Trace· 21 ways to learn something

The method library.
Pick the right instrument.

Twenty-one ways to move an assumption. Trace picks one for you based on risk and effort — but the library is always a click away.

01
Human

Structured Interview

To hear your customer describe pain in their own words.

Problem Fit45–60 min · ×8
02
Human

Jobs-to-be-Done

To reframe your problem around the underlying job.

Discovery60 min · ×6
03
Human

Mom Test Discovery

To check whether the problem is really felt, not polite.

Problem Fit30–45 min · ×10
04
Agent-assisted

Problem Storyboard

To map the current-state journey and locate true friction.

Discovery1–2 days
05
Agent-assisted

Landing Page Smoke Test

To probe demand before building anything real.

Solution Fit3–5 days
06
Human

Concierge MVP

To manually deliver the value before automating it.

Solution Fit2–4 weeks
07
Human

Wizard-of-Oz

To fake the AI while validating the UX.

Solution Fit1–3 weeks
08
Agent-assisted

Clickable Prototype

To test flow and clarity before any engineering.

Solution Fit2–5 days
09
Human

Pricing Elasticity Test

To find a price the market will actually pay.

Market Fit1–2 weeks
10
Agent-assisted

Fake Door Test

To measure real intent for a specific feature.

Solution Fit2–4 days
11
Agent-executed

Cohort Retention

To see whether usage survives the first week.

Market Fit4+ weeks
12
Agent-assisted

Channel Experiment

To test whether a channel can actually deliver.

Market Fit1–3 weeks
13
Agent-executed

Unit Economics Probe

To see if the numbers work at scale before scaling.

Market Fit2–4 days
14
Agent-assisted

Eval Set Build

To define what 'good enough' means for your model.

AI & Trust1 week
15
Human

Trust Calibration Study

To test whether users rely on AI the right amount.

AI & Trust2 weeks
16
Agent-assisted

Regulatory Scoping

To map the compliance path before spending on engineering.

Regulatory1–2 weeks
17
Agent-assisted

Agent Handoff Drill

To test multi-agent coordination and failure recovery.

AI & Trust3–5 days
18
Agent-executed

Competitor Teardown

To get oriented in the market without spending a week on it.

Discovery24 hours
19
Agent-executed

Market Sizing

To sanity-check whether the prize is worth chasing.

Discovery24 hours
20
Human

Willingness-to-Pay

To convert interest into a pricing signal.

Market Fit1 week
21
Agent-assisted

Sean Ellis PMF

To benchmark product-market fit against 40% threshold.

Market Fit1 week