Why I wrote this
Venture capital runs on judgment, relationships and a small number of huge wins. Software for it has to support fast decisions made with incomplete information, and most design advice was not written for that.
This guide is for designers building products for venture firms, their investors or the founders they back. It also helps if you are joining a firm in-house, where design often covers brand, the website and tools for the portfolio.
Venture looks like wealth management from the outside, but it behaves differently. Wealth management protects money that already exists. Venture bets on companies that might be worth a lot in ten years, and most of those bets fail. That changes what the numbers mean, who the users are and what a good screen looks like.
Read the first five sections in order. After that, jump to whatever you are working on.
How venture capital works
A venture firm raises a fund from outside investors, invests it in young private companies, and tries to return several times the money within about ten years. Four facts shape almost every design decision.
- The firm invests other people’s money. The people who run the fund are general partners (GPs). The people who put money in are limited partners (LPs): pension funds, university endowments, foundations, family offices, funds of funds and wealthy individuals. LPs commit money up front but only send it when the fund asks for it.
- Fees are set by the fund, not by clicks. A common structure is a management fee of about 2 percent a year to run the firm, plus carried interest of about 20 percent of the profits. Carry is where partners make real money, and it only pays out if the fund returns more than it took in.
- Funds have a long, fixed life. A fund usually invests in new companies for its first three to five years, then spends the rest supporting them and waiting for exits. Ten years is typical and extensions are common. A firm raises a new fund every few years, so it is usually managing several funds at once.
- A few wins pay for everything. Most startups fail or return little. One or two outcomes can return the whole fund. This is the power law, and it explains why investors chase outliers, move fast on hot deals and accept many losses.
How money is made and returned
Money comes back only when a company is sold, goes public, or the fund sells its shares to another buyer. Until then, the value of the portfolio is an estimate. Firms mark their holdings to a fair value every quarter, usually based on the latest funding round or comparable companies.
This creates the J-curve. In the early years a fund looks like it is losing money, because fees are paid and failures show up before the wins. Returns climb later, as winners raise at higher prices and eventually exit.
Venture capital, visualized
Why returns
bend before they rise.
In the early years, a fund can look like it is losing money. Costs and failures arrive first. The wins take time.
Illustrative cumulative net return
The venture capital J-curve
Fees arrive first
Management fees and operating costs begin before investments have had time to mature.
Failures show early
Losses become visible quickly. Breakout companies often take years to reveal their value.
Winners drive the curve
Later rounds, acquisitions, and public exits can turn a few outliers into fund-level returns.
Illustrative only. Actual fund performance varies materially.
Net return · 10-year horizon · Not to scale
Where firms differ
| Type | What it does | What it means for your design |
|---|---|---|
| Pre-seed and seed fund | Small checks into very early companies, many bets | Huge pipeline, little data per company, speed matters most |
| Early-stage fund (Series A and B) | Leads rounds, takes a board seat, deep involvement | Diligence, board work and founder relationships are central |
| Growth fund | Large checks into companies with real revenue | More financial analysis, closer to public market tools |
| Multi-stage firm | Invests from seed through growth across several funds | Many teams, many funds, permissions and reporting get complex |
| Corporate venture (CVC) | A company investing for strategy and returns | Internal approvals and links to the parent business |
| Solo GP and angel syndicate | One person or a small group, often on platforms | Very light tooling, mobile, fast |
| Fund of funds | Invests in venture funds, not companies | Lives in LP reporting and comparing managers |
At a glance
Money only comes back to investors when a company exits
One venture fund, simplified. A typical fund runs about ten years.
- Limited partnersPensions, endowments, family offices and foundations
- ↓ Capital calls↑ Distributions
- The fundRun by general partners
Fee: about 2% a year
Carry: about 20% of profits - ↓ Investments
- Portfolio companiesMost fail or return little. A few return the whole fund.
