Your company must decide on which revenue streams to scale up or down to hit a major goal in the future. You may not have a team ready with the scientific and data acrobatic chops needed to find a solution. You turn to your contacts for expertise and direction, and a week later, you and your newfound team present a promising glimmer. It would be nice if your success can be repeated for better workforce planning, but without months of data on each worker’s knowledge and their relationships with others, how can leaders efficiently create successful teams to solve big problems? Social network analysis is a powerful answer.
Collaborative culture for better workforce
Workforce planning is the process of supplying workers and skills to business needs, and social network analysis can be used to understand relationships among workers. Before such analysis can be effective for workforce planning, there needs to be a cultural foundation championed at least by company leadership that fosters knowledge creation and collaboration. Research done in 2002 by the IBM and social network scientist, Stephen Borgatti found that this foundation can be enriched by using social network analysis, which can then be used for workforce planning. Let’s touch a lite example that I call, “Expertise Reallocation”.
Better workforce through visual analysis
Each circle below represents a unique worker, and there is a line segment between two workers if they worked together on a data-related task. It seems that Molly is an expert because she shared tasks with seven workers, which is the highest of everyone. Management may ask Molly about her workload and unsurprisingly find that she thinks it is too heavy. Even if Molly can manage her workload, it’s likely not a good idea to have just one expert because that could render a bottleneck if Molly is unavailable. So, management decides to reallocate some of her workload to another possible expert. Who should they pick and how should they reassign Molly’s shared tasks?
There is a worker named Mike that shared tasks with three workers, which is the second highest in the bunch. So, management may reassign some of Molly’s tasks to Mike, resulting in the visual below. This new plan should yield better managed workloads for all.
Workforce assignment by optimization
This approach can be enhanced and used to create teams by first identifying team leaders like Molly and Mike. One approach is to first formulate an optimization model to identify possible team leaders and create teams with an objective of balancing workload. Technically, such a model could be a mixed integer program. Depending on the size and complexity of this model, the second step is to feed the optimization model into an algorithm to find a well-balanced set of teams. The Analytica Optimizer is one of the best platforms to model and solve such problems. Like other analysts, I think visually, and unlike other tools, Analytica allows you to organize your model into nodes (like boxes and ovals) instead of curtains of code. Talk about easy on the eyes.
The social aspects
Though all this mathematical stuff sounds good, you may be wondering how social network analysis can dig into workforce aspects like trust and influence in your company. In a 2010 paper Industrial engineering researchers from the University of Arizona discuss a powerful way of using social network analysis for workforce planning. Their core approach involves analyzing pairs of workers based on four social dimensions: trustworthiness, influence, reputation, and [geographical] proximity. Coupled with skills information for each worker and you’ll get quite a holistic view of company talent.
Modeling social dimensions
Applied to a large software development organization, the University of Arizona researchers start the process by having each worker evaluate another worker based on the four social dimensions on a scale, say from 1 to 5. Management decides the importance of each social dimension using a weighted scheme. A sum of products of the weights and values produces a compatibility measure for worker j on worker i, that is, a value Cij from 1 to 5 on how compatible worker j is to worker i (from the perspective of worker i). In short, you get this formula along with some conditions on the weights:
The “V”s are the values from 1 to 5 and the “w”s are the importance of each social dimension. This approach is flexible; if management wants to create teams wherein there is greater trust among workers, then greater weight can be placed on the trust weight. Data exploration methods can then be used to categorize workers based on the compatibility measures, which can yield a richer understanding of worker types. Plug this into an optimization model, likely a mixed integer program, decide or craft an appropriate algorithm to use, and you’ve got yourself a systematic process to create teams whose members will work well together.
To see how Analytica brings you the tools for better workforce planning, from analyzing data to creating optimal teams and more, start your free trial.