The not so lonely journey of a product issue through unitQ Monitor
Did you ever wonder how a quality issue becomes a quality issue?
Let’s start by defining a quality issue. A quality issue is any user friction with your product, where your product doesn’t work the way the user expects it to work. This includes software bugs obviously, but quality issues are more than that. Quality issues are also anything to do with the overall user experience. Quality issues are not matters of taste, like feedback about the site’s look and feel; nor are they issues like forgotten passwords.
Now let’s talk about the journey a quality issue takes through unitQ Monitor
Every quality issue follows the same pipeline before you see it in your unitQ Monitor. Quality issues can come from many sources: your customer ticketing system like Zendesk, app store reviews, and social media. unitQ Monitor tracks all this data for you. The first thing it does is clean the data. During this phase, we trash around 10 to 20% of data during cleaning to avoid introducing unwanted noise into the Quality Monitors, such as duplicate reports, spam, email headers, long response threads and other irrelevant data. unitQ Monitor tags this feedback as trash but does not delete it.
Next, unitQ Monitor cleans, normalizes and translates the remaining data
Cleaning involves removing all personally identifiable information, like names, addresses, email addresses, and so forth.
Normalizing involves applying structure to the data, like geographic data and user sentiment. User sentiment can be positive, neutral, or negative. Quality issues tend to fall into negative or neutral sentiment.
Then it translates content into English. Sometimes, product issues can slip through the cracks if you aren’t able to translate customer feedback into a language that you understand.
Now the machine learning kicks in, determining which of the remaining issues are actual quality issues
Most issues are not quality issues, rather they are feature requests, compliments or props, and password resets and other expected use cases. unitQ staff constantly reevaluates and retrains the ML model so it is always growing in accuracy.
The quality issues get assigned to one of hundreds of Quality Monitors, which are collections of quality issues that are all related to the same product defect.
At this point, these quality issues are now visible in your unitQ Monitor. Their presence can trigger alerts you’ve already set up; these alerts can notify your team through PagerDuty and Slack. You can also open tickets in Jira so your engineers can track and fix the issues.
Quality issues serve as canaries in the coalmine to make you aware of problems before they get too big, providing a way for you to make data-driven decisions so you can more accurately prioritize the issues you need to fix first, and fast. By acting quickly, your product quality should quickly improve, as should your unitQ Score.
To learn how unitQ is helping category-leading companies like Chime and Strava detect and fix urgent issues faster and incorporate crucial user feedback into their product decisions, request a demo today.