- ↓ Years later
- ExitSale, IPO or secondary sale
- ↑ Proceeds from the sale, back to the fund
Cash reaches LPs only after an exit. That is why young funds report mostly paper value, and older funds are judged on the cash they have returned.
The people you design for
A venture firm is small, often fewer than twenty people, but the roles are very different. Most products also reach outside the firm to LPs and founders.
| Person | What they want | What they fear | Design for |
|---|---|---|---|
| General partner | To find and win the few deals that matter | Missing the breakout company, or a competitor getting there first | Speed, mobile access, relationship history at a glance |
| Principal or associate | To source deals and do diligence fast | Bringing a weak deal to partners, or missing a red flag | Pipeline tools, research, memo drafting, notes that are easy to find |
| Platform or portfolio team | To help founders with hiring, intros and events | Being seen as overhead, not value | Network search, intro tracking, founder requests |
| CFO and fund operations | Accurate books, capital calls and reports on time | A wrong number in an LP report or a missed call deadline | Audit trails, approvals, reconciliation, clean exports |
| Investor relations | To keep LPs informed and raise the next fund | An LP losing confidence | Reporting, data rooms, LP CRM |
| Limited partner | To know how the fund is doing and when cash will be called or returned | Surprises, stale numbers, managers hiding bad news | Clear quarterly reporting, cash flow forecasts, comparable metrics |
| Founder | Money, help, and as little reporting work as possible | Losing control, or a board surprise | Short update forms, a clear view of their own cap table, quick answers |
The words you need to know
Venture has its own vocabulary, and investors use it fast. Learn these well enough to follow a partner meeting and label a screen correctly.
| Term | Plain meaning | Why a designer cares |
|---|---|---|
| GP and LP | The people who run the fund, and the investors who put money in | Two very different audiences, often in the same product |
| Fund size and commitment | The total LPs have promised to invest | The starting point for every fund-level number |
| Called capital and paid-in capital | The share of commitments the fund has asked for and received so far | LPs track how much is still to come |
| Dry powder | Committed money not yet invested | Tells partners how much room is left for new deals and follow-ons |
| Deal flow | The stream of companies a firm sees | The heart of most venture CRMs |
| Thesis | The kinds of companies a firm believes in | Drives filters, tags and sourcing |
| Term sheet | A short offer setting out the main terms of an investment | A milestone in the pipeline, often time-sensitive |
| Pre-money and post-money valuation | Company value before and after the new money goes in | Mixing them up gives wrong ownership numbers |
| SAFE and convertible note | Early money that turns into shares at a later round | Hard to show on a cap table until it converts |
| Priced round | A round that sets a share price, such as Series A | Resets valuations across the portfolio |
| Cap table | Who owns what share of a company | Needs careful design for many rounds and share classes |
| Dilution | An owner’s share shrinking as new shares are issued | Founders and investors watch it closely |
| Pro rata right | The right to invest in later rounds to keep the same ownership | Creates follow-on decisions and deadlines |
| Liquidation preference | Who gets paid first in a sale, and how much | Means ownership percentage alone does not show what someone gets |
| Mark or fair value | The fund’s estimate of what a holding is worth today | An estimate, not a price; always date it |
| Write-up and write-down | Raising or lowering a mark | Changes reported returns without any cash moving |
| MOIC | Multiple on invested capital: value divided by money put in | The quickest read on a single investment |
| TVPI | Total value to paid-in: everything returned plus current value, divided by money paid in | Includes paper value, so it can fall |
| DPI | Distributions to paid-in: cash actually returned divided by money paid in | The number LPs trust most, because it is real cash |
| IRR | Annual return that accounts for when money went in and came out | Very sensitive to timing, easy to flatter |
| Vintage | The year a fund started investing | Funds are only fairly compared within the same vintage |
| J-curve | Early paper losses followed by later gains | Explains why young funds look bad |
| Capital call | The fund asking LPs to send part of their commitment | A deadline with penalties for missing it |
| Distribution | Cash or shares returned to LPs after an exit | The moment LPs care about most |
| Exit | A sale, IPO or share sale that returns money | The end of an investment |
| Secondary | Buying or selling existing shares or fund stakes | A growing route to cash before an exit |
| Investment committee (IC) | The partners who approve investments | The end point of diligence and memo flows |
| Data room | A secure folder of documents shared during diligence or fundraising | Needs permissions, tracking and watermarks |
Principles
Venture tools serve people making high-stakes bets with too little information. These are the principles I would hold onto.
