Pretotyping for product teams

Make sure you're building The Right It — before you build it right.

validaid turns your boldest product bets into a living evidence tree. Run fast, cheap pretotypes, collect your own data, and let validated ideas earn their place.

Inspired by Alberto Savoia's The Right It · No passwords — sign in with a magic link

validaid · Fresh Beans evidence tree

Vision

Fresh local beans, delivered weekly — better coffee with zero effort.

Fake door proves demand
Validated
Signups pay $19/month
In progress
Starter box converts fence-sitters
Backlog

Grounded in a proven method

The discipline behind pretotyping — now a workspace

validaid puts The Right It into practice — the method created by Alberto Savoia, Google's first engineering director and self-described “Innovation Agitator.” His Law of Market Failure is the reason this exists:

“Most new products will fail in the market — even if competently executed.”— Alberto Savoia, The Right It

The fix isn't building better — it's making sure you're building the right it first. validaid is how teams run that check.

~90%

of new products fail in the market

Days

to first real data — not months

1

place for every bet & its evidence

Pretotyping plays you can run this week

Fake DoorConciergeMechanical TurkOne-Night StandPinocchioPretend-o-typeYouTube test

The method

Four principles, baked into every card

Most ideas fail because they were built right — but they were the wrong it. validaid puts Alberto Savoia's pretotyping discipline at the center of how your team works.

Pretotyping

Test The Right It before building It right

Validate that people actually want what you are about to build — before you sink months of engineering into it. Run fast, cheap, scrappy experiments that produce real signal.

YODA

Your Own Data, Always

Other people’s opinions, market reports, and gut feel don’t count. Every bet is judged on data you collected yourself, from real people taking real actions.

Skin in the Game

Measure commitment, not compliments

A nod is cheap. A signup, a click, a pre-order — that’s skin in the game. validaid scores how close each test gets you to genuine commitment.

Falsifiable bets

Write hypotheses you can be wrong about

We believe X, therefore we will Y, we are right if Z. A clear numeric threshold and timeframe means every experiment ends in a verdict — Validated, Plausible, or Busted.

The loop

From hunch to evidence in four moves

Every bet runs the same disciplined loop. The tree remembers what you learned.

01

Frame the bet

State a falsifiable hypothesis: what you believe, what you’ll do, and the measurable result that proves you right.

02

Run the pretotype

Move the card to In Progress. Run the fastest experiment that gets you your own data — a fake door, a concierge test, a landing page.

03

Record the data

A learning card captures the facts — observations only, no opinions — then the insights you draw from them.

04

Pivot, scale, or die

Set a verdict. Branch new bets from what you learned. Watch validated paths light up across the tree.

Always run the smartest test next

validaid scores every experiment on three YODA dimensions, 1–5, where lower is better. The total tells your team which bet gets you trustworthy data fastest, closest, and cheapest — so you never waste a sprint on an expensive test when a scrappy one would do.

TTD
Time to Data
How fast will you learn?
DTD
Distance to Data
How close to real commitment?
CTD
Cost to Data
How expensive to run?

Everything you need

A workspace built for evidence, not opinions

Evidence tree canvas

Your whole strategy as one visual tree, reading left to right from vision to evidence. Validated paths light up green.

YODA scoring

Rank every test by Time, Distance, and Cost to Data — so you always run the fastest, cheapest, closest experiment next.

AI pretotyping coach

Sharpen vague hypotheses, get YODA scores with reasoning, and surface hyperzoomed cheaper variants of any bet.

Team dot-voting

Surface the bets your team believes in most. One vote per person — skin in the game for prioritisation, too.

Tasks & comments

Turn each experiment into action. Assign tasks, capture discussion, and keep the evidence and the work in one place.

Multi-project workspaces

Run every product bet your company is making side by side, each with its own vision and evidence tree.

See it in action

From hunch to evidence, on one screen

Every bet is a card: state what you believe, run the cheapest test, record the data, set a verdict. Here's a real one going from idea to validated.

validaid · dashboard

Here's the evidence your team has collected so far.

3

Experiments running

7

Bets validated

12

Evidence collected

64%

Validation rate

TestValidated
We believe
coffee drinkers are frustrated enough with stale beans to switch
Therefore we will
run a fake-door landing page with a $50 ad budget
And measure
visitor → email signup rate on the landing page
We are right if
≥5% of visitors leave their email
YODA score
TTD 3DTD 4CTD 2= 9
Learning
Data (facts only)
8.4% of 612 visitors left their email in 5 days; the ads cost $42 in total.
Insight
Demand clears the bar — but an email is not a wallet. Test willingness to pay next.

Try it with your own bet

Before you ask

Questions teams ask first

Isn’t this just another Kanban board?

No. validaid is an evidence tree: bets branch from your vision, each test ends in a verdict, and validated paths light up green. It’s built to kill bad ideas cheaply, not to track tickets.

Do we need to know pretotyping already?

No. Every card walks you through the method — believe / will / right-if, YODA scoring, data vs. insights — and the optional AI coach sharpens vague hypotheses for you.

How fast can the team start?

About 60 seconds. Sign in with a magic link, land on a ready-made sample evidence tree, and frame your first real bet in minutes.

What does it cost?

Free to start, forever. Pro adds the AI coach, unlimited projects and teammates from $10/user/mo — see the full breakdown on the pricing page.

See plans & pricing validaid.app/pricing

Stop guessing. Start collecting your own data.

Spin up your team in seconds. Frame your first bet, run a pretotype, and let the evidence decide.