Lesson 81.1 proved the value returns. Now a stranger stands at the door: how much do they get to try before paying, and what makes them convert? This lesson compares the four honest doors — free trial, freemium, paid trial, demo — and forces you to price the free part like the distribution and support decision it is. (Product-track lesson: community-track readers, the same door logic applies to paid workshops and member previews in Lesson 83.4 — note what you defer.)

The vocabulary, in plain language

• Free trial: full (or near-full) product access for a limited time — typically 7 to 30 days — then pay or lose it. Shows the whole product against the clock.
• Freemium: a free tier that works indefinitely within durable limits — fewer projects, fewer seats, public-only output. A small permanent outcome, not a countdown.
• Paid trial: a low-priced, time-boxed entry — for example $9 for two weeks — that qualifies commitment before full price. Filters curiosity from intent.
• Demo: a guided walkthrough, live or recorded, for complex or high-touch value where self-serve trial would confuse. The seller shows; the buyer asks.
• Activation: the first moment the customer gets real value — first checked brief, first monitored topic, first shared workspace. Trials live or die here, not at signup.
• Paywall: the boundary where free stops and paid begins. Legible paywalls convert; foggy ones annoy.
• Conversion: a free or trial user becoming a paying customer. Measured as a rate over a cohort, not a vibe.
• Upgrade trigger: the event that makes paying the obvious next step — hitting the project limit, needing private workspaces, wanting the team in.

The mental model: each door teaches you something different. Trials teach whether the whole product impresses fast. Freemium teaches whether a small free outcome distributes itself. Paid trials teach who will spend anything at all. Demos teach what a complex buyer actually needs to believe.

Tradeoffs: four doors, four bets

Free trial — the whole product, whole-time. Best when activation is fast and the product explains itself in one session. Strength: the buyer experiences full value. Risk: they binge, extract the one-time outcome, and leave — or never activate and conclude "not for me" from an empty workspace.

Freemium — a durable small outcome. Best when a limited free job spreads by itself — one public brief a month, a single watchlist. Strength: permanent distribution engine. Risk: a permanent support and cost obligation, plus a free tier good enough that nobody upgrades.

Paid trial — the qualifier. Best when support costs are real or tire-kickers flood free doors. A small payment filters for intent. Strength: higher conversion downstream, fewer free riders. Risk: volume collapses — many who would have tried free never start.

Demo — the complex touch. Best when value needs configuration, trust, or a committee. Strength: objections surface live. Risk: it does not scale, and a demo-dependent product hides whether self-serve activation works.

There is no universally best door. There is only the door whose learning signal you need next and whose cost you can carry.

Free is distribution plus a support duty

Treat every free offer as two decisions stapled together. As distribution, free lowers the barrier and lets the product travel — sample briefs get forwarded, free workspaces get linked. As an obligation, every free user can file tickets, consume model and hosting spend, and expect the free thing to keep working.

Free is not automatically good growth. A free tier that costs $2 per user per month in model calls, with 5,000 free users and 40 payers, is a $10,000 monthly donation to churn. Before launching free, name who answers free-user support, what free costs per account, and what kills the free tier if the math breaks. The Stripe Billing documentation shows how trials, meters, and paywalls are enforced mechanically; this lesson decides whether the mechanics serve a business.

Desk example: three doors, three signals

Apply it to the source-backed Research Desk:

• Sample public brief: publish one checked brief openly. Signal: does anyone forward it, cite it, or ask "can you do this for my topic?" Tests distribution and credibility, not product use.
• Time-boxed private workspace (14-day trial): the visitor researches their own topic with full features. Signal: do they reach activation — first brief plus one saved source collection — within the window? Tests whether the whole product impresses fast.
• Limited saved watchlist (freemium): one watchlist, three monitored sources, public-only briefs, free indefinitely. Signal: do free users return weekly and hit the limit honestly? Tests whether a small durable outcome retains and triggers upgrades.

Each door answers a different question. Pick the one you most need answered, run one door well, and note what the others defer.

Exercise: map one access model end to end

Create ACCESS-MODEL.md for exactly one door — a stranger should see where free ends and why paying is natural.

# ACCESS-MODEL.md — [Project]

## Chosen door + why (one sentence)
- Free trial / freemium / paid trial / demo — because the question I need answered is: ___

## First useful action (activation)
- The user does ___ within ___ days/hours and sees ___ as proof it worked.

## Limit (the paywall, stated plainly)
- Free includes: ___ / free excludes: ___ / trial lasts: ___ / demo covers: ___

## Upgrade moment (the trigger)
- When the user ___, paying becomes the obvious next step because ___.

## Value preserved after the limit
- If they do not pay, they keep: ___ / they lose: ___ / export path: ___

## Support + cost impact
- Free/trial support owner: ___ / expected cost per free account: ___ / monthly free budget cap: ___
- Kill/adjust rule: "If ___ by [date], I will ___." 

## Deferred doors
- Not running yet: ___ — because: ___

Worked mini-example — Desk time-boxed trial: activation is a first brief on the visitor's own topic plus one saved source list within 48 hours. Limit: 14 days, two private workspaces, then private briefs pause. Upgrade trigger: the second watchlist alert arrives — "3 more monitored topics unlock on paid." If they do not pay, they keep exported PDFs and the public sample; private monitoring pauses but nothing is silently deleted for 30 days. Support owner is the founder, capped at 100 trial accounts this month; kill rule: "If fewer than 1 in 5 trials activate by day 3, I shorten onboarding before extending the trial."

Finish line: an ACCESS-MODEL.md with first action, limit, upgrade moment, preserved value, and support/cost bounds.

Verify quickly: ask a friend to play the stranger. Give them 60 seconds with the free door description. If they cannot say what they get, when it ends, and what paying unlocks, rewrite the paywall sentence until they can.

Common failure mode: the foggy paywall — "free plan with limited features, upgrade for more." Nobody upgrades to "more." They upgrade when a named trigger (second workspace, private briefs, team seats) meets a named outcome. Name both.

Check your understanding

1. In your own words, distinguish a free trial from freemium, a paid trial, and a demo. 2. Why is free access a support obligation as well as a distribution decision? 3. What five parts must ACCESS-MODEL.md make legible, and what does each prevent?

Next

The door is chosen. Lesson 81.3 prices what is behind it — tiers and billing units the customer can actually understand.