NeoNumber evaluates ideas before you try them and benchmarks their results afterward, so each decision helps you make the next one.
Review
See how well an idea stands before you act. Get the reasons, its biggest uncertainty and a useful next test.
A judgment of the plan, not a chance of success.
See a scored ideaReview
74 of 100 for how this idea stands today.
A judgment of the plan, not a chance of success.
Range 59–89, as unknowns get settled.
A useful next test
Invite a small customer group and check referral purchases before expanding.
A made-up example that shows how the method works. Nothing here was measured.
Result benchmark
See what happened, then compare performance with comparable work. Your measures and tradeoffs stay visible alongside the benchmark.
Without enough comparable results, you see your measurements and what is missing. Your original Review never enters the calculation.
How the two-part evaluation worksResult benchmark
Illustrative Result scorefor measured performance.
Compared with 100 simulated campaigns with the same channel, offer band and follow-up window.
How this benchmark score is calculated
Each measure is compared with the same measure in the reference set. We combine those positions using fixed weights. Your target does not affect the score.
These references are simulated to explain the method. Each range below contains 100 equally spaced values; the same numbered values belong to the same simulated record.
- Cost per new customer: $10–$49.60 in $0.40 steps
- Orders refunded: 1%–10.9% in 0.1-point steps
A made-up example that shows how the method works. Nothing here was measured.
your record
Keep the idea, the judgment and the outcome connected. Preserve versions, see corrections and take the lesson into your next decision.
Reviews and Results stay separate. They are never averaged into one number.
your record
Review
74 of 100
A judgment of the plan, not a chance of success.
Range 61–83, as unknowns get settled.
Result
Illustrative Result score
80 of 100
for measured performance.
Compared with 100 simulated campaigns with the same channel, offer band and follow-up window.
Your next decision:
Keep the customer-video approach. Test whether it works for a new audience before increasing the budget.
Review scores judge the idea. Result scores benchmark its performance. Compare like work within the same method; the two scores are never averaged.
A made-up example that shows how the method works. Nothing here was measured.
Follow one campaign from plan to Result
Put your next move to the test
Pick an example. See what you could send, the Review you could get and what you could learn from the Result.
A made-up example that shows how the method works. Nothing here was measured.
Let customers make the ads.
After: your Result
After 30 days and the refund follow-up
80 of 100 for measured performance.
Illustrative Result score, compared with 100 simulated campaigns with the same channel, offer band and follow-up window.
| What was measured | Outcome |
|---|---|
Cost per new customer | $18 |
Orders refunded | 3% |
How this benchmark score is calculated
Each measure is compared with the same measure in the reference set. We combine those positions using fixed weights. Your target does not affect the score.
| Measure | Position out of 100 | Weight |
|---|---|---|
Cost per new customer | 79.5 | 50% |
Orders refunded | 79.5 | 50% |
The weighted result is 79.5, displayed as 80. It is an index of the measures above, not the percentage of whole projects you beat.
These references are simulated to explain the method. Each range below contains 100 equally spaced values; the same numbered values belong to the same simulated record.
- Cost per new customer: $10–$49.60 in $0.40 steps.
- Orders refunded: 1%–10.9% in 0.1-point steps.
Low acquisition cost and few refunds put this campaign above the reference middle on both measures. Its $25 target is useful context, but does not determine its score.
Customer report from ad spend and first-order records.
Your next decision: Keep the customer-video approach. Test whether it works for a new audience before increasing the budget.
Review scores judge the idea. Result scores benchmark its performance. Compare like work within the same method; the two scores are never averaged.
See three ideas scored
Same goal, $2,000, 30 days. Which idea would you back?
A meal-planning subscription wants 80 new customers in 30 days, with a $2,000 budget. Here’s how three approaches could look before anyone spends it.
A made-up example that shows how the method works. Nothing here was measured.
Discount to everyone
45 of 100 for how this idea stands today.
A judgment of the plan, not a chance of success.
Range 30–60, as unknowns get settled.
Where this idea is strongest: A clear offer and a campaign the team can run.
Biggest uncertainty: A broad audience and deep discount leave little reason to expect lasting customers.
A useful next test: Test the offer with one specific audience before spending the full budget.
What was sent and how it scores
Spend $2,000 on social ads offering 40% off our $30 monthly meal-planning subscription. Target all US adults for 30 days. Aim for 80 new customers. We have not tested the message or discount yet.
| Dimension | Judgment |
|---|---|
Purpose | 65 |
Support | 25 |
Feasibility | 70 |
Value | 40 |
Downside | 25 |
Project/campaign method 1.0: the five judgments have equal weight. All five are assessed; the range includes a 5-point disagreement reserve and 10 points for future uncertainty.
Ask customers to bring a friend
74 of 100 for how this idea stands today.
A judgment of the plan, not a chance of success.
Range 59–89, as unknowns get settled.
Where this idea is strongest: A relevant signal from customers, with spending tied to paid sign-ups.
Biggest uncertainty: Requests to share may not translate into enough referrals to reach new buyers.
A useful next test: Invite a small customer group and check referral purchases before expanding.
What was sent and how it scores
Use a $2,000 budget over 30 days to offer existing US customers a $10 credit after a referred friend pays for our $30 monthly meal-planning subscription. Aim for 80 new customers. Twenty customers have asked to share it. Cap credits and check for duplicate accounts.
| Dimension | Judgment |
|---|---|
Purpose | 85 |
Support | 65 |
Feasibility | 70 |
Value | 75 |
Downside | 75 |
Project/campaign method 1.0: the five judgments have equal weight. All five are assessed; the range includes a 5-point disagreement reserve and 10 points for future uncertainty.
The ranges overlap, so 74 isn’t proof referrals will win. What each card gives you is the next test to run before committing the budget.
See the third idea
Show busy parents the benefit
68 of 100 for how this idea stands today.
A judgment of the plan, not a chance of success.
Range 53–83, as unknowns get settled.
Where this idea is strongest: A specific audience, a relevant problem and a concrete demonstration.
Biggest uncertainty: Interview interest does not establish willingness to pay.
A useful next test: Send a small group to the real checkout and measure purchases, not just clicks.
What was sent and how it scores
Spend $2,000 over 30 days showing US parents a short demo of our $30 monthly meal-planning subscription. Aim for 80 new customers. Ten interviews suggest weeknight planning is a problem; we have not tested whether the demo leads to purchases.
| Dimension | Judgment |
|---|---|
Purpose | 80 |
Support | 55 |
Feasibility | 75 |
Value | 70 |
Downside | 60 |
Project/campaign method 1.0: the five judgments have equal weight. All five are assessed; the range includes a 5-point disagreement reserve and 10 points for future uncertainty.
We check our own scores
Every Review is kept beside what actually happened. As outcomes come in, we publish whether higher-scored plans really did better, per method version, with the counts. Until there are enough, we say so.
How the two-part evaluation worksUse NeoNumber inside your product
Request Reviews and report Results from your own tools through the API, let AI assistants work with your Ideas through the local MCP server, and embed public scorecards on your site once the workspace owner agrees to publish them.
See what you can build