Pitch-deck heatmaps are the most exciting data you will have access to during a seed round, and also the most likely thing you will misread. A partner spends 120 seconds on slide 7 and you start composing the term sheet in your head. The next partner spends 30 seconds on slide 7 and you reorder the slide.
Both reactions are usually wrong.
Heat is a useful signal, but only when read against a small set of patterns and a clear sense of what heat is not. Here is the field guide we wish someone had written for us before we shipped our first deck.
What heat is.
Per-slide dwell is the number of seconds a viewer spends on a given slide before advancing. Most analytics tools — DocSend, Arx, others — give you per-viewer totals, average dwell, peak slide, and a heatmap visualization (red = high dwell, cool = low).
Heat is, mechanically, an attention signal. It tells you that the eyeballs spent time on this slide. It does not tell you whether the time was spent because the slide was compelling, because the slide was confusing, because the partner had a meeting interrupt them, or because the partner was making notes for the LP they will brief on you tomorrow.
The job of reading heat is to separate those four interpretations using context that lives outside the heatmap itself.
The four patterns.
1. U-shaped (or "intro-team").
Lots of time on the first two slides, drop-off through the middle, recovery on the team slide. This is the most common shape and the most easily misread.
What it usually means: the partner is doing thesis triage. They opened the deck to check whether your sector fits their fund's current focus, scanned the middle for any disqualifiers, and looked at the team to decide whether you are the kind of founder they back.
What you should do: do not redesign the middle. The middle of the deck is doing exactly what it should — surviving a 90-second skim. If the same partner opens the deck a second time and dwells through the middle, that is the meaningful signal. The first U-shape is triage; the second visit is engagement.
2. Top-loaded (or "polite quit").
Heavy reading of the cover, the problem slide, and the insight slide — then nothing. The viewer simply did not finish.
What it usually means: one of two things. Either your sector is not a fit (and the partner stopped at the moment that became clear), or your problem framing is too generic and the partner concluded you were a "another tool for X" without earning the rest of the deck.
What you should do: if you see top-loaded heat from multiple partners in the same week, your problem slide and your insight slide are doing the wrong job. Rewrite them. The rest of the deck is fine.
3. Middle-heavy (or "model dive").
Light intro, heavy time on the slides that contain the business model, the unit economics, and the revenue traction.
What it usually means: this is the diligence pattern. The partner already knows the space, already buys the problem, and is doing the only work that matters to them — checking whether your business actually works at scale. This is a high-quality visit.
What you should do: prepare for the second call. The questions you will be asked will be about your assumptions on CAC, payback, retention, and pricing. Have the model open in another tab.
A middle-heavy visit from a partner you've not spoken to is a meaningful inbound signal. Reply to that partner within 24 hours, with a calendar link.
4. End-loaded (or "ask review").
Light dwell across the body, heavy time on the financials, the use of funds, and the ask slide.
What it usually means: the partner has already decided they are interested at a thesis level and is evaluating the deal terms. This is often the second-visit shape, after a first U-shape. It is also occasionally the shape of a partner who has been forwarded your deck by a colleague and is doing a quick "is this priced sensibly" check.
What you should do: if you have a term sheet from another fund, this is the moment to mention it (carefully, not in caps). If you do not, this is the moment to ship the model, before they ask.
What heat is not.
Three things heat is often mistaken for, and is not:
Heat is not interest.
A partner who spent four minutes on your deck may be interested. They may also be writing a memo to their colleagues explaining why they are passing. Heat without conversion to a meeting is, at best, mild signal. We have seen 6-minute deck visits convert to a polite pass; we have seen 90-second visits convert to a term sheet. The distribution is wider than founders think.
Heat is not thesis fit.
A partner can read every slide of your deck, find each one compelling, and still pass because your sector is not where their fund is currently writing checks. Heat tells you they read the deck. It does not tell you whether they can do anything about it.
Heat is not partner identity.
Most deck-sharing tools track by email. They do not, by default, tell you whether the visitor was the partner you sent the deck to, an analyst at the same fund, or the partner forwarding the deck to a colleague. A 30-second visit from "an analyst at Catalyst" and a 4-minute visit from "the partner you spoke to at Catalyst" are very different signals. Most tools give you only the email, which means a partner forwarding to an associate looks identical to the partner reading it themselves.
Arx pulls the LinkedIn / Crunchbase role data into the viewer panel where it can, so the analyst vs. partner distinction is visible at a glance. Other tools require manual lookup.
The two-visit rule.
The single most reliable heat signal is the second visit.
A partner opens your deck. They spend some amount of time. The first visit is mostly noise — they could be triaging, eating lunch, or distracted by Slack. Real signal is generated when the same partner opens the deck a second time, days later, often at an unusual hour.
A second visit is, in our experience, a 60% predictor of a follow-up meeting. A second visit with a return to slide 7 (or whatever your central-insight slide is) is closer to 80%.
If you see a second visit, follow up within 24 hours. The window is short.
Comparing across investors.
Founders often try to compare heatmaps across investors and infer "this investor read it more carefully than that one." Resist this.
Different partners have different reading habits. Some partners scroll fast through every deck regardless of interest. Some partners dwell on slide 1 because they are reading carefully. Some partners speed-read the entire deck on a phone in three minutes and still write a check.
What is comparable across investors is the shape of their visit, against the patterns above. The dwell numbers themselves are noisy.
What to do during a round.
A practical workflow we recommend to founders running a current round:
A note on AI summarization.
The newer deck tools — ours included — now offer AI summarization of viewer behaviour. "This partner spent the most time on slide 7, returned twice, and likely has questions about market sizing."
These summaries are useful, and we use them. But they are summaries of a noisy signal. Read the underlying heat yourself, at least until you have calibrated your own intuition. The AI is doing pattern-matching on the same four shapes; you should be able to do it too.
The summary.
Heat is a useful, noisy, over-interpreted signal. The four patterns — U-shaped, top-loaded, middle-heavy, end-loaded — capture most of what you need to react to. The single most reliable signal is the second visit. Comparing absolute dwell across partners is mostly a distraction.
The deck is not magic. The heatmap is not magic. Together, they tell you a little bit more about who is paying attention than you would otherwise know. That is enough — it is not everything.