Customer Feedback and Support Log Analysis

Description

Systematically mines support requests, feedback forms, forums and front-line staff for pain points and clues, reading with explicit rules and tracking topics over time.

Support and feedback channels are a continuous, free stream of user evidence, but they are biased and easily misread. The method starts by listing every channel, including informal ones such as salespeople, and examining how each channel's design (tone of invitation, follow-up, auto-responses, difficulty of reaching you) shapes what arrives. Analysts read a broad sample of actual comments with their context, sort them into Muller's three types (actionable suggestions, statements of the obvious, and clues, usually the richest), and code for who users are, what they are trying to do, how they approach problems and what problems they have. Five reading rules apply: take the user's perspective, facts first, no jumping to conclusions, complaints point to problems rather than describe them, and remember the silent majority. Trends are tracked over time and compared with behavioural data.

Attributes

Loss type
Vestigial
Effort
Days

Steps

  1. List every feedback channel, including informal ones such as sales and account managers, and note how each channel's process biases what arrives.
  2. With help-desk software, review historical data on who comments and how often, then focus on recent weeks or a specific problem; without it, interview support staff and collect their stock answers.
  3. Read a broad sample of actual comments, not only answered ones, keeping context such as the feature or path the commenter was on.
  4. Sort feedback into actionable suggestions, statements of the obvious and clues, and code for who, goal, approach and problem.
  5. Apply the reading rules: user's perspective, facts first, no premature conclusions, complaints as pointers not specifications, remember the silent majority.
  6. Use topic and sentiment clustering to compare groups and track change over time; interpret a doubling of complaints after a release as a change for the worse, not twice as many unhappy users.
  7. Set a regular review slot, share findings with front-line staff, and compare perceived problems with usage data.

Sources

  • Observing the User Experience (Goodman et al., 2012)
  • Just Enough Research (Hall, 2013)
  • Usability Engineering (Nielsen, 1993)

Also known as

  • Customer feedback analysis
  • Support log analysis
  • Voice of the customer analysis
  • Support ticket analysis

Categories

  • Analytics & Behaviour
  • Product Analytics & Strategy