INT-07 · SEC. 09 Competitive Intelligence

Competitor Complaint Mining

Mine a competitor's negative reviews for weaknesses that recur, are recent, and a buyer would switch over.

FORMAT: goal
DIFFICULTY: advanced
TIME: 20 min
TOOLS: universal
MODELS: any
COPIES: 0 so far

When to use this

You're building a battlecard, comparison page, or objection-handling doc against a named competitor and want their real weaknesses, in their own customers' words, not the weaknesses you assume they have. Use it before you write a single "unlike them, we..." line.

The pattern

Pastes as plain text

Goal: build a ranked list of the real, recurring weaknesses of the competitor I name at the bottom of this prompt, backed by their own customers' words. Done means every listed weakness passes all three tests below. Anything that fails a test goes in a separate "heard once, unproven" list. Nothing gets quietly dropped and nothing gets quietly promoted.

Tests for a listed weakness:
1. Recurrence: complained about by at least 3 distinct reviewers or commenters, not 1 loud one.
2. Recency: at least one supporting complaint from the last 12 months. A weakness they fixed two years ago is not a weakness.
3. Exploitability: a buyer could plausibly switch products over it. "Support takes days" counts; "the logo is ugly" doesn't.

Method: search the competitor's reviews on G2, Capterra, TrustRadius, app-store listings if they have an app, and Reddit or Hacker News threads that mention them. Read the low- and mid-star reviews first; use 5-star reviews only for their "cons" field. For each candidate weakness, collect supporting quotes with source and date.

Output, ranked by recurrence count:
- The weakness, one line
- Evidence: source count, date range, and the two clearest verbatim quotes with source
- The discovery question: one question a sales rep can ask a prospect that gets the prospect to name this pain themselves

Then the "heard once, unproven" list, one line each. End with the single weakness you'd lead with and why.

Competitor name (plus URL if the name is ambiguous):

Real example output

37 negative and mid-star reviews read across G2, Capterra, and Reddit.

1. Billing surprises at renewal (9 sources, Aug 2025 to May 2026). "Renewal came in 40% higher with zero notice" (G2, 2-star, Jan 2026). "Auto-renewed us into an annual plan we never picked" (Reddit, Mar 2026). Discovery question: "How did your last renewal conversation with them go?"
2. Report builder slows past ~50k rows (6 sources, Oct 2025 to Apr 2026). Two quotes name export timeouts specifically. Discovery question: "What's your row count on the biggest report you run?"
3. Support escalations stall after first response (5 sources, all within 12 months). Discovery question: "When something breaks, how long until you talk to an actual engineer?"

Heard once, unproven: mobile app logs users out daily (1 App Store review); SCIM sync broke after an SSO update (1 G2 review).

Lead with #1: recent, recurring, emotionally charged, and it lands at the exact moment a customer is deciding whether to stay.

Why it works

The three tests are a filter applied per item, which is what keeps one angry Reddit thread from becoming your lead sales message. The "heard once, unproven" list preserves weak signal without letting it pass as fact. Turning each weakness into a question a prospect answers themselves beats a rep asserting it and hoping.