- Speed is a feature. A hot round can close in days. If your product makes an investor wait, click through tabs or retype data, they will go back to email and spreadsheets.
- Relationships are the real data. Who knows the founder, who introduced the deal, when someone last spoke to them. Make that history easy to see and easy to capture without typing.
- Capture, do not ask. Pull information from email, calendars, decks and notes. Every field you ask a partner to fill in is a field that stays empty.
- Support judgment, do not replace it. Investors are paid for their taste. Show evidence and patterns. Leave the call to them.
- Design for the outlier. Averages hide what matters in a power law business. Make it easy to spot the one company that is breaking out and the one that is in trouble.
- Be honest about estimates. Marks are opinions. Label them, date them and show the method, so nobody mistakes a paper gain for cash.
- Confidentiality is constant. Deal names, valuations and founder information are secret. A leaked term sheet can kill a deal. Assume every screen will be seen by someone who should not see it.
- Founder-friendly is a brand position. Firms compete to win founders. Anything a founder touches, from a portal to an update form, is part of the firm’s reputation.
- Taste shows. Investors judge companies partly on how they present themselves, and they judge tools the same way. Craft and restraint read as competence.
Designing with venture numbers
Venture data is sparse, late and full of estimates. A public stock has a price every second. A startup has a price maybe once every eighteen months. Good design makes that difference visible instead of hiding it.
Rules for venture metrics
- Always pair TVPI with DPI. TVPI includes paper value, DPI is cash. Showing only TVPI flatters young funds.
- Label every mark with its date and basis, such as “last round, March 2026” or “comparable companies, Q3 2026”.
- Show IRR with its date and whether it is gross or net of fees and carry. Net is what LPs actually get.
- Compare funds only within the same vintage. A three-year-old fund and a ten-year-old fund cannot be ranked side by side.
- Show ownership as fully diluted unless you say otherwise, and note when convertibles have not yet converted.
- Keep cost, current value and realized proceeds in separate columns. Blending them hides what is real.
- Mark missing data clearly. A company that did not send its update is different from a company with zero revenue.
- Expect mixed currencies and fiscal years in international portfolios, and label both.
Cap tables
Cap tables are one of the hardest things to design in this field. They have many rounds, share classes, option pools and instruments that convert later. Show a simple summary first, such as founders, investors, employees and pool. Let people expand into rounds and individual holders. Offer a “what if” view for a new round so people can see dilution before they sign.
Remember that percentage owned is not the same as money received. Liquidation preferences decide who gets paid first. A waterfall view that shows payouts at different exit values is often more useful than the percentages.
Portfolio views
| Question | View that fits | Watch out for |
|---|---|---|
| Which companies matter most? | Holdings sorted by current value or share of fund | A long list where the winners get lost |
| How is the fund doing over time? | TVPI and DPI by quarter since the start of the fund | Hiding the J-curve or starting the chart late |
| Where is money concentrated? | Value by sector, stage or company | Many tiny slices nobody can read |
| Who needs help or more money? | Runway in months and last raise date per company | Treating stale runway figures as current |
| What drove the change in value? | Bridge from last quarter’s value to this quarter’s: new investments, write-ups, write-downs, exits | Lumping everything into one “change” number |
Every chart needs a way to see the numbers behind it. Investors and LPs will check.
