Funnel screens

Supported screen kinds:

• intro - landing screen with headline and CTA
• question - single-question step (used for branching)
• content - informational content / value prop
• cta - call-to-action button
• form_collect - captures lead fields and creates a Submission
• redirect - sends the visitor to a URL
• outcome - final result screen (used with scoring)

Each screen has a position, a kind, a free-form config object, and an optional ab_variant_key. Screens with a variant key only render to sessions in that variant.

Scoring and outcomes

A funnel can score its visitors and send each one to the ending that fits. Scoring is optional: a funnel with no points and no outcomes behaves exactly as before.

• Points. Give a choice option a score: a number (added to the visitor's total) or a map of bucket to number, such as { "fitness": 5, "mind": 2 }, to keep separate subtotals. A block can also carry a block-level score map, counted when the visitor answered it (an input) or reached its screen (anything else). Points are numbers between -1,000,000 and 1,000,000; bucket names use letters, digits, hyphens and underscores.
• Outcomes. An outcome is a score band: a unique key, an inclusive score_min and score_max (leave one blank for no limit), an optional score_bucket to read one bucket instead of the total, and an optional screen_id that points at one of the funnel's outcome / redirect screens. Bands that read the same bucket cannot overlap. A score that falls in no band shows the funnel's regular ending.
• Server-side only. Formspring adds the points up when the visitor submits, from the points saved on the funnel. A score sent by the browser is never read. The visitor sees the matching ending, never their raw score.
• Where it shows up. The score, bucket totals and outcome key are saved on the submission, listed in the dashboard, the mobile app, the MCP tools and exports, and sent in webhooks (score, score_buckets, outcome) and to connected tools as _scoring.score and _scoring.outcome.key.

In the builder, the Preview tab has a "Simulate an outcome" panel: pick answers and see the score and the ending your current edits would produce. The score is shown only there, never to visitors. Funnel analytics shows how many submissions landed in each outcome.

Set points on an option in the builder's question settings and the bands under Settings > Outcome scoring, or send outcomes and option score values in a set_funnel_document call.

* * *

Screen order

The position controls the default flow. Visitors start at the first active screen and move forward unless branching or redirect behavior sends them somewhere else. Keep the first screen focused on the promise of the funnel, then ask only the questions needed to qualify or route the visitor.

* * *

Collection screens

form_collect is the bridge from funnel session to normal form submission. When it succeeds, Formspring creates the underlying Submission and includes funnel metadata in the payload. Put required lead fields here, and keep earlier question screens focused on qualification so the final collection step stays short.

* * *

Variant-specific screens

ab_variant_key lets you attach a screen to one experiment variant. Screens without a variant key are shared. A typical test keeps the intro and final collection screens shared, then swaps one question, proof screen, or offer screen per variant.

* * *

Content and redirect screens

Use content for proof, explanation, or reassurance between questions. Use cta or redirect when the goal is an external action instead of a form submission. Redirects still record funnel events, which makes them useful for tracking outbound bookings, payment links, or partner handoffs.