plumb
Sign inEvaluate an idea

The method

How Plumb measures an idea.

Everything that decides your verdict is on this page. The AI judges; a published formula does the math.

Step 1

The 10-point evaluation

Claude scores each of the 10 points from 0 to 10 using the written descriptions below, and says how much proof backs each score up. Points that matter more for whether a business survives count for more (weights).

01

Problem-Solving Skill

Real Problem

weight ×1.5

Does this fix a problem people urgently need fixed (a painkiller), or is it a nice-to-have (a vitamin)? Proof comes from conversations, complaints, reviews and forums.

0
Nice-to-have; no sign anyone struggles with this.
5
The problem sounds real and a few people complain about it, but they cope with workarounds.
10
A painful, frequent, costly problem, backed by conversations or public complaints; people already go out of their way to fix it.

02

Customer Understanding

Customer

weight ×1.5

Who has the problem, who pays, and how do they describe it? Based on real conversations, not guesses.

0
The customer is 'everyone' or not named; nobody knows who pays.
5
A customer group and payer are named, but based on guesses.
10
A narrow, specific group; you know who uses it and who pays, from repeated conversations.

03

Value Proposition Clarity

Value Proposition

weight ×1.0

Can you say in one sentence why someone should choose this over what they do today, including doing nothing?

0
Can't say why it's better than what people do now.
5
A clear pitch, but not very different from existing options.
10
A sharp one-sentence pitch with a clear, meaningful advantage over every alternative, including doing nothing.

04

Market Sizing & Timing

Market + Timing

weight ×1.0

How many people could buy this (market size: TAM, SAM, SOM), and what recent change makes now the right time: new technology, new rules or new habits.

0
A tiny or unknown market, and no reason why now.
5
A reasonable market; nothing special about the timing.
10
A large, growing market you can actually reach, with a clear recent change working in your favour.

05

Product-Market Fit Detection

PMF

weight ×1.0

Do customers pull the product towards them (product-market fit)? Signs: they keep using it, tell others, and would be very disappointed if it disappeared.

0
No users or signals yet, and no plan to measure them.
5
Early users or a waitlist, with some good signs or a clear plan to measure them.
10
Customers keep using it, bring in others, and come looking for it.

06

Execution Speed & Focus

Execution + Focus

weight ×0.8

Can a first version ship quickly while staying focused: one customer, one problem, one way of finding customers, one number to track?

0
Trying to do too much; a long time before anything ships.
5
Sensible scope; a first version is a few months away.
10
A very narrow first version that can ship in weeks, with one clear measure of success.

07

Business Model & Unit Economics

Business Model + Economics

weight ×1.5

How money flows: prices, profit on each sale (margins), the cost to win a customer (CAC) against what they pay over time (LTV), and monthly spending (burn). Does each customer make money or lose it?

0
No clear way to make money, or each customer costs more than they bring in.
5
A way to make money is defined; the numbers seem possible but aren't proven.
10
Proven prices, healthy profit on each sale, and each customer brings in far more than they cost to win.

08

Distribution & Sales

Distribution

weight ×1.3

How you'll reach customers (distribution): content, messaging people directly, partners, ads, and what it costs to win each one.

0
No way of reaching customers named; 'build it and they will come'.
5
Ways to reach customers are named but untested; a rough guess at the cost.
10
A proven way to reach customers again and again at a known cost, or special access to them.

09

Adaptability & Moats

Adaptability + Competitive Advantage

weight ×0.9

Can you change course when the evidence says so, and what stops others copying you (moat): a known brand, users who bring in more users (network effects), being hard to switch away from, data only you have, or doing it much better.

0
Easy to copy; nothing protects it.
5
Some advantage, like speed or knowing the niche well, but others could copy it in time.
10
An advantage that grows stronger over time, such as network effects, data only you have, or customers who can't easily switch.

10

Team Building & Systems

People & Systems + Capital Discipline

weight ×0.8

The right mix of people, written-down ways of working, a business that can run without the founder doing everything, and careful spending.

0
Key skills missing; no plan for how long the money lasts.
5
Core skills present with known gaps; money for a limited time.
10
A team with complementary skills, documented processes, and spending that grows only as revenue grows, with plenty of runway.

Step 2

The 7 signs it can grow (indicators)

Each is rated 1 to 5 and combined into the Can it grow? (scalability) score. For Manual Labor and Risk Level a high rating is bad, so those are inverted.

IndicatorWeight

Pain Intensity

1 = a mild annoyance, 5 = an urgent, costly problem that keeps coming back.

×1.3

Revenue Impact

1 = little or unclear money to be made, 5 = each customer pays a lot, or it directly earns the customer money.

×1.2

Manual Labor

higher is worse

1 = almost everything is automatic, 5 = every customer needs a lot of hands-on work.

×1.0

Tool Access

1 = needs tools, data or equipment that are unavailable or very expensive, 5 = everything needed is easy to get.

×0.8

Repeatability

1 = every job is custom, 5 = the same product or service is delivered the same way to many customers.

×1.1

Willingness to Pay

1 = people expect it to be free, 5 = people already pay for alternatives or have money set aside for this.

×1.3

Risk Level

higher is worse

1 = low risk from laws and rules, technology, the market or depending on others, 5 = very high risk.

×1.0

Step 3

The evidence rule

Hope is free; proof isn't. Each score is lowered when the claim behind it has no proof, then the scores are averaged into the overall score out of 100 (viability).

No proof yet (assumed)

×0.85

The founder believes it, but nothing submitted backs it up.

Some proof (anecdotal)

×0.95

A few conversations, personal experience or early signals.

Proven (validated)

×1.00

Real numbers, repeated customer conversations or paying users.

Checked against the web. Before scoring, Plumb searches the web for competitors, market data, demand signals and pricing, and checks your key claims. Independent evidence can raise a criterion to validated; a claim the web contradicts lowers it. Every finding links to its source, and only pages the search actually returned can be cited.

Step 4

The verdict

A fixed formula (not the AI) turns the scores into a verdict, so the same judgments always produce the same result.

Strong Go

Worth doing? (viability) ≥ 70 and Can it grow? (scalability) ≥ 60, with no hard gate triggered.

Conditional Go

Worth doing 50–69, or a Strong Go score held back by a low Can it grow score.

Pivot Recommended

Worth doing 35–49, or a hard gate on an otherwise passing idea.

No-Go

Worth doing below 35.

Hard gates

If Pain Intensity or Willingness to Pay is rated 2 or lower, the verdict is Pivot Recommended at best, however strong the rest looks. Vitamins get cancelled, and ideas nobody will pay for don't become businesses. The idea has to change shape, not just gather evidence.

Rubric version 2026-10-v4

See how your idea measures up.

About 4 minutes per evaluation.

Evaluate an idea →