisation depending upon its resources and capabilities. However, typical near-real-time fleet-based data (excluding live GPS data) from sources such as finance systems, fuel, servicing etc. are typically still measured in tens to hundreds of megabytes rather than gigabytes. The good news is that relational databases are perfectly adequate to manage these dataset sizes.
The real challenge is when datasets become much larger. An IBM whitepaper shows that vehicles can generate up to 1.3 gigabytes of data per hour from sensors such as GPS devices. This means an organisation with 500 vehicles will generate about 5.2 terabytes of data per day and approximately 1.9 petabytes per year. That’s a lot of vehicle and behavioural data.
The bottom-line is that organisations and fleet managers should have diagnostic processes put in place to align with fleet strategy. These processes should help define questions about business problems that can be answered with data. Once problems are identified then the right data needs to be acquired, understood and prepared for analysis. Only then will trusted analysis provide the greatest chance of forming key actions based on trends and strategic insights.
This article first appeared in IPWEA e-news at https://www.ipwea.org/blogs/intouch/2017/11/05/data-does-not-equal-strategy.