Research Economics and Management

With our research we focus on the themes of labor markets, aging, sustainability, innovation, market governance, banking, financial markets, entrepreneurship, marketing, and consumer behavior.

Smart Calling thanks to Data

You’re in the middle of dinner, hard at work, just taking your kids to bed, or working up a sweat in the gym, and then your phone rings. It’s somebody calling you about your health insurance, an internet provider or energy supplier trying to get you to switch over to them. You’re thinking ‘Not now!’, and you try to end the conversation as quickly as possible. Recognize it? Tilburg University alumnus Olcay Yucekaya has come up with a solution.

Olcay Yucekaya wrote his Master’s thesis on the subject how companies can pick more agreeable times to call (prospective) clients or customers, resulting in better service, better customer satisfaction and thus increased customer loyalty, and ultimately better sales figures for the company.

Olcay Yucekaya

The Time Factor 

The alumnus worked for years as an employee of the Den Bosch branch of the 2Contact call center, and as such was thoroughly familiar with the frustration he and his colleagues experienced when they called customers at times that were inconvenient to them. “It’s not much fun calling people if the last thing they want to do is talk to you. They hang up or get unfriendly. Besides that, it is counterproductive. I wanted to do something about that. I put it before Project Manager Pieter Prins and Operations Manager Joke Lammers-De Weers of 2Contact to write my thesis on the problem. Associate Professor George Knox helped me refine my ideas. He suggested using customer information to make calls more productive. That is what I started with. The next question was: What information am I going to use exactly? I tried to imagine what might be important factors. How to figure out the best moment to call customers? That is where I got the idea to subdivide workdays into half-hour slots.”

Characteristics and variables

He subsequently investigated if it was possible to categorize customers into groups on the basis of characteristics and variables such as age, family composition, and residence. “I incorporated data in a statistical model. Within this model, it is possible to apply a hierarchy, increasingly narrowing down the variable - of a person’s residence for instance: by province, by city, by postal code. That way you can compile very specific lists.” In his thesis, Yucekaya found that calls could be made more productive using this theory. “If you call at times most convenient to the customer, they are much more likely to want to talk to you, making the conversation much more pleasant. It makes customers more inclined to purchase things or to agree to payment arrangements.” This is borne out by the figures in his research.
Telephone numbers are automatically dialed a number of times, after which, if there is no answer, the call is ended. Normally, the percentage of numbers dialed that yield an answered call tends to be somewhere between 60 and 70%. In Yucekaya’s model, this was 93%. Besides this, call center employees turned out to have fewer but longer telephone conversations per hour, resulting in 31% more purchases. In his Master’s thesis, Olcay Yucekaya presented these results and more.

Call in the morning

Theory says to call customers is in the morning; the reality turns out to be somewhat different.  “Most people may want to be called in the evening around eight p.m., but most call center employees work in the morning. Besides this, some companies do not want you to call in the evening.  Also, my idea was to have a subdivision into half-hour slots while this was impossible in the system we were using. A subdivision into hours was possible.“ Yucekaya refused to give up. “The reason I picked this subject for my thesis in the first place was that I wanted to put my findings into practice. Why do a research project if you are going to shelve it afterwards? Other call centers make different choices if operating results are disappointing. They will have the dialer – an automated phone - call more customers for instance. That will result in more phone calls and more customers getting irritated, including phone calls where there may not be an employee available to take the call, so that the customer will hang up: an abandoned call. I think it makes much more sense to increase the efficiency of your calls.” It took quite a bit of time to work things out together and apply the research, but now framework Smart-Call™ is a fact.

Keep on learning

You might think that by now all call centers would have switched to Yucekaya’s new data-driven calling model, but nothing could be further from the truth. 2Contact is still the only call center using Smart-Call™. Meanwhile the passionate data scientist is working on improving his framework. “I have largely automated my own activities. Collecting and processing data, and generating lists now takes place virtually automatically. This leaves me more time to focus on the bigger picture. How can the model and the process be further improved and refined? Would it be possible, for instance, to have a computer call people and only have the call put through to somebody on the floor if the customer has indicated that they are willing to take the call. That would save a lot of time and frustration. I am also currently training others. During my time at the university I learned how to keep on developing myself, to keep on learning. I try to put that in to practice every day. I now want to explore genuine artificial intelligence. If there are students or researchers at the university currently working on this subject, I would like to invite them to please contact me.”

See also: Analytics Magazine INFORMS

More information: Olcay Yucekaya en George Knox

Linkedin Olcay Yucekaya


Author: Melinde Bussemaker