The core flows
Six flows show up in almost every venture product. They follow the life of a deal from first meeting to cash back.
1. Sourcing and deal flow
Firms see far more companies than they back, often a hundred or more for each investment. The pipeline is a CRM with stages such as sourced, first meeting, partner meeting, diligence, investment committee, term sheet and closed or passed.
- Make adding a company almost effortless, from an email, a deck or a name.
- Show who at the firm already knows the founder or the company.
- Flag duplicates. Two partners chasing the same company without knowing is embarrassing.
- Record why a deal was passed. Firms revisit passes, and the reason matters later.
- Let people tag by thesis, stage, sector and source.
2. Diligence and the investment memo
Diligence pulls together the deck, financials, market research, customer calls and references. The output is an investment memo that partners read before the committee votes.
Design for collaboration across a small team under time pressure: shared notes, a checklist of open questions, document links and a clear status. Keep the memo template flexible. Every firm writes them differently.
3. Closing
Closing involves the term sheet, legal documents, signatures and wiring money. Show what is signed, what is outstanding and who is waiting on whom. Treat wire details with the same care as any money movement: restate amount and recipient, and confirm through a second channel.
4. Portfolio monitoring
After the investment, the firm tracks each company through founder updates, board meetings and financial reports. Data arrives in every format, from email updates to spreadsheets to board decks.
- Make founder update forms short and reusable from one quarter to the next.
- Show what changed since the last update, not just the latest numbers.
- Highlight risk signals such as runway under twelve months or a missed plan.
- Track follow-on decisions and pro rata deadlines.
5. LP reporting and fund operations
Every quarter the fund sends LPs a report with performance, holdings and commentary. Fund operations also send capital call notices, distribution notices and annual tax documents.
Design the LP portal around three questions: how is the fund doing, what do I owe, and what am I getting back. Show calls and distributions on one timeline. Let LPs download everything as files, because many will load it into their own systems.
6. Fundraising
Raising the next fund means pitching LPs, sharing a data room and tracking commitments. The data room needs permissions per investor, watermarks, view tracking and an easy way to revoke access. The LP pipeline looks like the deal pipeline in reverse.
Confidentiality and compliance
In venture, secrecy is a bigger daily concern than regulation. A leaked valuation or deal name can cost the firm a deal or a founder’s trust. Rules still apply, and they shape fund operations and LP reporting most.
This section describes the United States in general terms. Rules differ by country and change often, so confirm specifics with the firm’s legal or compliance team before relying on them.
| Area | What it is about | What it changes on screen |
|---|---|---|
| Deal confidentiality | Term sheets, valuations and founder data are secret, often under NDA | Access by deal team, no previews in notifications, watermarks, careful sharing |
| Information barriers | Firms may hold competing companies or know private details about public ones | Walls between deal teams, logs of who viewed what |
| Adviser registration and reporting | Many venture advisers have lighter rules than other fund managers, but still have duties | Records kept, disclosures to LPs, approvals on what gets sent out |
| Marketing and performance claims | Claims about returns made to prospective LPs must be fair | Gross and net returns shown together, periods labeled, disclosures beside the claim |
| Investor eligibility | Private funds are generally limited to investors above wealth or sophistication thresholds | Eligibility checks before an LP sees fund materials |
| Identity and money laundering checks | Funds must know who their LPs are | Onboarding forms, document upload, checks on new investors |
| Fund documents | The partnership agreement sets reporting, fees and LP rights | Reports must match what the agreement promises |
| Privacy | Founder, employee and LP personal data must be protected | Masking, access control, data retention rules |
How to work with legal and finance
- Ask early which data each role may see. Permissions are the first design decision, not the last.
- Treat the audit trail as a feature. Who changed a mark and why should be one click away.
- Keep LP-facing numbers locked once a report is published. Corrections should be visible as corrections.
- Make sharing deliberate. Warn before sending anything outside the firm, and show who will see it.
