Venture Capital Research for Founders: A Practical Guide

Venture capital research explained for founders: how to score investor fit, read portfolios and avoid bias so your shortlist earns real replies.

Venture Capital Research for Founders: A Practical Guide
Do not index
Do not index
Global venture capital investment fell sharply after its 2021 peak. That shift reflected a change in the market cycle, not a verdict on the quality of every startup seeking funding.
For a founder with 11 months of runway, the practical lesson is to treat investor research as bias-aware fit testing. Build a ranked, evidence-backed shortlist, and give each name a reason you can explain. Check that reason against the available evidence before spending time on outreach.
The shortlist is the result. The research is the work of testing each potential match.

What Venture Capital Research Covers and Why Timing Matters

Venture capital research has three parts: discovery, fit testing and verification. Discovery identifies investors who may fund a company at its stage, in its sector and in its country. Fit testing compares each investor's focus with the startup's actual needs. Verification checks whether the evidence still holds, since a fund's published investment thesis can become outdated.
A list alone is a map without a route. Research turns it into a ranked shortlist, with a clear reason behind each target.
Timing matters because deal activity responds to interest rates, public-market valuations, technology cycles and investors' appetite for risk. A quieter market does not automatically mean a weaker company. It may mean investors are taking longer to commit, or applying different expectations to a round.
The three tasks address different questions: Who might invest? How closely does each investor fit? What evidence supports that judgment? Keeping them separate makes it easier to spot weak assumptions. A portfolio match, for example, is a useful clue, not proof that an investor is currently making similar investments.
If you need to check recent activity, Gritt's public feed of recent investments by investors in its database offers one way to see which companies those investors have backed. It is a record of investments, not a guide to how funding rounds are structured.
Choose where to focus your research based on what you already have: no names calls for discovery; a long, unranked list calls for fit testing; a ranked list with unconfirmed assumptions calls for verification.

Finding Investors Worth Researching

The long list comes first, and its quality depends entirely on where the names came from. Work four discovery sources, and record which source produced each name.
Portfolio pages. Go to investors who fund at your stage and read the companies they list by name. A portfolio page is the closest thing to a public statement of what a fund believes, because someone signed a cheque for each logo.
Public deal announcements from the last 12 months. These show who is writing cheques now, not who wrote them in 2021. A fund that hasn't announced a deal in 18 months may be deploying from an old vehicle, raising a new one, or winding down. You can't tell from the website.
Accelerator and co-investor lists. Syndicate partners reappear across rounds. If the same three funds keep showing up alongside your target investor, that cluster is a network you can map and reference.
Warm referrals from founders who raised recently. Ask what the partner was like in the meeting, not just who they were. A referral that produces a name and a personality profile is worth ten names from a database.

Apply hard filters before any scoring

Two filters do most of the work: stage and country. A pre-seed company shouldn't spend an hour researching a fund whose smallest cheque is Series A, and a Stockholm startup shouldn't prioritize a fund that only backs U.S. companies.
Structured attributes beat keyword matching. Investment stage, cheque size, sector expertise, relevant portfolio companies, follow-on capacity, geographic activity and governance preferences are all fields you can filter on. A keyword search for "AI" will happily return a growth fund that has never written a seed cheque.
A database-driven search on stage and country typically produces a long list of around 40 names. That's the right size: large enough to test, small enough to research properly. You can run that kind of structured screen with investor search tools built for early-stage founders.
Don't contact anyone yet. Record each name with its source and the date you found it. When you later cut an investor from the list, the trail tells you whether your filter was wrong or your thesis was.

