Utah tech unicorn Qualtrics is jumping on the machine learning bandwagon with a trio of smart features built into its platform under the brand iQ.
Qualtrics builds software which allows brands like Adidas and BMW to collect and analyse customer experience information through surveys and feedback forms.
Launched back in July, Qualtrics iQ is made up of three elements:
- iQ directory - a data warehouse for experience data, which CEO Ryan Smith describes as: "A single system of record for experience data, all in one place across the organisation." Experience data is what Qualtrics calls customer or employee feedback information collected in its platform.
- Stats iQ - a tool for democratising statistical analysis for Qualtrics users.
- Predict iQ - an AI-powered tool for making predictions for things like customer churn rates.
This all sits within the core Qualtrics XM platform, which has been built to allow anyone within a business to gather insights for things like customer or employee satisfaction, product testing, brand or pricing research; analyse the needs and trends, report it back and drive continuous improvement, all in a simple web-based platform.
Now it wants to make the platform smarter by embracing machine learning features to make suggestions to users. It's not a unique strategy - Salesforce has done a similar thing with its Einstein brand - but it's smart for Qualtrics to help its users get actionable insights from their data.
Chief executive Ryan Smith and head of marketing Kylan Lundeen demoed the iQ technology in London today at its X4 Europe event. Smith started out by sharing an anecdote about the late David Rockefeller's famed rolodex.
A Wall Street Journal journalist got access to the billionaire's rolodex after his death at the age of 101, reporting that it contained the contact information for 100,000 people, with many including rich information about that person and their previous interactions.
"He meticulously logged the experiences he had with each person he met, from JFK to Nelson Mandela. Writing milestones in their life and details of every encounter. In other words, he was building the ultimate experience database," Smith said.
"Last year when we launched the platform it was phase one, it lets you get all customer experience and employee and brand and product experience under one single technology," he added. "However our best brands manage all four on that platform.
"Today we are launching phase two, which is why we want the data on a single platform because it's all about adding intelligence to these experiences on the platform and we call this Qualtrics iQ."
Qualtrics iQ features
Digging into the tools a little further, the idea behind the iQ directory is to have a smarter repository of this data so that users can analyse and cut it up in new ways without some of the heavy lifting normally involved, all with the primary aim of avoiding unnecessary or frustrating customer experiences. For example, there is now a button in the directory where users can find and group similar customers (high engagement, low satisfaction, for example).
Users can now set permissions to control how often a certain customer, or cohort of customers, can be contacted to help avoid unnecessary contact being made.
"With directory you can set permissions for whole organisation so no one creates bad experiences for that customer, so no one is asking questions you know the answer to or reaching out too many times," Lundeen said.
Directory also has features which will customises survey questions on the fly, according to Lundeen.
Then with Stats iQ users can start to analyse what customers are saying to offer better experiences in the future.
For example, Qualtrics saw that people who don't ski also want to spend more time with industry leaders at its annual Salt Lake City summit. "So that's an easy trade off," Lundeen said. "Take the ski day into an additional breakout day for attendees at our summit next year."
Using conjoint analysis within the Qualtrics platform, Smith claims to be able to get this whole process done in under an hour.
"The question creation, experimental design and randomisation happens behind the scenes, entirely automated and an hour later results trickled in," he said. "Started and built and sent it in under 30 minutes by simplifying the entire process down to four steps: define, refine, distribute and analyse."
Other features include iQ text, which allows users to do text analysis for things like visitor comments to start to derive patterns and sentiment analysis.
There is also a Driver iQ tool for 'key drivers', for pinpointing which variables matter the most for customers and suggesting actions, such as improving the loyalty of repeat passengers on an airline.
Finally, with PredictiQ customers can start to do forecasting within the Qualtrics platform using neural network algorithms. Giving the example of customer churn rates, Smith said: "It works by looking at last five years of churn data we have on an event. It pulls that data, [experience] data of feedback and the operational data in the directory and in seconds we have a table with a churn prediction for every single person and the ability to take action for each person or the whole cohort of people and give you a way to intervene."