For a few years, CX research tech is booming, empowered by emerging tech. Despite the exploding variety of tech, the state of execution is in a mixed state. Companies are looking for orientation.
What are the benchmark implementation top companies are deploying? Is there any vendor-independent guidance?
CX.AI teamed up with University Ostfalia and TMRE to give answersand to track progress year by year with this study called “The statse of CX analytics”.
We interviewed 155 enterprise insights professionals from around the world qualitatively and quantitatively. These are the five biggest learnings
#1 Getting leadership but-in is the biggest pain
The biggest pain of CX insights teams is not finding the fanciest tech in town. Its not to quantify the customer journey, to improve response rate or to drive impact. The biggest challenge is to gain the internal buy-in of stakeholders.
One reason is that researchers do not professionally manage the process of getting buy-in. 40% are doing nothing systematically in this regard. Just 27% have a marketing plan for insights in place.
#2 NPS is king. Stability can be an issue
There are many different outcome metrics in CX research, but the bulk of companies leverage NPS. Some participants complained about the instability of the NPS score, but there are a few leverage available modeling methods to fix that. 17% are utilizing multidimensional rating as a means to get a stable outcome measure.
#3 Surveying still has room for improvement
One-third are still using CATI (phone) to gather feedback. 23% of companies are already using advanced tech to get better feedback, such as active listening, audio or video feedback.
The biggest room for improvement: 40% are still asking different questions to promoters vs. distractors to get better feedback. This is blocking the way to find drivers through modeling properly.
#4 Text analytics on the rise – but not satisfactory yet
36% still do not systematically analyze their customer feedback, apart from eventually reading them. Most text analytics is “unsupervised”, leaving room for improvement. Although 47% are already leveraging text analytics, the majority is not satisfied with granularity or accuracy.
#5 Driver analysis and impact evaluation is evolving
57% are using no or not appropriate methods to evaluate what’s important. Just 30% running drivers on text analytics data. That’s a method that combines short-surveys, with qualitative insights and predictive and prescriptive insights.
Only 40% perform a proper analysis to explain the impact of CX on the bottom line. Just 13% use causal machine learning – a tech proven to outperform conventional modeling by factor two.
Conclusion
The State of CX Analytics is advancing. It is evolving in measurement, surveying methods, text analytics, and proper interpretation with modeling. However, the biggest “construction site” is still the art and science of getting buy-in. This is what insight leaders need to work on:
- Set up a marketing plan. “Insights” is a product like any other service. It remains a shelf warmer unless adequately marketed. It starts with researching the target audience’s interests.
- Do your homework. Use tech to get your scores stable and eliminate the risk of questioning the outcome itself.
- Simplify with tech. If your tech can read, categorize, and understand by finding what predictably leads to outcomes, you can simplify to focusing on the supercritical topics.
- Align with interests. Understanding is biased by interests and beliefs. Use available analytics tech to prove impact to every particular interest. Prove the impact of CX on churn, revenue (for CFO), word-of-mouth (for CMO), and operational costs (for COO).
- From reporting to gamified planning: Leadership needs planning help not just insights.The more insights can be experienced e.g., by interactively simulating impact, the better it is understood.
The report is provided by cx-ai.com and can be obtained by writing an email to Frank@cx-ai.com
p.s. this in-depth online course is free for enterprise insights professionals and discusses many of the points of the conclusion https://www.cx-ai.com/cx-analytics-masters
