The 9 Investor Questions That Destroy Rounds — Friday Notes Audio
The Pattern No One Talks About
9 of the 14 market-size claims in a typical Series A deck fail diligence. Not because the data is fabricated. Not because the founders were dishonest. Because the evidence behind each claim — the original source, the methodology, the publication date, the category match — cannot be verified by an analyst in the time available.
This is the pattern we see across 214 verdicts and Sentinel audits: the deal does not collapse at the meeting. It collapses afterward, when the LP update or the closing memo needs to include the sourcing, and someone traces the numbers back to the beginning.
“The round collapsed not because the claim was wrong — but because it could not be traced to a primary source in the two minutes available before the next question.”
The 9 Questions, Each Explained
These are not hypothetical risk areas. Each one maps to an investigation trigger observed across Sentinel audits and PRISM due-diligence stress tests. They are ordered by frequency of occurrence — the first three account for the majority of late-stage failures.
“Where does this market size number come from, and what methodology did they use to arrive at it?” The most commonly asked and most commonly failed question. Top-down percentages (“we are taking 1% of a £4B market”) are the trigger. A bottoms-up model built from segment × conversion × ARPU is the only defensible answer.
“Which benchmark report did this retention figure come from, and does the product category actually match?” Founders pull the nearest-looking benchmark and apply it regardless of category fit. An analyst knows the difference between mobile social, B2B SaaS, and desktop productivity retention curves. The categories look similar on paper; the underlying user behaviour is fundamentally different.
“What churn assumption is this LTV built on, and where did that churn number come from?” LTV is the most reverse-engineered number in any deck. Founders often build it from an industry average rather than actual cohort data. An investor who has seen 200 decks can recognise the median B2B SaaS churn figure being used as a first-principles output.
Q4 — Competitive differentiation: “Have you looked at [named competitor] — they launched three months ago and appear to be in your space?” Any “only” or “first” claim is verified against Crunchbase, ProductHunt, and a targeted operator search within 60 seconds. An unnamed competitor is a question, not a comfort.
Q5 — CAC channel assumption: “What is this CAC based on, and is that channel actually open to you at this stage?” CAC figures built from aspirational channels (national TV, enterprise sales at Y CAC) before you have operated in them are among the most common financial-model failures.
Q6 — Customer evidence quality: “How many paying customers is this NPS score from, and how were they selected?” An NPS of 72 from 14 customers who were selected by the founder is not a statistically meaningful data point. Investors know this. The question probes whether the sample is representative or self-selected.
Q7 — Regulatory status: “What specific licences or approvals does this require in [jurisdiction], and what is the current status?” Particularly acute in fintech, health, and AI governance. A business that needs regulatory approval it has not yet begun pursuing has a CAC problem it hasn’t noticed yet — the cost of acquisition includes the cost of compliance.
Q8 — Team credential depth: “Has the team done this before, and if not, what replaces that experience?” Not a disqualifier; a prompt for evidence. Domain expertise, domain advisors with real involvement, or direct market experience all answer it. Vague team slides do not.
Q9 — IP ownership: “Is all the intellectual property clean, and do the assignment agreements reflect that?” Late discovery of IP gaps is a closing killer. Any code written by a founder before company incorporation, any contractor without a clear work-for-hire agreement, is a potential liability that surfaces in legal diligence.
How Analysts Actually Investigate
The investigation is faster and more systematic than most founders expect. An analyst with access to your deck will run the following checks before the first partner meeting, typically in 30–60 minutes.
Market size: pull the cited report (if one is cited), verify the exact figure, check the methodology section, and confirm the product sub-category matches yours. If no report is cited, search for the figure in the obvious databases. If it cannot be found in 90 seconds, it is flagged as unverifiable.
Retention metrics: identify the benchmark in the deck, locate the source report, check the product category classification in the methodology section, and compare it to your product category. The mismatch between “mobile apps” and “enterprise workflow SaaS” is invisible to a founder and obvious to an analyst with the report open.
If a claim cannot be traced to a named, dated primary source within 90 seconds, it will be flagged as unverifiable. This is not rigidity — it is the operational reality of LP-update memos, which require sourcing for every material claim. A claim that cannot be sourced is a claim that cannot appear in the memo.
The Citation Standard
A citation is not a link to a blog post that mentions a statistic. A citation is a reference to a primary source document — the original research, the regulatory guidance, the audited filing — with enough information to allow anyone to verify the exact figure in its original context.
The minimum citation standard for a fundraising deck: named source (author or organisation), document title, publication date, specific page or section, and the methodology used to arrive at the figure (if it is a derived estimate rather than a direct measurement).
“A secondary source that cites a primary source that no longer says what the secondary source claims is the most common form of citation laundering. Analysts trace back to primaries. Always.”
Source laundering — citing a secondary source rather than the primary — is the single most common citation failure in Sentinel audits. A blog post summarising a Gartner report, a deck citing another startup’s investor update, a press release interpreting a research paper: all of these are secondary sources that may have misquoted, rounded, or recontextualised the original data. An analyst will trace back. The question is whether the primary source confirms the claim.
Pre-Diligence Self-Audit
The most effective preparation for investor diligence is to conduct it yourself, in advance. Run the analyst’s checklist against every quantified slide before you send the deck.
Get the Friday Notes dispatch
Due diligence intelligence, verification frameworks, and what we’ve learned across 214 verdicts. Every Friday at 09:00 GMT.
- TAM: Build a bottoms-up market size model. Remove all top-down percentage claims. For every TAM figure, identify the methodology and confirm the primary source document exists and is current.
- Benchmarks: For every benchmark (retention, NPS, CAC, churn), identify the source report, open it, confirm the product category in the methodology section exactly matches your product.
- LTV: Trace LTV back to the churn assumption. If the churn figure is from industry data rather than your own cohort data, say so explicitly and note the confidence interval.
- Competitors: Search Crunchbase, ProductHunt, and LinkedIn for every adjacent player. If you have named competitors in the deck, your analyst will check whether you have missed any.
- CAC: Confirm the CAC figure is from a channel you have operated in or from a directly comparable company. Flag it if it is aspirational.
- Regulatory: Confirm the specific licence or approval status in your target market, with a named source and timeline.
- IP: Confirm all code, designs, and IP have a clear chain of title from creator to company, in writing.
Building the Diligence Layer
A diligence-ready deck is not one that is harder to challenge. It is one where every material claim has a traceable evidence chain behind it, and where the founder has already confronted the questions before the meeting.
The Sentinel verification service exists precisely for this: a named human analyst reviews every quantified claim, applies the citation standard, tests each one against the 9 questions above, and delivers a signed verdict before the deck leaves the building. The result is not a perfect deck — it is a deck whose weaknesses are known, documented, and addressed before an LP’s analyst finds them.
The founders who close the fastest are not the ones whose decks are impervious to challenge. They are the ones who already know every weakness, have a documented answer for each one, and deliver those answers in the meeting before the question is asked. The dynamic shifts from defence to presentation. That shift is worth every hour of pre-diligence work.
96% of founders who ship to market after a GO verdict from ThriveFinity do so within 60 days. Not because the verification removes uncertainty — it doesn’t. Because it removes the uncertainty that was manufactured by unverified claims, and replaces it with a clear, evidence-graded picture of what is known, what is inferred, and what must be tested. That is what investors are looking for. Give it to them before they ask.