Why Prototype the Experience Before You Fund the Build?

I keep watching teams schedule a product build to “see if people like it,” polish UI before a behavior hypothesis exists, and treat stakeholder applause at a demo as desirability proof. Those are expensive costumes. Experience prototypes falsify whether people can finish the job with dignity — cheaply.

Cheap evidence sits in the middle between insight and scale — see 6 Stages Most Organizations Skip Between Insight and Scale. Instrument learning with something like The Experiment Canvas™. And do not confuse a clickable demo with learning — prototype-as-finish-line is ceremony in 6 Design Artifacts Worth Keeping — and 6 That Are Ceremony.

If you haven’t felt the experience, you haven’t prototyped it — you’ve only scheduled a build.

1. How Does Concierge Delivery Prototype Experience?

Method: Deliver the end-to-end experience manually for a small set of humans — humans are the system.

Falsifies: Desirability and job fit — will they hire this outcome done this way?

Run lean: Cap at N customers or employees; script the promise; log every step and friction.

Kill/continue: They return, refer, or abandon a workaround — or they ghost after one try.

2. What Is Wizard of Oz Experience Prototyping?

Method: Present an “automated” or agentic experience while a human performs the work unseen.

Falsifies: Interaction trust — clarity, control, and competence feel — before you build the agent.

Run lean: Chat, voice, or UI shell; human operator; record where people hesitate or demand a human.

Kill/continue: They complete with confidence — or trap, retell, or escape the channel. For the trust contract when agents later act for real, see 6 Trust Pillars for Agentic Customer Experience.

3. How Do Fake-Door / Offer Tests Work Without Building?

Method: Offer the experience — landing page, button, email, QR, shelf talker — before it exists; measure intent.

Falsifies: Demand signal without building fulfillment.

Run lean: Clear promise; easy signup; honest “not ready yet” follow-up; no dark patterns.

Kill/continue: Qualified interest above threshold — or curiosity theater with zero follow-through. Fake doors that shame people are trust violations, not prototypes.

4. Why Run a Paper or Sketch Walkthrough?

Method: Walk real users through screens, cards, or paper steps that stand in for the product.

Falsifies: Flow risk — where the job breaks, where language fails, where dignity costs spike.

Run lean: Low fidelity on purpose; one job; watch hands and faces; don’t defend the sketch.

Kill/continue: Time-to-first-success improves — or they invent a workaround mid-walkthrough. Better framing questions before you freeze a build live in 10 Design Questions That Beat a 40-Page Requirements Document.

5. What Is a Service Rehearsal Prototype?

Method: Tabletop, then live rehearsal of the service with real role owners — frontline, backstage, partner.

Falsifies: Operating-model risk — orphan seams, unclear ownership, policy collisions.

Run lean: One journey; timers; “who owns this seam?” cards; stop when a seam has no owner.

Kill/continue: Seams named and staffed — or the rehearsal collapses into “IT will figure it out.”

6. How Does Moment-of-Truth Enactment Prototype Feeling?

Method: Role-play or staged enactment of the critical emotional beat — recovery, consent, first win, bad news.

Falsifies: Emotional job and stakes — what success and betrayal feel like.

Run lean: Real customers or employees if possible; otherwise trained proxies plus later validation; debrief feelings, not features.

Kill/continue: People say “I’d trust that” — or the room goes quiet at the dignity cost.

7. What Is a Manual-Backed Facade?

Method: Ship a thin front — form, chat, page — backed by spreadsheet or ops humans, not a platform.

Falsifies: Promise vs capacity — can you keep the experience promise at small scale?

Run lean: Explicit capacity cap; promises you can keep; log unpaid labor and exception types.

Kill/continue: Ops can sustain with dignity — or heroes burn out keeping the costume alive. Before you fund the bigger pilot, use 11 Questions Before Funding Any Innovation Pilot.

8. How Does a Pop-Up Front Door Prototype Channel Fit?

Method: Stand up a temporary physical or digital front door where the job already happens.

Falsifies: Context and channel fit — will they enter here, not only in your preferred portal?

Run lean: Hours or days, not months; observe wrong-door and escape; capture verbatim jobs.

Kill/continue: Traffic converts to completed jobs — or people walk past to the old path.

9. How Do You Prototype Experience With Human-Scripted Conversation?

Method: Run the conversational experience with a human following — and adapting — a script across SMS, chat, voice, or desk.

Falsifies: Dialogue design — can conversation finish the job without loops or shame?

Run lean: One intent; escalation rules; measure completion and “felt heard.”

Kill/continue: Job done in one conversation — or an escalation storm and retelling tax. Prototyping without a named behavior is a classic design-thinking misuse — see 7 Ways Design Thinking Gets Misused.

10. Why Rehearse Recovery Before You Scale?

Method: Deliberately break or simulate failure; rehearse undo, apology, refund, rebook, and human handoff.

Falsifies: Trust under failure — control, care, and accountability when the happy path dies.

Run lean: One failure mode; powered recovery band; time-to-make-right; named accountable human.

Kill/continue: Recovery restores trust — or “the system decided” leaves nobody askable.

What Should You Ask Before the Next Build Request?

Five questions for experience prototyping:

1. What named behavior are we falsifying?
2. Which of these ten methods is the cheapest honest test?
3. What is the kill/continue date?
4. Who feels the experience — real humans or stakeholders only?
5. What must we not build until the experience proves out?

Mantra: Prototype the experience. Build the product only when the feeling and the job survive contact with humans.

FAQ: Prototyping Experience Without Building a Product