The Investor Fit Scorecard Founders Should Build

A useful fit scorecard separates six decision dimensions: industry, investor capabilities, financial terms, working style, technical understanding and governance. Score them independently. A strong sector match cannot make up for a fund that invests at the wrong stage or expects terms your company cannot accept.
Dimension
Signals to check
Where to verify
Industry environment
Sector focus and relevant category investments
Investor website and portfolio page
Investor capabilities and network
Operating expertise, co-investors and ability to support later rounds
Partner biographies and investment announcements
Financial evaluation
Typical cheque size, ownership target and investment stage
Public deal announcements and founder reports
Organisational fit
Hands-on or hands-off style and board expectations
Partner interviews and founder references
Technological assessment
Understanding of your product domain and technical approach
Portfolio company backgrounds and partner biographies
Legal and governance
Preferred investment instruments and governance terms
Public term sheet examples and legal counsel
Financial fit deserves careful scrutiny. A headline valuation does not show how proceeds may be distributed in an exit. The NVCA model term sheet explains liquidation preference mechanics. With a 1x non-participating preference, an investor can take the original investment back first or convert to common stock. Participating preferred can return the preference before the investor shares in the remaining proceeds. Ask about the investor's typical terms, then review any proposed documents with counsel.

How to weight and score the rows

Score each dimension from 0 to 2: zero for no evidence, one for partial or ambiguous evidence, and two for clear, verified evidence. Add the scores for a total out of 12. Use 8 as a working threshold for deciding which investors merit further research, not as proof that a match will succeed.
Adjust the emphasis by stage. For a technically complex seed company, an investor's understanding of the product may deserve more weight. In later rounds, ownership expectations and governance terms may matter more. Keep the underlying scores visible so a change in emphasis does not conceal weak evidence.
Attach a source link or name a conversation for every score of 2. Record what the evidence supports and when you checked it. A portfolio page may confirm an investment, for example, but it cannot establish how involved the investor was or what terms they offered.
A broad label such as FinTech can hide important differences. Payments infrastructure, lending, insurtech and regulatory tooling are distinct markets. Score the fit against the specific problem your company solves, then compare the evidence across investors. The result is a ranked shortlist with reasons you can inspect, rather than a spreadsheet sorted by familiar sector labels.
Fill out the scorecard for five investors. Compare which dimensions separate the stronger candidates from the weaker ones, and revisit any score that rests on an assumption rather than evidence.

Reading Portfolios and Geography to Test Real Overlap

Portfolio overlap is the strongest observable proxy for thesis fit, but only when you read it in the right order. Most founders read it backwards: they find one comparable company, get excited, and skip the checks that would have told them the investment was in 2019 and never followed on.
notion image
Run three steps in sequence.
Find three named companies that solved a problem close to yours at a similar stage. Three, not one. A single adjacent logo could be an outlier, a favour to a friend, or a partner's personal bet. Three is a pattern.
Check whether the investor followed on in those companies. Follow-on investment is the clearest signal of conviction you'll find in public data, because it means the fund put more money in after seeing the company's actual numbers. A fund that leads seed rounds and never follows on is telling you something about how it behaves after the wire clears.
Compare recent deals with older ones. A portfolio is a historical record, not a current strategy. If every comparable company was funded before 2022, the thesis may have moved. Look for at least one deal in the last year that fits your category.
Geography works the same way: it's a scoring feature, not a gate. Headquarters, active regions and portfolio concentration by city shape ecosystem fit, which covers things a thesis match can't: customer access, hiring networks, acquisition relationships and syndicate partners.
Evidence from a study of 15,111 investments by 629 venture-capital firms across 5,905 startups and 362 metropolitan areas found that affiliation with prominent VC-financing clusters improved startup M&A success, and that startups in industry strongholds showed greater M&A likelihood, according to this MDPI analysis of venture capital geography. Geographic co-location with VC hubs didn't significantly affect IPO exits but did hinder M&A outcomes when absent.
That's a specific, useful finding. If your realistic exit is acquisition, proximity to a financing cluster carries more weight than it would for a company aiming at an IPO. Remote investing and cross-border syndication can overcome distance, so don't exclude a fund for its postcode. Down-weight it unless your first customers cluster in that market.
Here's the checklist for your top ten names. Three overlapping companies, one follow-on example, one deal from the last year, one geographic reason. If a name can't clear four lines, it's not a target yet, and an email that names a specific portfolio investment is the difference between a reply and a delete.

