In our previous blog post, we discussed machine learning and how machine learning models can be applied for Customer Success.
In this excerpt, we discuss how predictive analytics can help Customer Success teams and why a good CSM solution should have both predictive analytics and rule-based alerting capabilities.
How Predictive Analytics Helps Customer Success
Machine learning unlocks the value hidden in the data captured by a CSM solution, and identifies the factors that drive various outcomes. Using machine learning to predict customer behavior has tremendous benefit vs. depending solely on rule-based alerts.
Rule-based alerts rely on the SaaS vendor’s intuition as to what customer behaviors signal the likelihood of churn, expansion, conversion, etc. Many vendors struggle to define these behaviors, but even if they can come up with a handful of rules, that’s only a handful of factors that are being evaluated to predict an outcome.
Machine learning evaluates hundreds to thousands of factors, and identifies which ones are the best predictors for each type of outcome. Whereas rules are based on intuition, machine learning is real data science.
CSM solutions that incorporate machine learning are continuously capturing data, and they update their predictive models based on this new data.
This means that the models get more accurate over time.
Rules never improve in accuracy unless you change them, and changing a rule is no guarantee of increased accuracy. Remember, rules are typically based on intuition so a new rule could be less accurate than the old one.
The dynamic nature of a machine learning model also means that it can automatically adapt to changes in overall customer behavior. Customers may act differently over time as product offerings change, and a client base can evolve or the macroeconomic landscape shifts. Rules are static and assume that the future will be like the past. As time goes by, rules may become even less accurate than when first applied.
Accurate predictions of customer behavior can dramatically improve the productivity and effectiveness of a Customer Success team. Accurate predictions mean there are less “false alarms," which waste time that could be better spent on customers who actually need attention. And missing out on alerts can result in lost business (e.g. no up-sell alert) or lost customers (e.g. no churn alert).
In short, predictive analytics enables Customer Success Managers to spend more time with the customers who need their attention the most.
It's Not Machine Learning vs. Rules... You Need Both
While the benefits of using machine learning to predict customer behavior are significant, a good CSM solution should still allow for rule-based alerts.
After all, not everything is about predictions.
There are many scenarios where CSMs want to monitor specific customer behavior, and receive notifications about certain things that customers do, or don’t do.
For example, vendors can use rules to specify milestones during an onboarding stage, and receive alerts if customers are not achieving those milestones. This gives the Customer Success team a chance to reach out to a customer who may be struggling during a critical and impressionable phase in their lifecycle. Other rule-based alerts may look for too many support tickets or the number of times that customers hit the “help” button.
Advanced CSM solutions incorporate both predictive analytics and rule-based alerting to cover all scenarios. They are more accurate at predicting customer behavior, while monitoring specific scenarios that require the attention of the Customer Success team.
- - - -
In the next installment, we'll talk about the steps involved in implementing machine learning models as well as what to consider when evaluating CSM solutions with predictive analytics.
Want to learn more?
Download our eBook: Predictive Analytics for Customer Success.