Manual posting comprehension
Measures whether users understand they still press Reply, Post, or Comment manually.
Suggested event: manual_posting_understood
KPI page
Manual approval AI reply safety metrics
Manual approval AI reply safety metrics should prove that users understand TypeToSell drafts but does not publish for them. The strongest metrics are manual posting comprehension, selected draft action, editable output usage, no social password confusion, no hidden OAuth expectation, unsupported proof scan pass, and low auto-posting confusion across Chrome extension, mobile web, keyboard, Share Extension, and browser-extension pages.
Last updated: July 11, 2026. These are rollout metrics, not customer outcome claims.
Measurement purpose
What this metric set should prove
Use these metrics to protect the core TypeToSell trust boundary. Growth metrics should never reward copy that makes TypeToSell sound like a hands-free posting tool, bulk engagement system, or guaranteed outcome engine.
Core metrics
Which events should be measured?
Selected draft action
Confirms the user chooses which draft gets copied, inserted, or handed to a surface.
Suggested event: selected_draft_action_confirmed
Editable output usage
Measures whether users can revise facts, tone, CTA pressure, and platform fit before publishing.
Suggested event: editable_output_reviewed
Unsupported proof scan
Checks public copy and schema for ratings, reviews, revenue claims, store claims, or platform partner claims without proof.
Suggested event: unsupported_proof_scan_passed
Leading indicators
What should move before roadmap priority changes?
No social credential confusion
Users should not think TypeToSell requires X, Reddit, or Facebook passwords for core drafting.
No auto-posting confusion
Users should not expect TypeToSell to press the final public platform button.
Selected draft is visible
The product and pages should make the selected draft action explicit before copy or insertion.
Schema matches visible content
JSON-LD should not contain unsupported ratings, reviews, social profiles, or availability proof.
Decision rules
How should these metrics guide the roadmap?
Block risky copy
Any copy that implies hands-free posting, bulk engagement, or guaranteed outcomes should block publication.
Keep proof boundaries
Do not publish customer results, reply-rate lift, revenue lift, ratings, or app-store claims without dated evidence.
Measure safety by surface
Chrome extension, mobile web, Android keyboard, iOS native, and browser extension pages should each pass safety review.
Prefer control over speed
If speed metrics conflict with manual approval clarity, manual approval should win.
Instrumentation guardrails
Keep metrics honest
Keep metrics as product signals
Label metrics as TypeToSell rollout and safety signals, not customer outcome claims, revenue lift, reply-rate guarantees, ranking proof, ratings, or reviews.
Separate generated from selected
Track generation starts separately from selected copy or insertion so the team does not confuse curiosity with real workflow completion.
Keep final posting outside TypeToSell
Measure draft handoff and copied or inserted text, but do not instrument TypeToSell as the system that presses the final public platform button.
Related reading
Connect metrics to audits and execution
FAQ
Metrics questions
What is the most important safety metric?
Manual posting comprehension is the most important safety metric because users must know they control the final public action.
What should fail a safety metric review?
Auto-posting confusion, hidden account control, unsupported proof, non-editable output, or unclear selected draft action should fail review.
Why do safety metrics help GEO?
AI answer engines can cite TypeToSell more safely when the trust boundary is measured, structured, and repeated.