Why Two Good Investors Score the Same Deck Differently: Thesis as a Re-Weighting, Not a Disagreement

A founder forwarded me two rejection emails last month, nine days apart, about the same deck. The first, from a pre-seed angel, said the company felt "too early — no real traction yet, but a strong team." The second, from a growth investor she'd been introduced to further down her list, said close to the opposite: "Team seems capable, but we'd need to see the metrics hold at scale before we could underwrite this." Same deck. Same numbers on every slide. Two investors reading it as if it belonged to two different companies.

It wasn't two different companies. It was one company, read through two different theses. And — this is the part that doesn't fit the folklore — both investors were right.

Founders are taught that a good deck earns a consistent yes and a bad deck earns a consistent no, as if a score were a property of the document, like a word count. It isn't. A score is a property of the deck and the fund reading it, and almost nobody explains the second half of that equation.

Why doesn't a pitch deck have one correct score?

Because the deck carries several independent pieces of evidence, and a single number can't represent all of them without someone choosing how much each one counts.

Every investor — whether they'd phrase it this way or not — is reading a deck across a small, consistent set of underlying signals: who is building it, how big and urgent the problem is, how defensible the product is, how fast it's growing, how the unit economics hold up, how exposed it is to failure, and how convinced the reader is that this specific team, in this specific market, is the one that wins. NUVC's own scoring model formalises this into seven dimensions — team, problem/market, solution/product, traction, financials, risk/fragility, and conviction — precisely because that's close to how an experienced partner actually reads a deck, whether or not they'd list it that way out loud.

An overall score is a weighted combination of those seven readings. Change the weighting, and the number changes without a single word of the deck changing. That isn't measurement error. That is the entire mechanism by which "fit" exists between a company and a fund.

What does an investor's "thesis" actually mean, mechanically?

Founders hear "thesis" and picture a paragraph on a website — sector, stage, cheque size. That's the thesis a fund markets to its limited partners. It is not, reliably, the thesis the fund applies when your deck is actually open on the screen.

We tested the gap directly. Across 7,800+ real investor thesis statements and 162 VC deal memos from 15 ANZ funds explaining exactly why they invested, what a fund says it prioritises and what its decisions reveal it prioritises diverge sharply. Product execution predicts investment outcomes about as strongly as any single signal in that dataset, yet it shows up in roughly 4% of written theses — the least-stated of the major dimensions. Conviction, the partner's gut read that this is a winner, comes out as the strongest single predictor in the deal memos, ahead of team, ahead of market. Stated thesis and revealed thesis are two different re-weightings of the same seven dimensions. (Full breakdown in The Say-Do Gap.)

That's what a thesis actually is, underneath the marketing language: a weighting function, applied consistently to whatever deck lands in front of it. Not a taste. Not a mood on the day. We've mapped this into nine distinct evaluation lenses — from the team-heavy read of an angel to the financials-heavy read of a private equity fund — in a separate piece on investor archetypes. This post isn't about what those nine lenses are. It's about why the re-weighting underneath them is the mathematically correct thing for each of those investors to be doing.

It's worth separating this from a different, harsher filter that happens earlier. A fund's anti-thesis — the sectors and models it flatly won't touch — is binary. Hardware, for a software-only fund, doesn't get re-weighted low; it gets excluded before any dimension is scored at all. Re-weighting is what happens to a deck once it's already inside a fund's stated scope. It decides how well a deal scores, not whether it gets read.

How does the same company score differently to an angel, a growth fund, and a family office?

Hold one company's underlying facts fixed at each stage of its life, and three genuinely different, genuinely correct investors will still see three different faces of it.

• At pre-seed, an angel is evaluating a promise. There's no traction to underwrite yet, no financial model worth the paper it's printed on, often not even a finished product. What exists is a founder and a market thesis. The rational move for a cheque that size, at that stage, is to weight the evaluation hardest toward team — founder-market fit, resourcefulness, the pattern of someone who has already solved an adjacent problem. That isn't naivety. It's the only evidence that actually exists yet.
• At growth stage, the underwriting logic inverts. The team has already been proven — surviving pre-seed, seed, and a Series A is itself the proof. What a growth investor is underwriting now is whether the machine scales: revenue growth rate, net revenue retention, gross margin trajectory, CAC payback. A growth fund that kept weighting its evaluation toward "founder story" at this stage would be pricing a risk it can no longer see, and ignoring the one risk it actually faces — whether this specific engine keeps compounding at this specific size.
• A family office weights differently again — toward risk and durability, almost regardless of stage. Family office capital is close to irreplaceable. It rarely comes from limited partners who've priced in an 80% portfolio failure rate; it's often a family's own principal. Asking "what's the downside, and can we survive it" before anything else isn't conservatism dressed up as strategy. It's the mathematically rational response to capital you cannot simply raise again next vintage.

None of these three is evaluating a different company. All three are looking at the same seven dimensions of the same deck. What differs is which dimension the specific structure of their capital makes it rational to weight hardest — and that structure is fixed long before your deck ever reaches their inbox. You cannot pitch your way out of it. You can only learn which lens you're walking into.

