Guide
Pitch deck for pre-seed: what to show when you have little or no revenue
With little revenue to show, investors test the plan instead: the ask, the burn behind it, the market, the unit economics and the model that holds them together.
By Amer Habeeb. Updated . 8-minute read.
Every number below comes from Fernbrook Analytics, a fictional company. You can change its price in the live model on the home page.
The figures in the blocks marked “Real work, shown with permission” are the one exception. They come from The Real Sharkz, a real pre-seed deck shown with the founder's written permission, quoted exactly as published.
What changes at pre-seed
At seed, a deck can lean on revenue that already exists. At pre-seed it usually can't. There may be a product, a waitlist, a few pilots, or only a plan. Investors know that, so they test the plan instead.
With no history to point to, almost every number on your slides is a forecast, and a forecast is only as believable as the arithmetic behind it. That arithmetic lives in a short model.
This guide works through it on Fernbrook Analytics, the fictional sample company on this site. Fernbrook is raising a seed round, so its amounts are larger than most pre-seed plans. The checks are the same at any size.
Real work, shown with permission
Where it helps, I also point to a real pre-seed deck, The Real Sharkz, whose product slide states traction as zero until the first real campaign. You can page through all 16 slides of The Real Sharkz deck.
Put the ask up front
The ask is three things: the amount, the months it lasts, and what you will have done when it runs out. Many decks keep it for the last slide, so the investor reads every other slide without knowing the size of the conversation.
Put it near the front, in one sentence: the amount, the runway in months, and one or two milestones written as numbers your model produces. “Grow the business” is not a milestone. A customer count by the end of a named year is.
Real work, shown with permission
The Real Sharkz deck puts the ask on the cover. Slide 1 states ₹2.07 Cr and “24 months, fully budgeted” before any story, so an investor knows the size of the conversation from the first second. A cover ask works only when a later slide shows the working, and in that deck slide 15 does.
The test is simple. Could an investor repeat your ask back to you after the first two slides?
Burn × months = the ask
At pre-seed the amount should come from one calculation: monthly burn times the months you need to reach the next milestone, plus a margin you can name. Then the amount is arithmetic an investor can check.
Burn is the cash that leaves the bank each month after customers pay you. Take it from the cash flow, not the income statement, and take it year by year, because burn usually rises as you hire.
Real work, shown with permission
The Real Sharkz deck shows the calculation in its simplest form. Slide 15 states ₹8.63 lakh a month × 24 months = ₹2.07 Cr, and its five use-of-funds lines add up to 100%.
On your own deck, multiply monthly burn by the months on your ask slide. If the result is not your raise, one of the three numbers is wrong.
Count the market from your own tables
A pre-seed market slide often quotes an industry total from a research report. That says the category is big. It does not say how many customers you could sell to, or how much of that your plan needs.
Count it from the bottom up instead: the customers you could sell to, times your price, times 12 months. Then put your plan next to it. The share your plan needs is the number an investor can argue with.
Real work, shown with permission
The Real Sharkz market slide, slide 8, is counted the same way, from the founder's own tables: listings × completion rate × average raise. It also states the hard constraint, about 964,000 active backers by Year 3, instead of hiding it. Naming the constraint yourself reads as confidence.
Unit economics, with the failing case shown
With little or no revenue, unit economics are the closest thing a pre-seed deck has to evidence. The cost to win one customer is sales and marketing spend divided by the customers won that year. Payback is that cost divided by the gross profit one customer brings in each month. Use gross profit, not revenue.
The step most decks skip is the failing case. Every unit-economics slide rests on one assumption, usually the cost to win a customer. Show what happens if it is wrong. An investor will run that case anyway, so let them see you ran it first.
Real work, shown with permission
The Real Sharkz unit-economics slide, slide 12, puts the failing case next to the working one. At the planned ₹52 blended acquisition cost, lifetime value is 3.2 times cost. At ₹250, when every backer comes from ads, it is 0.66 times.
Answer the hard question early
Every pre-seed company has one question an investor asks first. For a marketplace, how the first side gets supplied. For a regulated product, whether it is allowed. Leave it for the meeting and every slide is read with it in mind.
Find the one answer that, if wrong, would end the conversation, and put it near the front, in the headline.
Real work, shown with permission
The Real Sharkz deck answers its hard question on slide 4. A reward-based platform stands or falls on whether it sells securities, so the slide says what a backer gives, what a backer gets, and why no security is offered.
Make the team slide and the costs agree
At pre-seed the team is most of what an investor is backing. The team slide is also where a common mistake hides: it and the cost line are written weeks apart, and nobody multiplies them. People on payroll × average salary is the floor of your payroll, before taxes, software or rent. If it sits above the operating costs in your model, one slide is wrong.
Run this multiplication every time the team slide changes. To have it and the other checks in this guide run on your own deck, send me the deck. You get one written page back, and every point names the slide, the number and the fix.
Keep a short model behind the deck
A pre-seed model does not need fifty tabs. It needs one assumptions tab, three years, and three statements that agree. Every number on your slides should be a cell in it. Then, when an investor asks “what if you charge less?”, you can answer in the meeting, not a week later.
To build one from a blank sheet, follow how to build a financial model for a startup. If an investor is likely to ask for the model, send it with the deck.
A 12-slide pre-seed outline
Twelve slides is enough for most pre-seed raises. This order puts the ask and the hard question near the front:
- Title: what the company does, in one line
- The ask: amount, months and milestones
- Problem: who has it, and what it costs them
- The hard question, answered in the headline
- Product: what exists today, including zero traction if that is the truth
- Why now
- Market, counted bottom-up
- Business model: price and gross profit per customer
- Unit economics, with the failing case
- Team, with salaries that match the costs
- Financials, with the balance check
- Use of funds: burn × months, adding up to the ask
The rebuilt Fernbrook deck uses ten slides in much the same order: see all ten slides of the Fernbrook sample deck. For the seed-stage version of these checks, read the six questions investors test in a seed deck.
Check yours
- The ask states the amount, the months and one milestone as a number.
- Monthly burn × those months, plus a named margin, equals the amount.
- The market is customers × price × 12, with the share your plan needs.
- The cost to win a customer and the payback are on a slide, with the failing case.
- People × average salary fits inside each year's costs.
- Every number on a slide is a cell in a model whose balance check reads 0.00.
Have your pre-seed deck read
If you would rather have the deck rebuilt from your own numbers, see what the $95 pitch deck package includes.
Send the deck you have, and your model if there is one. Within 2 business days (India time) you get one written page: where investors will push hardest, two fixes you can make today, and what already holds up. It is a consistency check, not an audit.
Want to see the format first? See the full sample
How this guide was made: written with Claude from the method I use on every read and the Fernbrook sample model. The self-test recomputes every Fernbrook number on this page; other figures link to their source.