INTERVIEW – HOW TO PREVENT CUSTOMER CHURN USING PREDICTIVE ANALYTICS – CRISTINA CRISAN TRAN
By NextLevel Pharma - July 16, 2018

Leading up to NextLevel Life Science’s 6th Edition MedTech Commercial Leaders Forum 2018, we are conducting interviews with selected members of our prestigious speaker panel to learn more about their thoughts on this vital issue.

*Opinions below are those only of the individual and do not reflect upon corporate strategy or positioning.

For more information regarding NextLevel Life Science’s 6th Edition MedTech Commercial Leaders Forum 2018click here!

Cristina Crisan Tran, Vice President Global Customer Marketing & Education, Straumann

NextLevel: How would you describe the work that you do at Straumann?

CCT: Straumann is a global leader in implant and restorative dentistry as well as oral tissue regeneration. I am Vice President for Global Customer Marketing and Education, which comprises market research, medical marketing, medical advisory, training and education as well as strategic solution management. The latter includes indication-based solutions, e.g. for the edentulous patient, and value-added services to increase the efficiency and competitiveness of the dental practice (e.g. loyalty programs and patient communication tools).

Prior to Straumann, I used to work for Sonova, a global market leader in hearing aid solutions. I spent almost 10 years at the company holding different senior leadership roles in international sales and marketing. My lecture is based on a concrete situation I faced and had to master in my role as Director of Global Field Marketing. At that time, we suffered from customer attrition. It took us a while in HQ to realize that existing loyal customers were churning, because the company was growing nicely. However, this was mainly thanks to the acquisition of new customers.  Therefore, we needed to understand why it was happening and how to stop that trend. That is when I worked with predictive analytics methods to figure out why customers were churning and what kind of behaviours or attributes defined a customer’s churn probability. At the end of the initiative, we were able to define churn scores for each customer that we included into our CRM system. Afterwards, we defined a program together with the local sales organizations to prevent customer churn.

NextLevel: Can you talk a little bit more about your experience in using predictive analytics with big data?

CT: Yes, like many companies we had a lot of data in our ERP and CRM systems. Usually, the challenge is that companies do not know how to use the wealth of data they possess. It helps when you know what you are looking for. Predictive analytics is all about using existing data to identify patterns and behaviours that allow you to predict future trends.

In order to address the customer churn, we initially interviewed customers who substantially dropped sales to better understand their reasons and get a couple of insights to look for in the data. Afterwards, we looked at the patterns of those customers that churned.

Some data points we looked at were the number of interactions with the sales rep, what products they bought, if they were participating at our roadshows, marketing  and educational events, where they were located, if their sales rep had changed in the last year, if they had any complaints logged at customer service, etc.  We also considered potential weather-related sales drops in the respective country or area. For example, in some US cities during winter and heavy snowfall, sales would drop dramatically since elderly patients would not go out to see their hearing care professional.

Based on these insights, we built a predictive model with churn scores (probability to churn). We then addressed each of the customers at risk according to their ranking by value for the company and churn score. What is amazing is that in the end very often customers just needed a little bit more attention.

There was also the issue of the “silent customer.” We tend to solve the problems of the customers that are complaining, but these are not the ones that will drop first. They are complaining to give you a chance to do it right. Around 90% of customers that move away, and there are studies for this, are doing this silently without any prior complaints. We also considered this insight when we built our predictive model.

NextLevel: How did you define an approach internally toward the customers’ churn scores?

CT: We did these analytics in the bigger markets, for example, US, Germany, and France, because they had the most data available. It is important when you want to use predictive analytics that you have sufficient data in your CRM and ERP systems.

After one identifies the customers who are likely to churn, a specific approach has to be defined including the entire customer-facing team, but especially the sales reps. Further, it is important to know that a churn score is just a probability. You could compare it with playing the lottery. You have a probability to win, but it does not mean that you are going to be the lucky one. That is the same with the churn scores. There is a probability of a customer to churn, but it does not mean it will automatically happen just because the churn score is high.

Therefore, we were trying to give these customers more attention, involve them more, and ultimately change the parameters that were influencing the likelihood to churn. There were many commonalities across borders; however, the program that we had in place to address churn was adjusted to the local market needs.

NextLevel: What type of other insights did you get from the predictive analytics?

CT: Many insights came out beyond churn management and churn scores. For example, the number of visits needed to keep a customer happy was especially important. The medical devices space often has direct sales, which is very expensive and consequently requires you to allocate the resources wisely. We figured out that to keep a customer it was sufficient to visit them 7 times per year, but to win over a new customer from competition it was necessary to visit them an average of 12 times per year.

We used the insights from this data to restructure our go-to-market as well as the target and visit planning of the sales force. We could also decide better on the most important triggers to pull in order to attract new customers. There were many positive side effects of doing this exercise with predictive analytics. The insights that we won went way beyond churn management.

NextLevel: What are you looking forward to most at our event in London?

CT: I would like to learn what trends are emerging, what other companies are doing, and what could be applicable for my area. Further, I am very much looking forward to meeting people and networking with them.

For more information about this MedTech Commercial Leaders Forum please click here!