While the term algorithm implies complexity, this isn’t necessarily the case, as an algorithm is simply a set of rules used to solve a problem.
Unfortunately, algorithms aren’t always that simple. As the complexity of a supply chain algorithm increases, so does its technical requirements, calling for advanced programming skills, access to sophisticated programming languages and an ability to access and organize data.
This is why this field has been historically dominated by highly qualified data scientists and operation research teams, all of which are experts with years of experience in advanced analytics solutions.
As programming software has progressed from complex machine code through to high-level machine independent languages, it’s become easier to write algorithms. This has, to an extent, reduced the dependence on elite data scientists and led to the emergence of what may be termed citizen data scientists, individuals possessing a combination of supply chain knowledge and analytical skills.
The Age of Big Data
One of the characteristics of the digital age has been the explosion in the volume of structured and unstructured supply chain data. It’s believed the amount of data in circulation by 2025 will be ten times greater than what it was in 2017. Additionally, it’s estimated that as much as 80% of this data will be unstructured, meaning data that isn’t organized in an easily accessible relational database. Being unstructured doesn’t mean this data is of no use, but exactly the opposite. Unstructured data contains massive amounts of potentially useful information and is accessible through sophisticated analytical tools.
A characteristic of data in the digital era is that it’s fast moving with a high velocity. This means you need to access and process this data quickly, while it’s still current, to reap its benefits.
Opportunities Offered by Big Supply Chain Data
Supply chain data that can be accessed and analyzed created tremendous opportunities for optimizing and redefining your supply chain. Big data analysis can identify previously hidden trends and determine how they relate to outcomes. Real-time interpretation of information contained in big data can create strategic and competitive opportunities. Added to that, optimization techniques allow you to identify inefficiencies and improve operations.
A particularly powerful application of advanced analytics, known as prescriptive analytics, involves the creation of an intelligent model of the business. Using this technique, it’s possible to determine trends, predict what will happen and ask what-if questions to identify the right supply chain decisions.
How Analysts Create Solutions to Supply Chain Problems
Historically, the first step was always to identify the supply chain problem and establish what question needs to be answered. It’s essential that key issues are clearly understood and that the analyst understands the organization’s supply chain.
The analyst then determines the type of algorithm needed, usually some form of regression analysis. Factors to consider include the number of independent variables, types of dependent variables and the mathematical formula or curve which best represents the problem. This process is referred to as creating a model representing the problem. There’s also the need to identify the data sources, determine how to extract the data and prepare it for use.
Using a high-level programming language, the analyst will write one or more algorithms to create a mathematical model. Then, using solver software, determine the best solution or answer, taking care to ensure the model accurately reflects the real world and that answers are feasible.
In the right hands, this algorithmic approach is extremely effective. Downsides are the time taken to create and validate models and that users are dependent upon the analyst’s skills for further development and maintenance.
The Need for Professional Data Scientists in Supply Chain Analytics Jobs
This approach requires a high degree of programming and analytical skills. Analysts need advanced knowledge and experience with third- and fourth-generation programming languages such as Python, R, or SAS, as well as how to access, clean and prepare structured and unstructured data. Additionally, they need advanced mathematical skills and a thorough knowledge of the different types of analytics.
It’s not surprising, then, that there’s a significant shortage and that data scientists are in demand. Since 2013, the demand for data scientists, also sometimes referred to as operations research (OR) specialists, has more than tripled. While 2013 may seem like a long time ago, it should be noted that it takes several years to gain the experience and training needed to qualify as a data scientist, and current indications are that the number of job postings far exceeds the number of qualified applicants available.
So, if you are in the market for a data scientist, it may be wise to review your options and consider one of several other alternatives.
How Analytics Platforms Reduce Dependence on Highly Skilled OR Specialists
In an effort to reduce dependence on data scientists, a number of vendors are marketing analytics platforms. Some are packages that offer standalone solutions to specific problems, while others allow users to define and solve problems.
Analytics packages are easiest to deploy because all that’s needed is to configure them to your organization, a process that’s well within the skill set of an IT professional or citizen data scientist. Their disadvantage is that you may have to change the way you work to align your business with how the package works.
This isn’t an issue with business intelligence / analytics platforms that offer powerful analytics capabilities and an ability to define almost any problem. Many incorporate features that streamline data linking as well as user dashboards that simplify input and outputs. While reducing workload to a degree, the majority of these platforms still require some advanced third- and fourth-generation programming knowledge and expertise. While much of this work can be performed by in-house experts, many applications still require the services of a data scientist.
Democratization of Supply Chain Analytics Jobs With 5th Generation Programming Languages
Up to this point, the major changes in analytics jobs revolved around lessening the dependence of the organization on fully-fledged data scientists. But they also helped create opportunities for in-house IT and managerial personnel with the aptitude to work with and understand advanced analytics.
All of this changes when using fifth-generation programming languages. These languages focus on identifying constraints and allow users to create visual, code-free models that accurately represent the real-world supply chain.
Fifth generation programming languages have several distinct advantages over earlier languages, including:
- Code-free coding
- Visual model representation
- Easy to update and expand
- Simple scenario construction
- Built-in dashboards
- Simplified data connection
Although fifth-generation coding is easier, it’s essential to understand there’s still a crucial need for appropriate analytical skills and experience. However, in most instances, this is something that’s well within the capabilities of citizen data scientists.
Future Skill Sets for Supply Chain Analytics Jobs
It’s apparent that the requirements of supply chain analytics jobs are changing. While there will always be scope and a need for professional data scientists, they are no longer as essential, especially when using fifth-generation prescriptive analytics platforms such as Enterprise Optimizer.
At the same time, supply chain professionals should learn and develop analytical skills that allow them to interact with supply chain models.
In this way, it’s possible to put supply chain modeling software where it should be: directly in the hands of supply chain professionals. With the ability to instantly run what-ifs and determine optimal decisions, a fifth-generation prescriptive analytics model opens up numerous possibilities for optimizing supply chain management and reducing supply chain costs.