Introducing oCX, the New, AI-Generated CX Metric


Whenever a company is seeking to evaluate the quality of its customer experience, nothing is more important than the opinions of the customers who actually have the experience. This is why virtually all metrics designed to measure CX quality today are derived from surveys administered directly to customers.

But oCX, or “Observational Experience,” is a metric that accurately gauges the quality of a company’s customer experience without using a survey at all. oCX is based on an AI-powered technology that identifies and analyzes the judgments of individual customers about the customer experience, as these judgments are made known through the unsolicited text comments made to friends and others in social media, or on review sites, or via other online mechanisms. In essence, oCX is capable of accurately measuring the quality of a company’s customer experience based on simply observing the opinions rendered by its customers “in the wild,” so to speak.   

Using AI tools to decode the sentiments and emotions behind written-text comments, oCX predicts the response that each observed customer would likely have provided had they been asked to rate their own “likelihood of recommending to a friend,” on a scale of 0 to 10, as shown in the “Sample Reviews” box, below. With these individual ratings, oCX then computes an NPS-like score that has been shown to be a very accurate representation of the actual score the business would have achieved with a survey, if it had fielded one. 


Turn unstructured feedback into action with oCX 


Since oCX is not based on structured, quantitative survey data but instead relies on interpreting unstructured data represented by written text, the oCX metric can provide a richer and more authentic reflection of the actual customer experience a company delivers. And while unstructured data already represents 80% to 90% of the data available today, it is growing at an astounding rate, in excess of 50% annually, so oCX will inevitably become an indispensible assetwhen it comes to extracting genuine insight from this sort of unstructured customer experience data. [i]

Moreover, because it is derived from such unstructured data, oCX may also present a clearer way to understand Gen Z consumersthe dialoguers,” according to a recent McKinsey report, who not only make their decisions based on what others think and say on Instagram, Youtube and Tiktok, but also have more fluid expectations of the companies they deal with on a daily basis.[ii]

“When a customer gives you feedback (good and bad), it’s a gift. Finding a way to analyze and measure that feedback is crucial to understanding how customers react to your products and services. The new oCX (Observational Customer Experience) metric is intriguing. This is the type of metric that can help you operationalize the feedback and reviews to create a better customer experience for your future customers.”

Shep Hyken

Customer Experience Expert and New York Times & Wall Street Journal Bestselling Author

How is oCX calculated?

Our team of data scientists has devised a sophisticated approach that assigns a numerical score ranging from 0 to 10 to each written customer review. This method leverages advanced techniques such as Gaussian Mixture Model and sentiment analysis, seamlessly harmonizing text-based feedback with the corresponding Observational Experience (oCX) score. This methodology distills the unique essence of each customer’s perspective into a single, meaningful number. Through thorough analysis of a substantial volume of comments, we have achieved a remarkable level of confidence in consistently identifying this alignment. By differentiating scores from 0 to 6 and those from 9 to 10—reminiscent of NPS computations—we extract an NPS-like overall score. This methodology empowers us to genuinely observe the refined quality of customer experiences within their natural contexts, unbound by the constraints of conventional surveys.

Learn your oCX Score

 Currently oCX reports are available for brands in online food delivery and online marketplaces for download.  We are adding new verticals every month, feel free to reach out to us to learn about our new reports.

If you are analyzing different forms of text based data such as surveys, reviews, social media, tickets and forums and you would like to track experience of your users across journeys for different products and services, oCX can help you automate and get in track in days.  To learn more about oCX and inquire about your brand’s oCX score, contact us here.

Related Resources

Consumer Perspectives: Best Fintech Apps


  • Gain insights on global fintech user sentiments through our analysis of over 500,000 reviews in 2023
  • Discover front runners in customer experience and what it takes to be a top fintech brand
  • Discover Observational Customer Experience (oCX) score, transforming written feedback into a standardized numerical rating

Introducing oCX, the New, AI-Generated CX Metric


  • oCX, or “Observational Experience,” is a unique metric that assesses the quality of a company’s customer experience without relying on traditional surveys.
  • This metric is powered by AI technology that analyzes individual customer opinions expressed in unsolicited text comments on social media, review sites, and other online platforms.

Etsy vs Amazon: An Analysis of Holiday Shopping 2023


Dive into the guide for your holiday shopping ! Discover the game-changing trends in consumer spending, unravel the mysteries behind Amazon’s global dominance, and explore Etsy’s unique charm in supporting small businesses.Whether you’re seeking the latest trends or looking to support local creators, this guide promises to enrich your holiday shopping experience.

Request a Demo

Seeing is believing

Learn how to apply AI to analyze all the various CX ”signals” generated by your customers via surveys, text, complaints, social media, and other interactions. Discover how to quickly identify and flag the most important problems and opportunities, and then better prioritize your investments.