Testing Selection and Similarity Bias in the Evidence

notion image
A portfolio records the companies a fund backed, not the companies it considered and rejected. Fund multiples, portfolio rankings and deal volume therefore describe a selected group. Treat those results as evidence about a fund's choices, not as a full picture of how similar startups fare.
A study using administrative data on Norwegian new firms found that venture-backed companies had a sizeable valuation premium at birth, and that the premium doubled over the following five years. The authors interpret this as evidence that estimates of VC returns can be substantially biased upward when selection is not corrected, according to this Cambridge study of selection bias in early-stage investment returns. Some of the difference was already present before investors committed capital.
For your shortlist, this means a strong portfolio may reflect screening access or founder networks, as well as what the fund contributes after investing. Ask what the funded companies have in common, and consider which comparable companies are absent from the record.

Similarity bias narrows the door

A 2026 American Economic Association paper on disagreement among venture capital judges reports that VCs can disagree significantly when evaluating the same startup. The paper links some of that disagreement to VC-founder homophily bias: investors may favour founders who resemble them. Two judges can assess the same deck differently for reasons that do not reflect the company's underlying prospects.
Large funds and experienced managers continue to attract much of the available capital. That concentration does not mean every founder has equal access to it. A broad market can still produce a narrow set of opportunities for founders outside familiar networks or profiles.
Use three checks to test your assumptions.
Compare a fund's stated thesis with its actual investments. Look at whether its funded founders and companies reflect a range of backgrounds and business models. A pattern can suggest what the fund tends to select, though it cannot prove why.
Ask peers about response patterns and rejection reasons. “They never replied” and “they replied quickly and passed on market size” point to different issues. Separate access from the reasons a fund gives for declining.
Track your own replies and meetings by investor type. Your outreach record includes outcomes that portfolio databases omit. To compare investor directories before building your research list, review these Crunchbase alternatives for investor research.
Treat each portfolio pattern as a hypothesis, then compare it with your own evidence before spending time on outreach. A ranked shortlist is more useful when every name has a reason for its place, and the limits of that reason are clear.

A Worked Example From Long List to Ranked Targets

Take a pre-revenue AI company raising a seed round. The workflow below is the same for any sector, but AI makes the constraints vivid because capital is concentrated in a handful of categories.
notion image
Private-market data shows why the filtering matters. Carta reported that companies on its platform raised $49.3 billion across 1,179 venture rounds in Q3 2026, and that AI companies received about 57% of capital invested at Series B and later, while AI represented 44% of cash raised at seed and Series A, down from 48% in 2025, per Carta's State of Private Markets Q3 2026. Round counts fell slightly while total capital rose, which means larger average rounds and a more concentrated market.
For an AI founder, that produces a specific trap: plenty of funds will take a meeting, but a meaningful share of the capital only arrives at Series B. You need to know which kind of fund you're talking to before you spend the hour.

The list shrinking, step by step

Step 1: start with about 40 names filtered by stage and country. Before anything else, check that your company information is accurate. Match quality depends entirely on what you enter. If you describe yourself as Series A-ready when you're pre-revenue, or name the wrong sector because it sounds more fundable, the shortlist will be wrong in ways you can't see. Gritt matches investors using the information you provide, so inaccurate company information produces inaccurate matches.
Step 2: remove any investor whose stated stage starts at Series B or later. This single cut is brutal and liberating. For the example company, roughly half the list disappears.
Step 3: score the survivors on the six dimensions, 0 to 2 each. Most land between 5 and 8 out of 12. That's normal, and it's the sound of a list becoming a plan.
Step 4: check portfolio overlap for the top 15. Two or three named companies in the same problem space, at least one follow-on, at least one recent deal.
Step 5: apply geography as a score. If the first ten customers are in Germany, a Munich-based fund gets extra weight regardless of its postcode on a map.
Step 6: sort into three tiers. Lead candidates, angels and follow-on investors, and a watch list for funds that fit but don't lead at your stage.
The result is a shortlist of around 12 names, each with a one-line reason and a source link. Twelve is a working number: enough to run parallel conversations, few enough that you can write an individual email to each one.
A non-AI company runs the identical process and should expect less competition in its category, plus a lower risk of interest that evaporates at the next stage.