Why does a one-size score hurt both founders and investors?

For a founder, an un-weighted, generic score does something specific and unhelpful: it averages away exactly the information you need. A company that's a strong fit for the right growth fund and a genuine miss for an angel can collapse to a flat, unremarkable middle score the moment you flatten every lens into one number. That average describes no real investor's decision — it describes a fund that doesn't exist. Reading a generic score as "how fundable am I, full stop" is answering a question nobody with a chequebook is actually asking you.

For an investor, the cost runs the other direction, and it's arguably the more expensive mistake. A screening tool that scores every deal the same way — team-weighted, or market-weighted, or whatever the default happens to be — isn't a neutral instrument. It's an implicit thesis, usually not the fund's own, quietly substituted in underneath the ranking. A generalist screen doesn't sharpen a specialist fund's judgement; it dilutes it back toward the average deal, because it's optimised for no one's actual mandate. The value of a thesis-aware screen — one that re-weights the same dimensions the way a fund's own capital structure requires — is that it expresses that fund's conviction back at scale, instead of smoothing it into consensus.

This is, not incidentally, why "spray and pray" fundraising and "score everything the same way" screening fail for the same underlying reason. Both treat a multi-dimensional evaluation as if it collapses to one number that means the same thing to everyone reading it. It doesn't, on either side of the table.

What should you actually do with a score, on either side of the table?

If you're raising: don't read a single score as a verdict on your company. Read it as a verdict from one lens, and ask which lens produced it. A weak score from a growth-weighted read of a pre-revenue pre-seed company isn't damning — it's the wrong lens applied too early. The same deck, read through a team-and-market lens, may be genuinely strong. Your job isn't to inflate the number; it's to find the fund whose weighting function matches what your company actually is right now.

If you're screening: don't treat a generic ranking as objective. Ask what it's implicitly weighting, and whether that matches your fund's real capital structure, cheque size, and risk tolerance — not the categories on your website, the thesis language written for LPs. NUVC's Deal Lens exists for exactly this: it takes the same seven-dimension score every deck already has and re-weights it against a fund's specific mandate, in milliseconds, with no extra model call — so the number in a pipeline reflects that fund's actual thesis, not an average of everyone's.

The deck genuinely doesn't change between the angel's inbox and the growth fund's. What changes is legitimate, mechanical, and knowable in advance. Score for that — on both sides.

See your score re-weighted for the investor types you're actually pitching → nuvc.ai

Frequently asked questions

Why do different investors give different scores on the same pitch deck?

Because a score is a weighted combination of several evaluation dimensions — team, market, product, traction, financials, risk, and conviction — and different investor types legitimately weight those dimensions differently based on their stage, cheque size, and capital structure. An angel weights team hardest because there's no traction yet to evaluate; a growth fund weights traction and unit economics hardest because the team is already proven; a family office weights risk hardest because it deploys capital it can't easily replace. The deck doesn't change. The weighting does.

Is a lower score from one investor a sign my startup isn't fundable?

Not necessarily. A lower score from an investor whose thesis doesn't match your current stage — a growth-weighted read of a pre-revenue pre-seed company, for example — reflects a mismatch of lens, not a verdict on the company. The same deck can legitimately score well through a different, appropriately-staged lens.

What is an investor's "thesis," mechanically?

An investor's stated thesis — sector, stage, cheque size — is usually LP-facing marketing language. Their revealed thesis, the one that actually predicts their decisions, is a specific weighting of the same evaluation dimensions every other investor uses. Analysis of 7,800+ investor theses and 162 real VC deal memos shows the two can diverge sharply — product execution, for instance, predicts outcomes strongly while appearing in only a small share of written theses.

Should a VC fund use a generic startup score to screen deals?

Not on its own. A generic, un-weighted score describes an average investor that doesn't exist, which makes it a weaker predictor of any single fund's actual decision than a score re-weighted to that fund's specific mandate. A thesis-aware screen expresses a fund's conviction; a generic one dilutes it toward the average deal.

Sources

1. Gompers, P. A., Gornall, W., Kaplan, S. N., & Strebulaev, I. A. (2020). How do venture capitalists make decisions? Journal of Financial Economics, 135(1), 169-190.
2. NUVC Research, "The Say-Do Gap" — analysis of 7,800+ investor thesis statements and 162 real VC deal memos from 15 ANZ funds. nuvc.ai/blog/say-do-gap-what-vcs-actually-evaluate
3. NUVC Research, "The 9 Investor Archetypes" — evaluation-lens mapping derived from 4,988 investor thesis statements. nuvc.ai/blog/9-investor-archetypes-how-vcs-actually-weight-your-pitch
4. NUVC Research, "The Anti-Thesis" — explicit exclusion analysis across 7,800+ investor thesis statements. nuvc.ai/blog/anti-thesis-what-vcs-explicitly-wont-fund
5. NUVC scoring methodology: calibrated on 1,400+ real VC investment decisions, validated against 200+ verified funding outcomes.