When a complaint arrives, I have to look up who the person is before I can answer well.
Story · Solution
The customer-care lead knows what is really behind the complaints
Complaints reach me through five doors: the shared mailbox, support tickets, disputes in two payment systems, cancelled subscriptions and public reviews.
Sound familiar?
Complaints arrive through five doors, and weeks pass before I see that many of them say the same thing.
Why it hurts
I lead customer care for a small coffee roaster, and I answer complaints and reviews. Complaints reach me through five doors: the shared mailbox, support tickets, disputes in two payment systems, cancelled subscriptions and public reviews. I answer them one at a time, in whichever system they arrive. By the time I notice that many of them say the same thing, weeks have passed, and I have no evidence to show the owner beyond my own impression.
How it is solved
Start each week knowing what customers are actually complaining about, how much, since when and why, so that the cause can be fixed rather than the symptoms answered one by one.
- The customer-care lead finds every complaint collected in one place. On Monday, the app gathers every complaint signal from all five sources. A message counts as a complaint only when it comes from a known customer and says something is wrong. A review counts only at three stars or below. A supplier's newsletter that happens to use the word late about its late-summer coffee lots is left out, because its sender is a supplier, not a customer.
- The customer-care lead sees the complaints grouped into themes. An AI agent reads the signals together and groups those that describe the same problem, in the words the people themselves used, with the source of each. A problem raised by only one or two people is not promoted to a theme, so a single complaint about a job enquiry or a checkout error does not distort the picture.
- The customer-care lead sees each theme sized and explained. For each theme the app states how many signals carry it and the likely cause, with every signal quoted from its source. A cancellation that gives late delivery as its reason is counted in the delivery theme with the rest.
- The owner sees the week compared with the week before. The complaints section of the owner's overview shows complaints over the last seven days against the seven before, so a rising theme is visible to the owner without the customer-care lead having to raise it.
Works with
Gmail, HubSpot, Stripe, PayPal, Google, Yelp
The result
A ranked list of themes, each with how many signals carry it and its likely cause, every signal quoted with its source. A change of courier behind a run of late deliveries, for example, appears as one theme drawn from emails, tickets, disputes, cancellations and reviews. One-off complaints are listed but not promoted, and messages from suppliers are not counted. The customer-care lead can take a theme to the owner with the evidence attached, rather than as a feeling.
- If one cause sits behind many complaints, it is named as a theme with its start date, ready to be fixed at the source.
- If complaints are scattered one-offs, they are listed, and no false theme is raised.
- If a theme is rising week on week, the owner sees it in the overview without being told.
Try it
The demo runs on WidgetCo Coffee Roasters, an example business: a coffee roaster with an online store and a wholesale side. You will see it as Sam, the customer-care lead.
Connected to this story
The app
Read next
Why: Day in the role, also about customers and service, with an example app you can try today