Tag, count, cost: how to find real customer problems in your field data

Most hardware companies already know what their customers struggle with. The evidence sits in service reports, warranty claims, support logs and reviews. The problem is that nobody adds it up, so product decisions get made on the latest angry email instead.

Here is a method you can run in one afternoon that turns that scattered evidence into a ranked list of problems with numbers behind them.

Engineer sorting service reports and returned parts

Where to look

  1. Warranty and return data: what fails, how often and what it costs.

  2. Service reports: what technicians actually do on site, and how long it takes.

  3. Support logs: what people call about, including "no fault found" cases, which often point to usability problems.

  4. Sales notes and lost deals: why customers chose a competitor.

  5. Public reviews: unfiltered customer language.

You don't need all five. Two good sources are enough to start.

Step 1: Tag

Pull 50 to 100 recent records into a spreadsheet. Give each one a short label written from the customer's point of view: "door hard to close", "error code unclear", "filter change takes too long". Avoid technical root causes at this stage. You're describing what the customer experiences, not what broke.

Step 2: Count

Merge labels that mean the same thing, then count how many records share each one. Sort from most to least frequent. Within an hour you'll usually see that a handful of problems account for most of the records.

Step 3: Cost

For the top five, estimate the cost per case: service time, spare parts, handling, a lost sale or a bad review. Multiply by frequency. The ranking often changes. A rare problem with a high cost can outrank a frequent cheap one.

Illustrative example. In a sample of 120 service records for a cordless vacuum, 34 concern battery runtime, 29 hair tangling around the brush roll, 21 a dustbin that is hard to empty and 12 a charging dock that tips over. Battery cases cost more each, but reviews and early returns show the brush problem hurts sales more.

Three traps to avoid

  • The loudest customer. One key account can dominate the conversation. Check whether its complaint also shows up in the data.

  • Engineering labels. "Capacitor failure" is a cause. "Stops working after 18 months" is what the customer feels. Keep both, but rank by the customer problem.

  • Survivorship. Data only shows customers who stayed and called. Those who quietly switched to a competitor are invisible, which is why sales notes and interviews matter.

Turn the top problem into a problem statement

Write your top problem in four parts: who has it, what happens today, the impact with a number, and the evidence source. For example: "Owners in households with pets cut hair off the brush roll about once a week. It appears in 24% of sampled service records and is the most common theme in low-rated reviews. Source: service data, five interviews, 200 reviews."

Notice what's missing: any mention of a solution. That comes later. A rough number with a clear source beats a strong opinion in every product meeting.