Designing AI features for investors
AI is already reshaping venture work: finding companies, summarizing decks, drafting memos, reading founder updates and answering LP questions. The upside is real. So is the risk, because a confident wrong answer can send a partner into a meeting with the wrong facts.
- Cite everything. Every claim in a summary or memo should link to the deck page, email, call note or data source it came from.
- Separate facts from opinions. “Revenue grew 40 percent last year” comes from a document. “Strong team” is a judgment. Style them differently, and be careful about AI forming judgments at all.
- Watch for bias in sourcing. Models trained on past deals can repeat past patterns about who gets funded. Make scoring explainable, let people see why a company was ranked, and never auto-reject.
- Say when data is missing or old. Private company data is thin. “No revenue figure found” beats a guess.
- Keep people in charge of decisions and outbound messages. AI drafts the memo, the founder email or the LP answer. A person reviews and sends it.
- Protect confidential data. State plainly whether deal and portfolio data are used to train models, and keep data from one firm or deal away from another.
- Show change, not just state. The most useful AI in monitoring points out what moved: a missed target, a new hire, a sudden drop in runway.
- Hand off to a human well. When AI cannot answer, route the question to an analyst with the context attached.
A useful test: would a partner be comfortable repeating this output to a founder or an LP, using only what is on the screen? If not, the design is not finished.
Research with investors who have no time
Partners are among the hardest users to schedule. Plan around that.
- Start with associates and fund operations. They do most of the hands-on work, know where the pain is and are easier to book.
- Ask for fifteen minutes and come with something concrete. A prototype with realistic data gets real reactions in a short slot.
- Watch the real workflow. Sit with someone preparing for a partner meeting, a board meeting or a quarterly LP report.
- Look at their actual tools. Most firms run on spreadsheets, email and a CRM bolted together. The workarounds show what is missing.
- Talk to founders and LPs separately. Their needs conflict, and they will not be candid in front of each other.
- Use realistic, anonymized data. Test with a portfolio of forty companies, three funds, mixed rounds and missing updates.
- Respect confidentiality in your notes. Remove company names, valuations and people from anything you share.
- Be honest about small samples. Five firms give you direction, not proof. Firms vary a lot by stage and size.
Common mistakes and a checklist
Mistakes I would warn anyone about
| Mistake | Why it hurts | Do this instead |
|---|---|---|
| Treating marks like prices | Paper gains look like real money | Date every mark, show its basis, pair TVPI with DPI |
| Asking partners to enter data | They will not, and the CRM goes stale | Capture from email, calendar and documents |
| Designing like a public markets terminal | Startup data is sparse and slow; live tickers make no sense | Design for quarterly updates and missing data |
| Burying relationship history | The most valuable context is invisible | Show who knows whom and the last touchpoint |
| One view for GPs, LPs and founders | Each sees too much or too little | Separate views with strict permissions |
| Ignoring the founder experience | Late, incomplete updates and a worse firm reputation | Short, reusable forms that respect founder time |
| Comparing funds across vintages | Misleading rankings | Compare within vintage, show fund age |
| Showing IRR alone | Easy to flatter with timing | Show IRR with multiples and the date |
| Leaky notifications and previews | Deal names on a lock screen or in a shared channel | Generic notifications, details behind sign-in |
| Designing only the happy path | Real portfolios have down rounds, write-offs and silent founders | Design write-downs, missing updates and failures first |
Before you ship
- Every mark shows its date and basis
- TVPI always appears with DPI
- IRR and returns state gross or net and the as-of date
- Ownership says whether it is fully diluted
- Missing and late data each have a designed state
- Permissions tested for partner, associate, ops, LP and founder
- Nothing confidential shows in notifications, previews or exports by default
- Data entry for partners is optional, not required
- LP numbers match across the portal, the PDF report and exports
- AI output links to its sources and is labeled as AI
- Tested with realistic, anonymized portfolio data
