Frustration Hunter

Find what's frustrating your users, ranked by impact, grounded in real session evidence.

When to Use

• "What are the biggest frustrations on our site right now?"
• "What's causing friction on the checkout flow for mobile users?"
• "What got worse this week vs last week?"
• "Show me the top rage-clicked elements across our product"
• "Find sessions where users are dead-clicking on the settings page"

Use free fullstory:discover_groups and fullstory:get_opportunity_stats for live signals. Discovery accepts one scope at a time (page_id, domain, app_name, segment_id, or funnel_id); use a segment for device/user filters and the live schema for time ranges. Use computed metrics for exact custom windows.

Workflow

Step 1: Discover opportunities

Call fullstory:discover_groups for the requested window (default last 7 days, using the live schema). This returns the top live frustration signals — rage clicks, dead clicks, form abandonment, error spikes — each with a user count and page location.

If the user asks about a specific page or flow, scope the opportunities to that area. If they ask "what got worse," set compare_to_previous=true.

Step 2: Get live stats

For the top 3-5 opportunities, call fullstory:get_opportunity_stats on each. This returns:

• Exact user count (not an estimate)
• Device split (mobile vs desktop percentage)
• Page concentration (which pages it happens on)
• Rate-of-change vs prior period (is it getting worse?)
• Frustration rate vs site average

Present the top frustrations as a ranked list:

1. Rage clicks on "Apply Promo Code" — 412 users, 73% mobile, /checkout
2. Dead clicks on cart icon — 287 users, Safari-specific, /products
3. Form abandonment at password field — 218 users, 22s avg dwell before leaving

Step 3: Pull sessions for the worst offenders

For the top 1-3 signals, build a metric describing the observed frustration, compute it, and call fullstory:get_sessions(metric_id="<id>"). Review 3-5 returned sessions per signal.

If StoryAI Premium lookup tools are available and the user wants pre-ranked stored opportunities, use get_opportunities (time_range="7d", or 24h, 14d, 30d), get_opportunity, and get_sessions_for_opportunity. get_managed_funnels is also Premium-only. Otherwise continue with live discovery and metric-backed sessions; do not require an upgrade.

For each session, use the session-context agent to read the transcript with a specific task:

• "Did the user's action succeed or fail? What happened after the frustration event?"
• "What element did they rage-click and did anything change on the page after?"
• "Was there an error in the console when this dead click happened?"

Step 4: Synthesize

Look for patterns across sessions:

• Same element failing across users → likely a bug
• Users clicking non-interactive elements → likely a design issue
• High abandonment after a specific step → likely a UX flow problem
• Errors only on specific browsers/devices → likely a compatibility issue

Present findings with: what's happening, how many users, where it concentrates, what the session evidence shows, and what the likely root cause is.

Deeper cuts

If the user wants dimensional breakdowns beyond what get_opportunity_stats provides (browser version, custom user property, etc.):

1. Build a metric for the specific frustration event: fullstory:build_metric(query="rage clicks on Apply Promo Code", output_type="top_n") with the desired grouping dimension
2. Compute it: fullstory:compute_metric(metric_id)
3. If scoping to a cohort, attach a segment first: fullstory:update_metric(metric_id, segment_id) then compute

Guidelines

• Default time range: last_7_days. Ask before using a different window.
• Start with 3-5 opportunities. If the pattern is clear, stop.
• Always ground findings in session evidence — numbers tell you "what" and "where", sessions tell you "why."
• When presenting, include both the metric (user count, device split) and the qualitative finding (what users actually experienced).
• If an opportunity has a sharp rate-of-change (getting worse fast), flag it prominently.