content online focuses on how it might be applied in Product or Marketing — the 2 most typical fields where data scientists create great value. Nevertheless, working at a startup, I’ve needed to work with many more functions outside these two. Data exists across the corporate, and the fact is that each department can profit from data science and Analytics to enhance efficiency and drive business value. In this text, I’m going to debate certainly one of those less-covered topics — data science for the Customer Support (CX) team.
I remember the primary time I used to be pulled into a gathering with the CX team, I used to be completely clueless. I didn’t know what to anticipate or how data could actually help them. But now I actually have worked with the team for over three years as their Data Science partner, from the early days after we barely had any data reporting to now, after we are deeply embedded within the function and support data-driven decisions. Within the sections below, let me undergo the common data science use cases in CX.
1. Metrics Tracking
Before you’ll be able to improve anything, you may have to measure it — and CX isn’t any exception. Constructing metrics can also be a superb strategy to establish trust together with your stakeholders.
For CX specifically, some common metrics include:
- SLA (Service Level Agreement): That is the commitment or goal for a way quickly the client support team responds to customer contacts. For instance, “reply to all chats inside 3 minutes.” It’s critical to observe whether the team all the time complies with the SLA. It is often measured as the share of support interactions that meet this goal.
- TTR (Time to Resolution): SLA cares about whether each interaction was done in a timely manner, while TTR measures the whole time it takes to resolve a support ticket — including all of the forwards and backwards. Imagine you, as a user, reached out to customer support via email for a product query. They responded quickly each time you messaged them, but not one of the replies actually solved the query. On this case, SLA would look good, but TTR could be long. That’s why we’d like each to finish the story.
- FCR (First Contact Resolution): Ideally, the client shall be supplied with what exactly they’re in search of within the very first conversation. Subsequently, FCR is designed to measure the share of support tickets which can be resolved while not having follow-ups. Naturally, a low FCR is correlated with a high TTR.
- CSAT (Customer Satisfaction Rating): The above metrics are all internal measures of how quickly we get back to our customers and solve the problems, while CSAT is a direct external measure of how satisfied customers are with the support they received. It is commonly captured via a survey after a support ticket is resolved, with an issue like “How satisfied were you with the support you received?” (rating 1 to five).
- Contact Rate: We care concerning the quality of the service, nevertheless it is equally vital to know what number of support cases are generated. An amazing strategy to normalize the case volume is to calculate the Contact Rate because the
variety of cases / variety of energetic customers. This tells us how often customers encounter issues and wish help, so it is usually a measure of product friction.
In fact, there are various more metrics we have now built for the CX team, however the above metrics should provide you with a superb first glimpse into what data matters to the CX team. They, in fact, are organized and presented in dashboards so the team can monitor the performance and dive into certain case types, teams, or customer segments. At my company, the information team also co-hosts a weekly metrics review meeting to identify trends, surface insights, and drive discussions.
Now that we have now all these metrics, how we could utilize them to drive changes? That’s where the actual power of knowledge science is available in. See the next use cases.
2. Workforce Management
Each customer support interaction ends in labor costs in addition to technology costs, overhead costs, and other operational costs that include it. Subsequently, it’s critical to accurately monitor capability and forecast future support demand for staffing and planning.
The info team can provide a number of value here:
- Forecasting contact volume: It is a complex but high-impact task. It first requires cross-functional collaboration to get the fitting assumption of customer growth projections and adjust the contact rate expectation given product launches and enhancements. Then, data scientists can utilize data toolkits like time series models to bake in all of the assumptions and predict the support case volume.
- Capability planning: Once we get a superb prediction of contact volume, the following query is what number of support agents we’ll need to take care of a superb level of service. This requires scenario simulation of agent performance and availability, and optimization of the agent shift schedules to make sure we meet SLAs without overstaffing.
3. Process Improvements
Data is just not only helpful to trace the team performance, but it might also drive real process improvements. Just to provide you just a few examples that I actually have seen:
- TTR evaluation: TTR is only a random large number without making sense of it. The info team can analyze TTR to discover drivers of long resolution time and use that to tell process improvements. For instance, if the onboarding-related cases often take an extended time with many back-and-forths, this might imply that the CX team needs more training regarding the present onboarding process, or the onboarding flow is over-complicated, so customers continually find it confusing. If the cases coming from email often have a protracted time to resolution with a low CSAT, possibly we should always allocate more resources to reply the e-mail queue to hurry up the responses, or provide higher tooling support to assist agents draft their emails.
- Support tiering strategy: Not all customers are of equal value to a business. Subsequently, a typical practice is to create support tiers amongst customers and prioritize the contacts from top-tier customers. The info team may also help provide you with the tiering system based on customer value and monitor the effectiveness over time.
- A/B testing of support flow: Where should we put the live chat button? make the support center more discoverable for purchasers? Is a certain auto-reply email format higher than one other? A/B testing method helps us answer these support flow design questions.
- Self-service enhancements: The perfect world of customer support isn’t any human support needed 🙂 Though this is sort of unimaginable to succeed in, the information team may also help to catch up with. For instance, we checked out what type of questions users didn’t resolve via the assistance center. This informs what latest topics must be added to the assistance articles and the way the assistance center search function must be improved.
- Chatbot improvements: Chatbot is a typical tool to reply customers’ questions without routing to real agents. Especially on this AI era, we have now seen significant improvements in chatbot quality and availability. Our data team has played a critical role in two rounds of chatbot vendor evaluation with the CX team — organising the information pipeline, A/B testing of various chatbot options, evaluating chatbot performance, identifying the low-performing contact categories, and helping fine-tune the bots to realize a greater chatbot containment rate.
4. Customer Feedback Evaluation
Last but not least, support contacts generate a fantastic amount of text data — they arrive directly from the shoppers and could be used to know customer pain points and product gaps.
- Case categorization: Support cases could be categorized manually by the CX team or with a rule-based framework, but the information team may also help to automate this step, especially with AI’s power today. With easy prompt engineering, most LLMs today can categorize each case based in your product context with decent accuracy.
- Text evaluation: Except from categorization, AI can take the entire case transcripts to summarize and discover the client pain points. My team collaborated with the engineers to construct an internal AI product called “Voice of the Customers” that processes all case details through LLM and surfaces essentially the most common customer complaints in each product area. It is a perfect opportunity to bring CX insights to the entire company and shut the feedback loop with product and marketing. We’ve got seen it getting used actively in product roadmapping.
Working with the CX team has been an unexpected but rewarding a part of my data science journey. From tracking team performance, supporting capability planning, to optimizing internal processes, and improving customer experiences, data science can really transform how the client support team operates.
