Predictive analysis utilizes many different statistical techniques, examining historical data in an effort to project future behavior. While past behavior may not be 100% indicative of future behavior -- think of financial market analysis and how hard it is to pick the next great stock performer -- it can help contact centers work more smoothly and efficiently.
In this case, predictive analytics techniques can be wonderfully useful. However, it's essential to know there's a balance between "science" and "art" that each organization must discover to make its predictive analytics deployments successful.
Historically, contact centers have been using a form of predictive analytics in the area of workforce management for many years. Workforce management analyzes relationships between key pieces of historic data to project future staffing requirements.
With the increase in computing power, the retention of more data and enhanced statistical techniques, predictive analysis can now be used during specific "moments of truth" in the contact center.
Workforce management: The old standby
Predictive analysis is the fundamental concept behind workforce management systems that project future demand of resources by analyzing historical data. If a contact center does not have the staff in place to handle customer inquiries in a timely manner, overall customer satisfaction decreases -- and that, in turn, drives down loyalty and customer retention.
The science. Workforce management looks at historical data regarding volume of calls, average handle time and other factors to project future call volume and gauge staffing requirements.
The art. Historical data may not be representative of the future. For example, in a certain year, the Fourth of July may fall on a Monday, and the following year, it may fall on a Tuesday. When looking at each year, either additional assumptions must be added to the model prior to engaging predictive analytics techniques, or the business must utilize judgment to make the appropriate adjustments to the outputs of the models once the forecast is developed.
Upselling: Getting more from each engagement
Customers often lack knowledge of additional relevant products or services that may be available from a company. Analytics can put the right suggestions in front of employees at the right time.
The science. By analyzing historical information, predictive analysis may identify situations in which customers who have used a specific product return to purchase a complimentary product. If the current caller has a pattern of behavior that is similar to previous customers, there's a greater likelihood that caller may be interested in purchasing the same complimentary product. The current interaction provides an opportunity to proactively educate the caller of a potential need.
The art. Upselling is tricky in the contact center. Prior to embarking on an upsell opportunity, the contact center agent must resolve the customer issue to the caller's satisfaction. Once resolution of the issue is complete, only then should the agent pursue an upsell opportunity if it makes sense. It takes training and practice to use the cues that analytics can provide.
Customer retention: Building loyalty
On many occasions, a change in behavior -- such as buying less often -- or a significant event -- such as approaching the end of a contract -- can indicate that a customer may discontinue doing business with an organization. Retention efforts often occur too late in a relationship with a customer. An interaction with the contact center may provide an opportunity to engage the customer before it's too late -- and help build upon an ongoing relationship.
The science. By analyzing historical information, predictive analytics techniques may identify patterns in which customers with similar behaviors have left an organization or a significant event is approaching. If a customer is approaching the end of a contract, a contact center agent can broach the subject with the customer and attempt to extend the contract.
The art. For effective customer retention, judgment must be used to ensure that the relationship between the customer and organization is still valued by both entities. If a customer has reduced or stopped using a specific service as a result of moving out of the service area, a company must allow that customer to leave without hassle.
Fraud intervention: Stopping loss sooner
Fraud can cost organizations a tremendous amount of money -- and once it occurs, it can be very challenging to discover the criminal and recover losses. Predictive analytics techniques can help uncover potentially fraudulent activity by keeping an eye on trends that go against the norm.
The science. By analyzing historical information, predictive analysis may be able to identify spending patterns that are inconsistent with a customer's previous activity. For example, if a customer normally makes one or two purchases in a month for $500 each and all of a sudden that customer makes 10 purchases in a two-day time period, red flags should be raised and a special group of contact center agents should reach out to the customer to confirm the purchases were indeed made by that person.
The art. Don't assume the customer is guilty. The outbound contact should be treated as a courtesy call and the contact should instill confidence in customers that the organization is looking out for them.
Making it work
The idea of predictive analytics use in contact centers has been around for a long time, starting with workforce management. But as the technology evolves and becomes more accurate, predictive analysis is becoming more and more useful for other tasks. As predictive analytics techniques continue to develop, there are many new opportunities to increase their use in the contact center for both inbound and outbound calls.
It's critical to always add the human factor of customization and common sense to implementations of these increasingly powerful systems. It's not the tool that guarantees success. Always keep in mind the necessary balance between the science of predictive analytics and the art of how and when to utilize that tool's capabilities.
How to make better business decisions with predictive analytics tools
The advantages of predictive marketing analytics
The public sector puts predictive analytics to work
Key facts about predictive analytics processes and techniques