Blog, Pitch decks, 6 min read
What investors actually look at in a pitch deck
Most decks are read in under four minutes. Here is what gets read in that time, and what gets skipped.
A pitch deck does not get read the way you wrote it. You built it front to back over three weeks. An investor opens it between two meetings, scrolls fast, stops twice, and decides whether to reply. DocSend has been publishing data on this for years and the pattern has not really changed: the average deck gets a few minutes of attention, and that time is not spread evenly.
So the question is not "is my deck good". The question is "what survives a three minute scroll".
The four slides that get read
In practice, attention clusters on the same handful of slides.
- The problem, because it tells them whether this is a market they care about.
- The product, because they want to see the thing, not read about it.
- The traction or numbers, because this is where the reading slows down.
- The team, because at early stage they are mostly betting on you.
Everything else is supporting material. Competition grids, go to market plans, five year projections: useful in the meeting, skimmed in the scroll. That does not mean cut them. It means do not spend your best thinking on them while the traction slide is a screenshot of a spreadsheet.
One idea per slide, and say it in the headline
The most common fixable problem we see is a slide whose headline is a label. "Market" tells the reader nothing. "Indian D2C beauty is 8 billion dollars and growing 20 percent a year" tells them something they can react to.
Write every headline as a sentence with a verb in it. If you can read only the headlines top to bottom and follow the argument, the deck works. If the headlines are a list of nouns, the deck needs the speaker to make sense, and the speaker is not in the room when it matters.
Numbers are a design problem, not a spreadsheet problem
Founders paste tables. Tables are where good numbers go to die. A 40 percent month on month growth rate in a table reads the same as a 4 percent one. Drawn as a chart with the one comparison you want made, it lands in a second.
Pick the single comparison each chart exists to make, then remove everything that is not that. Two lines, not eight. One colour for the thing you want looked at, grey for the context.
What to cut
- The mission statement slide. Nobody has ever funded a mission statement.
- Long bios. Name, what they did before, why it matters here. Three lines.
- The hockey stick with no assumptions behind it. It reads as a guess, because it is.
- Anything you would skip past yourself when presenting it live.
The version you send is not the version you present
If you are in the room, the slides can be sparse because you are the narration. If you are sending it, the deck has to narrate itself, which usually means a bit more text and a clear reading order.
Keep two versions. A live version with fewer words and bigger type, and a send version with enough context to survive being forwarded to a partner you have never met. Most decks fail because they are one file trying to do both jobs.
A quick test before you send
Open your deck on a phone. Scroll it at the speed you would scroll a link a colleague sent you. Then close it and write down what you remember. If what you remember is not the argument you wanted to make, the problem is the deck, not the reader.