Organising Your Research Record and Outreach Pace

Research collapses without a record. One row per investor, one source of truth, no threads scattered across three inboxes and a notes app. Carta's investor relations tools describe a single source of truth for investor ownership information, contact details and cap-table data, with the ability to send updates, share pitch decks and see whether recipients opened or read messages. You don't need enterprise software to apply the principle. You need one table.
Field
Why it matters
Example
Stage
Filters out mismatched cheque sizes
Seed
Sector and geography
Confirms ecosystem fit
AI, Stockholm
Portfolio overlap
Basis for a specific opening line
Two named companies in your category
Last contact date
Prevents repeat outreach too soon
14 March
Response status
Separates opens from real interest
Replied, meeting booked
Next action
Keeps follow-ups on schedule
Send updated deck after the partner call
If you already keep notes in a tool you trust, the local-first notes workflow from Acres is a reasonable place to structure research that has to stay searchable when you're offline or between tools. Sector, country and stage filters in a searchable investor database can populate the first three fields quickly, which leaves you to fill in the judgement calls.
An open is not interest. Replies, meetings, diligence requests and commitments are the outcome measures. Tracking opens tells you whether your subject lines work, which is a diagnosis tool, not a scoreboard.

The sending rules most founders never read

Since February 2024, senders delivering more than 5,000 messages in a 24-hour period to personal Gmail accounts are classified as bulk senders, according to Google's Gmail sender guidelines. Those senders must authenticate with SPF, DKIM and DMARC, keep spam rates low, and offer one-click unsubscribe, with requests processed within two days.
Your campaign will stay far below 5,000 messages. The useful part isn't the threshold, it's what the rules reveal: sending infrastructure and recipient engagement directly affect where your email lands. A founder who blasts 400 untargeted emails is teaching Gmail to distrust them.
Three habits that protect the inbox.
  • Set a daily cap you can personalise. Five, twelve, thirty. Whatever number lets you write each email individually.
  • Remove bounced addresses after every batch. One invalid address in a thread tells providers your list is stale.
  • Review complaints and replies weekly. Complaints mean your targeting slipped. Silence across the board means your subject line or your list did.
Personal, investor-specific messages reduce complaints because they're less likely to be marked as unwanted. That's reputation management, and it compounds.

Confirm Your Shortlist and Start the First Batch

Approve a first outreach batch of 5 to 12 investors who clear the scorecard threshold and show specific portfolio overlap. That's the decision, and everything before this point was preparation for it.
Lock the threshold at 8 out of 12 and remove anyone below it, even the ones you like. Then confirm your company information is accurate, because every downstream match depends on those inputs. Draft one individual email per investor that names a specific portfolio company and explains the thesis link in a sentence.
Gritt drafts individual emails for matched investors using each investor's portfolio companies and the overlap with your startup, referencing those specific investments in the message. You approve each one, or set auto-approve, and only approved emails send from your own Gmail, so replies arrive in your inbox. Gritt doesn't guarantee that you'll raise money, receive a reply, get a meeting, or reach any particular investor, and neither does any other tool. The evidence and the emails are yours.
Review every draft before it leaves. Nothing should go out that you haven't read.
Set a review date two weeks out. On that date, check replies, bounces and scorecard scores, then decide whether to widen the list or rewrite the opening line. Most campaigns improve on the second batch, not the first.
Gritt.io turns this workflow into a running process: it matches your company to early-stage investors in a database of 45,000+ verified angels and VCs, drafts a separate email for each match that names the portfolio companies behind the fit, and sends only what you approve from your own Gmail. If you've built the scorecard and you're ready to test your shortlist against real conversations, visit Gritt.io to generate your first campaign and preview every email before you send one.

Ready to raise funds?

Join other 9.800+ startup founder now!

Subscribe

Written by

Gritt Team
Gritt Team

Fundraising guides from the team behind Gritt: 45,000+ investors, and personalised outreach sent from your own Gmail.