Rule-based alerts vs. predictive alerts

Rule-based alerts take advantage of your best ideas of what to track for your customers. You can create a rule that specifies the conditions for when to trigger an alert. You can create multiple rules for a single alert type (e.g. churn), each detailing a set of conditions that need to be met to trigger an alert for your customer. 

Rule based alerts help you track customers based on specific behaviors or criteria you care about or believe are likely to lead to churn or conversion. They can also be leveraged to facilitate the customer lifecycle management by tracking key milestones that your customers need to achieve according to a specific timeframe. 

Limitations of rule-based alerting can be that you don’t know what behavior to look for to generate an alert, so creating rules is, at best, a guess. And you can create a handful of rules to trigger alerts, but 2 or 3 rules means you’re looking at 2 or 3 signals - not a highly accurate way to predict customer behavior.

Rules are also static, which means they don’t improve in accuracy unless you change them - which itself isn’t a guarantee they’ll become more accurate. They also don’t adjust to changes in customer behavior, such as when a product evolves and offers new functionality.

Types of alerts that are supported by the rule-based system in Natero:

  • Churn
  • Conversion
  • Expansion
  • Renewal
  • Engagement

Predictive alerts are based on Natero's state-of-the-art machine learning algorithms, which will proactively alert you to customers that require attention by learning from the vast amount of historical data of your customers.

Different from the rule-based system which triggers alert based on a handful of pre-defined conditions; the machine learning system is true data-driven platform and offers greater accuracy vs. those that rely on intuition or guesses.

A predictive alert will be triggered when Natero's machine learning model identifies certain behaviors or patterns of a customer (e.g. activity level, late payment, usage trend) that are most relevant indicators of churn, conversion or expansion. Normally these factors are not apparent or easily captured in a simple rule.  

Predictive analytics are only useful when enough data is gathered and consumed by Natero. As more data comes in, Natero’s machine learning models automatically adjust for changes in customer behavior and become more accurate in their ability to predict customer actions, with no effort required from you. 

Types of alerts that are supported by the machine learning system in Natero:

  • Churn
  • Conversion
  • Expansion 

 

Read our blog article on when to use rule-based alerts vs. predictive alerts.

Have more questions? Submit a request

Comments