Key Highlights
•
A new AI framework turns customer churn risk scores into specific, cost-effective actions a company can take to retain each individual customer. This reveals a major gap: a highly accurate churn prediction model is useless if the factors it uses (like customer behavior) can’t actually be changed by the business.
Source →
•
The study found that it’s much easier to find actionable ways to retain customers in financial services (like banking) than in digital services (like streaming). This is because financial services have clear levers like interest rates or fees, while digital services often rely on hard-to-change user engagement habits.
Source →
•
For analyzing bilingual speech where people switch languages, the standard unit of “a word” is flawed because switches almost never happen in the middle of a spoken phrase. A more accurate unit is the “intonation unit,” which is a natural chunk of speech marked by pauses and pitch changes.
Source →
•
Using words to measure code-switching probability inflates the number of potential switch points and makes the results less precise. By switching to intonation units, researchers can get a much clearer and more accurate picture of how and where people actually mix languages in conversation.
Source →
•
A study in Estonia found that when people have cybersecurity problems at home, they mostly ask friends and family for help, not professionals. This informal help is often slow and can be inaccurate, creating a security gap despite Estonia’s advanced digital infrastructure.
Source →
•
The research highlights a need for official cybersecurity support services for the public, as well as better education to help people solve common problems themselves. It also found that women are more likely to depend on others for cybersecurity help, while men are more likely to expect the advice they get to be wrong.
Source →
Stay curious. Stay informed — with
Science Briefing.
Always double check the original article for accuracy.
