OnClarity Insights
How do we get from monthly reports to seeing today’s problems today?
Asked by a product leader at a US consumer platform.
5 min read
The short answer
Stop waiting for the report and measure time to know: the hours from the first customer signal to the issue raised with a named owner. Check every conversation daily for new or rising issues, route each one to an owner against a threshold, and review what is open each week. The monthly report then confirms what you already knew.
Why is the monthly report always late?
Not because the team is slow. Because every step waits on the one before it.
A customer mentions a problem in a chat on the 3rd. Others mention it in calls and app reviews over the next week. Nobody connects them, because each conversation is filed under a category somebody chose a year ago. The month closes. The analysis takes another week or two. The issue appears in the deck as a rising theme, gets discussed, and is handed to someone to look into. By then the customers have been explaining it for a month, and many of them have stopped explaining it at all.
Most teams respond by making the dashboard faster. That rarely helps, because the delay does not sit in the dashboard. It sits in two places: noticing that a pattern exists, and getting a named person outside customer service to own it.
The cost of finding out late is real. In SQM Group’s research, 41% of customers had to contact the organisation more than once about the same problem. And the customers who give up on you do not stay quiet: in the 2025 National Customer Rage Study, one in four complainants posted about their most serious problem on social media, and 43% said the company never responded.
What are the four clocks of time to know?
Every problem passes four points. The gaps between them are where the days go.
Signal. The first time a customer mentions the problem, in any channel.
Detected. Someone, or something, recognises it as a pattern rather than a one-off.
Owned. A named owner, usually outside customer service, accepts the issue.
Fixed. The cause is removed and contacts about it fall.
Time to know runs from signal to owned, measured in hours. Fixing is the next job; you cannot start it until someone owns the problem.
Illustrative example, not customer data.
How do you shorten each clock?
Signal: listen where customers already talk.
Read conversations, reviews and public posts, not only surveys. Many customers never fill in a survey about the problem that annoyed them.
Keep every conversation, not a sample. The first three mentions of a new problem are easy to miss in a sample.
Detected: look every day, against a baseline.
Check daily for issues that are new or rising compared with the last few weeks.
Set a threshold per issue type, so a spike in payment failures is raised faster than a spike in feature requests.
Look for topics nobody has named yet. The categories you already have only find the problems you already knew about.
Owned: give every issue type a named owner outside customer service.
Agree owners for the usual sources of problems: product, billing, logistics, policy.
Agree an acknowledgement time, for example one working day for high-value issues.
Hand over evidence with the issue: example conversations, volume and what it is costing.
Fixed: write the outcome down before the fix ships.
Decide what success looks like, for example contacts on this issue down 40% within 30 days.
Track contacts on the issue until they fall, and close it only when the drop holds.
How do you measure time to know from Monday?
Take the last three issues that reached your leadership team. For each one:
Find the first customer who mentioned it, in any channel. That is the signal.
Find when someone first called it a pattern. That is detected.
Find when a named owner accepted it. That is owned.
Count the hours between each.
The biggest gap is your first thing to fix. In most teams it is between detected and owned: the problem was noticed, then sat in a deck or a channel with nobody responsible for it.
What is the monthly report for now?
The monthly report still matters. It shows trends, it is what the board reads, and it is where you prove that fixes held. What it should not be is the place where you discover problems. If the monthly report regularly tells your leadership something new, time to know is too long.
The honest trade-off
Faster knowing has costs. Alerts without owners create noise, and teams quickly learn to ignore them. Daily detection needs someone to triage what it finds, which is real work. Some issues need a few days of data before you can tell them from noise, so “today” sometimes means “this week”. And knowing faster is wasted if nobody outside customer service acts. Start with owners and thresholds, then add speed.
What to do this week
Pick the last three issues that reached leadership.
Find the first customer mention of each one, in any channel.
Time the four clocks for each issue, in hours.
Name an owner outside customer service for your top five issue types.
Set one daily threshold for your most expensive issue type.
Questions people ask
How do you analyze voice of customer data?
Start from conversations rather than surveys: support chats, calls, emails, app reviews and public posts. Group them into issues by meaning rather than by keyword or old categories, compare each issue’s volume with the previous weeks, and send new or rising issues to a named owner with example conversations attached.
How can I automate customer feedback analysis?
Automate the reading and grouping, not the decision. Software can read every conversation, group mentions of the same issue and flag what is new or rising against a baseline. People should still decide which issues matter, who owns them and what outcome counts as fixed.
What is time to know?
Time to know is the number of hours from the first customer signal about a problem to the moment a named owner accepts it. It covers two gaps: noticing the pattern and getting someone responsible for it. Report the median and the slowest 10% each month.
Sources
SQM Group, Call Center Customer Experience Research (41% of customers contacted the organisation more than once about the same problem), accessed Oct 2026.
W. P. Carey School of Business, Arizona State University, and Customer Care Measurement & Consulting, 2025 National Customer Rage Study, press release 2 Dec 2025.
The timeline above is illustrative. It uses example dates, not OnClarity customer data.
