In our second blog article on Predictive Analytics, we will start to examine what it takes to create a clean, unbiased, statistical approach to predicting 3 specific challenges that affect virtually every contact center:
- Optimal agent performance patterns
- Agent absenteeism
- Agent attrition
Data Science is essentially about identifying and creating the highest-quality data set, then searching for patterns. The objective should be to help business users make important data-driven decisions, prove or disprove educated guesses with unbiased data, predict the future, and optimize outcomes.
A predictive analytics model aims at solving a business problem or accomplishing a desired business outcome. Simple enough. However, researching and agreeing upon the desired outcomes for a data analysis project, or rather, the lack thereof, is a huge pitfall in the entire process.
One of the reasons is the dilution of the meaning of a desired outcome - which is not to be confused with the accuracy of the data model.
Hypothetically you can build an accurate model to solve an imaginary business problem — but it’s a whole other task to build a model that contributes to attaining business goals in the real world.
Defining the problem or the business question you want your model to solve is a vital first step in this process. A relevant and realistic definition of the problem will ensure that if you’re successful in your endeavor of building this model and once it is used, it will add value to your business.
In addition to defining the business objectives and the overall vision for your predictive analytics model, you need to define the scope of the overall project. In the context of workforce optimization for a contact center, the following examples would be representative of typical challenges:
- Determining which employees are the best fit for specific scheduling requirements (time of day, tasks on hand, subject matter).
- Nail down the factors that adversely affect employee performance (skill matching, work environment, team culture).
- ...or perhaps validate a long-held assumption - is low employee engagement causing high attrition rates in your contact center?
The 3rd objective is challenging, since the vast majority of business leaders hold their own judgement in high regard. Illusory superiority is a blocking issue when it comes to eliminating cognitive bias in a data-driven project.
In the case of #2, where organizations have completed an analytics project to determine the root cause for a general lack of motivation, the results have ranged from simply communicating more effectively from the top-down - for instance, adding a quarterly all-hands to effectively communicate and reinforce the corporate missions and objectives - to providing individual contributors with raises across the board. Think about it: if a 2-3% increase in payroll can reduce a 15% attrition rate by, say, 6%, with an associated savings in HR overhead that far surpasses the remuneration adjustments, a positive impact will be felt at both the employee and corporate levels.
Finally, once your model is developed, at least 5 questions come to mind on next steps for deployment:
- What are the consequences of predicting the wrong solution?
- What is the cost of a false positive?
- How will the model be deployed?
- Who is going to use the model?
- How will the output of the model be represented?
Our next blog article on predictive analytics will continue to focus on deployment, and we will also introduce "inputs" and "validation".