Key takeaways
- The checklist asks whether a reader can find something in the deck, such as who has the problem or what a revenue figure means, not whether the news is good.
- Weights change with stage: the problem and team carry most points for an idea-stage deck, traction and financials for a deck with revenue.
- Partly counts as half an item. Unanswered items count as No, so the score only rises when someone has checked.
- A complete deck can still describe a weak business, and a sparse deck can describe a strong one. The score measures the document.
How to use the pitch deck grader
- 1Pick the company's stage. Idea means no product in use yet, pre-revenue means a product or pilots without revenue, post-revenue means paying customers.
- 2Open the deck next to the checklist. Give the deck a name if you plan to copy the report.
- 3For each item, answer Yes if a reader could find it without asking the founder, Partly if it is there but vague or missing a detail, and No if it is absent.
- 4Read the score, the points per section and the list of what to add first. The list is ordered by how many points each gap costs at this stage.
- 5Copy the report to share with a co-founder, or to keep beside your notes on the deck.
How the pitch deck grader scores a deck
Each section of the checklist has a set of items. A section's score out of 10 is the share of its items met, with Partly counting half. Its points are that score times the section's weight, divided by 10. The total is the sum of the points, out of 100.
The weights are the stage weights in the deck analyzer Delsivo builds on, which comes from Nautis's deck analysis pipeline. The checklist items under each section are Delsivo's own, written for this tool. The bands (under 50 gaps to fill, 50 to 74 mostly complete, 75 and above complete) are our own too, not an industry standard.
| Section | Idea | Pre-revenue | Post-revenue |
|---|---|---|---|
| Problem | 25 | 15 | 5 |
| Solution | 20 | 20 | 10 |
| Market | 20 | 15 | 10 |
| Team | 20 | 10 | 5 |
| Roadmap and milestones | 10 | 5 | Not scored |
| Ask and use of funds | 5 | 5 | Not scored |
| Validation | Not scored | 20 | Not scored |
| Go-to-market | Not scored | 10 | Not scored |
| Traction | Not scored | Not scored | 30 |
| Financials | Not scored | Not scored | 20 |
| Competition | Not scored | Not scored | 10 |
| Path to scale | Not scored | Not scored | 10 |
Example: an idea-stage deck that meets three of the four Problem items and half of one Team item scores 25 × 0.75 + 20 × (0.5 ÷ 3) = 22.1 out of 100.
Why traction needs a definition, a period and a date
The traction items ask for more than a number because a number alone can't be checked. "$40K revenue" could be one month or a year, booked or collected, recurring or a one-off project. A reader who has to guess will either discount the figure or ask, and both slow the conversation down.
The same goes for validation before revenue. Twenty pilots means something different when twelve of them are unpaid, so the checklist asks whether paid and unpaid commitments are separated.
| Vague | Checkable |
|---|---|
| Strong revenue growth | MRR from subscriptions: $22K in June 2026, $31K in September 2026 |
| 20 enterprise customers | 8 paying customers on annual contracts and 12 unpaid pilots, as of October 2026 |
| Huge market | 4,000 mid-sized logistics firms in our first region × $12K a year = $48M |
| Great retention | Net revenue retention of 104% over the 12 months to September 2026, measured on customers who started before October 2025 |
What the grader does not do
- It doesn't read your file. You answer the checklist, so the result is only as careful as the answers.
- It doesn't check whether figures are true, consistent or sourced well. Completeness is not accuracy.
- It doesn't judge design, writing style or slide count. A plain deck with every item scores the same as a polished one.
- It doesn't say whether to invest, and it isn't investment advice. It grades the deck as a document, never the company as an investment.
- It uses one checklist for every sector. A hardware or health company may need items it doesn't list, such as certification or trial plans.
How Delsivo reads a deck
Delsivo's deck analysis is planned for the pilot. By design it reads the PDF itself, pulls out company facts and claims with the slide number and the exact quote, flags slides it couldn't read well, and computes scores in code from criterion scores using the same stage weights. Deck quality is kept separate from thesis fit, so a well-made deck never reads as a good fit on its own. People review every result. See how the deck analyzer works or how slide citations work.