How we grade a plan.
Every plan goes through the same review before it's published. The reviewer didn't write the plan, and the job is to find what's wrong with it.
Why we review everything.
Early on, our quick screens kept looking better than the businesses turned out to be. Rebuilding them month by month, with local prices and a hired manager in the costs, brought the numbers down every time. So now nothing is published until a second person has checked it.
| Idea | First screen | After rebuild and review |
|---|---|---|
| Private client company | 15.5% | |
| Nashville make-ready | 23% | |
| Pest control, Houston | 25% IRR at a 4.0x sale | |
| Hydro-excavation | Negative |
The eight checks.
| Check | What the reviewer does | To pass |
|---|---|---|
| V1 Arithmetic | Reruns the model scripts and compares every table and headline number to the output. Confirms the balance sheet balances and sources equal uses. | Everything matches, or it gets fixed. |
| V2 Driver audit | Lists the ten inputs that move profit most, such as price, volume, wages, rent, churn and the exit multiple. Records each one's source and date. | At least six of the ten are sourced, and none of the top three is a guess. |
| V3 Local prices | Looks for a primary local source for the top three inputs: published price lists, competitor websites, job postings, listings, government rate schedules. | Attempted for all three, with links and dates. |
| V4 Benchmarks | Compares margins, revenue per employee, churn and acquisition cost against public filings, franchise disclosures and trade data. | Anything more than 25% off is explained or corrected. |
| V5 Return realism | Recomputes investor returns with a sale, without one, and over ten years. Flags anything over 30% IRR. | High returns are explained, and the no-sale case is shown. |
| V6 Licensing | Checks every license, permit and ownership limit on the state's own website or statute. | Every required license is named, or listed as an open item. |
| V7 Consistency | Checks that the funding amount, dates, unit counts and names agree everywhere in the plan. | No conflicts. |
| V8 Red team | Writes the five strongest reasons an investor would decline. | Each one has an answer, a fix, or a stop sign. |
How every number is labeled.
Each important input in a plan carries one of three labels, so you can see how much weight it will bear.
- Verified
- From a primary source, dated within the last two years. A price list, a government schedule, a filing.
- Secondary
- From an aggregator, an article or a vendor's own marketing.
- (U)
- Unverified. An estimate we haven't been able to confirm yet. Marked wherever it appears.
What the grades mean.
Grade A
Passes all eight checks, and the top inputs come from primary sources.
Published.
Grade B
Passes the arithmetic, return and consistency checks. Remaining gaps are printed in a box on page one.
Published, with the gaps disclosed.
Grade C
Fixable problems: numbers off, inputs unverified, licensing unclear.
Goes back for one revision. A second C becomes an F unless a partner overrides it.
Grade F
Verified inputs break the plan, or the base-case investor return is under 10%.
Set aside. Not published or sold.
The order things happen.
- Research and screen. A market file and a nine-point test: size, margin, recurring revenue, competition, whether a hired manager can run it.
- Full plan. Fifteen to thirty pages and a monthly five-year model.
- Review. The eight checks above, by someone who didn't write it.
- Partner decision. Publish, revise, or set aside.
- Published. On the board, with its grade and known gaps.
What we don't do.
- We don't give personal investment advice. The plans are general research, sold to anyone at the same price.
- We don't take a percentage of money you raise, and we don't introduce investors.
- We don't sell shares in any business we write about.
- We don't replace your own diligence, your lawyer or your accountant.
See a reviewed plan for yourself.
The pest control plan shows the reviewer's verdict, the known gaps and every downside case.