A Guide to Conducting a Cohort Analysis as a VC

very portfolio tells a story. The problem is, most VCs aren’t listening closely enough. We track IRRs, celebrate markups, and skim quarterly updates, but rarely do we zoom in to ask how and why our investments perform over time. 

That’s where cohort analysis comes in. 

It’s not just for SaaS founders or growth teams. Done right, it’s one of the most revealing tools in venture capital, a way to surface patterns, pressure-test assumptions, and make better investment decisions across the lifecycle of a fund.

Why Cohort Analysis Matters in VC

Cohort analysis is the practice of grouping things - users, customers, investments - by a shared starting point, then tracking how they perform over time. While it’s widely used in product and marketing teams to study customer retention, it’s just as valuable for venture capital firms.

In VC, cohort analysis lets you move beyond surface-level performance. Instead of asking “How’s the portfolio doing?” you start asking “How’s the 2018 vintage tracking compared to 2016?” or “Are our seed investments in climate tech maturing faster than our fintech bets?” It gives you a way to measure not just outcomes, but momentum, and whether your thesis is actually compounding value over time.

Done right, cohort analysis reveals the quality of what you're building, not just the quantity. It helps you see if you’re improving with every fund cycle, catching red flags early, and backing the right companies at the right time. In a game where power laws drive returns and timing shapes outcomes, cohort thinking can turn hindsight into foresight.

Defining a VC Cohort: What Are You Grouping and Why?

Cohort analysis only works if you’re grouping the right things, and in venture capital, that depends on what you're trying to learn.

At its core, a “cohort” is just a set of investments that share a common characteristic. Group them right, and you can start to see real patterns - where returns are compounding, where they're stalling, and what’s driving the difference. Here are a few of the most effective ways VCs structure cohorts:

By Vintage Year

The most common cohort type in venture. Vintage year cohorts group investments made in the same calendar year, or within the same fund vintage. This is especially useful for tracking how timing impacts returns. Are companies backed in 2018 outperforming those in 2020? Do downturn-era vintages recover slower but return more? Vintage curves can answer these questions and guide fund pacing, allocation, and risk exposure over time.

By Investment Stage

Group deals by the stage at which you invested - Pre-seed, Seed, Series A, and beyond. This helps you understand where your firm’s strengths lie. Are your seed-stage bets consistently reaching Series B? Is your Series A strategy actually producing breakout winners? Stage-based cohorts also highlight capital efficiency and survival rates across inflection points.

By Sector or Investment Thesis

If your firm runs multiple theses or targets diverse verticals, it’s useful to cohort by theme. Think - fintech vs. climate tech, or AI-first vs. marketplace models. Sector-based cohorts can show whether certain bets consistently underperform or if particular themes are reaching liquidity faster. This is especially valuable when refining your thesis or presenting sector-specific results to LPs.

The key is clarity. Whatever lens you use - time, stage, or theme - make sure the cohort tells a story worth tracking. Done well, it gives you a framework to compare, learn, and improve with every cycle.

What Are the Key Metrics to Track During a Cohort Analysis? 

Cohort analysis is only as powerful as the metrics you choose to track. When evaluating a portfolio, or comparing cohorts across time, stage, or sector, you need a clear blend of quantitative indicators and qualitative signals. Together, they offer a full picture of how your investments are maturing, compounding, or falling short.

Here are the core metrics every VC should consider when running cohort analysis:

Capital Deployed

Start simple - how much money was put to work across the cohort? This baseline lets you normalize outcomes and compare ROI across different groups. If two cohorts returned the same dollar amount but one required half the capital, the story is very different.

Follow-on Rates

A critical signal of downstream investor validation. What percentage of companies in each cohort raised a subsequent round within a given time window? Rising follow-on rates often point to quality selection and market pull, while drop-offs can suggest timing misfires or poor initial signals.

Step-ups / Markups (TVPI Over Time)

Track how the Total Value to Paid-In (TVPI) ratio evolves across periods for each cohort. A sharp early markup might look great, but flatlining afterward tells another story. Watching this metric over time reveals whether value is truly compounding or just spiking on paper.

IRR Per Cohort

IRR helps measure capital efficiency and time-based return for each cohort. While it can be noisy, especially early, it’s useful when comparing cohorts at consistent time intervals (e.g., IRR at 3 years post-investment). It’s also helpful when benchmarking fund vintages.

Exit Multiples

When exits do happen, track the gross multiple of invested capital (MOIC) per deal and average it across the cohort. This tells you how high the ceiling was for that vintage or thesis, and whether you’re capturing big outcomes or settling for singles.

Time to Exit

Measure the average time it took for companies in the cohort to reach liquidity events (acquisition, IPO, or shutdown). Faster isn’t always better, but patterns here can reveal market dynamics or strategy flaws. Are certain sectors consistently slower to mature? Are your seed deals taking longer to convert?

Attrition / Shutdown Rate

A vital but often underreported metric. How many companies in each cohort have failed outright or become dormant? Survival rates over time help distinguish real portfolio depth from inflated vanity metrics. A high TVPI with high attrition may still be a fragile cohort.

The goal is to step back and see what the numbers say together and not to obsess over each number in isolation. Cohort metrics turn a set of deals into a story of strategy, timing, and learning. And in venture, that story is often where the real edge lies.

How to Structure a Cohort Analysis as a VC

Cohort analysis isn’t just about slicing data, it’s about setting up a repeatable framework that reveals how your investments perform over time. Whether you’re analyzing fund vintages, sector plays, or stage-specific bets, here’s how to structure it step by step.

Step 1: Choose Your Cohort Criteria

Start by defining the group you want to analyze. This depends on what you’re trying to learn.

• Example: All seed-stage investments made in 2020
• Other options: All Series A investments in fintech, all climate tech deals across 2018–2021, or portfolio companies from Fund II

Step 2: Gather the Right Data

The quality of your analysis hinges on clean, complete inputs. Pull data from:

• Internal CRM systems and deal tracking tools
• Investment memos and quarterly portfolio updates
• Cap tables, valuation reports, and fundraising records. 

Make sure your data reflects the full lifecycle: entry valuations, follow-on rounds, write-downs, and exits.

Step 3: Normalize Data

This is where most analyses go sideways. Standardize definitions across all deals before comparing them.

• Adjust for currency differences across geographies
• Normalize round labels (e.g., Seed vs. Pre-Seed, Series A vs. Bridge)
• Align timeframes - measure “Year 1” consistently from the date of investment, not the calendar year

This ensures you’re comparing true like-for-like performance across the cohort.

Step 4: Visualize the Cohorts

Numbers only get you so far; visualization is what makes patterns click.

• Use line charts to show TVPI growth over time. These highlight J-curve dynamics and help compare value creation velocity across vintages.

• Build waterfall charts to break down capital returned vs. remaining. They make it easy to see how each cohort contributes to overall fund performance.
  

Done well, this structure gives you a dynamic view of where value is building, where it’s leaking, and what to refine in the next cycle. Cohort analysis isn’t a one-off, it’s a discipline.

Tools to Use to Conduct a Cohort Analysis

You don’t need a massive tech stack to start running meaningful cohort analyses, but the right tools can save time, reduce errors, and reveal insights faster. 

Here's a breakdown of what you need and what’s available.

Tools for Synthesizing and Structuring Data

For early-stage firms or smaller teams, simple tools like Google Sheets or Excel can get the job done. These are ideal for manually structuring cohorts, building retention tables, tracking follow-on rounds, and visualizing capital deployment. Add conditional formatting, pivot tables, or basic charts to surface trends without friction.

VC-Specific Tools for Grouping and Analyzing Cohorts

As your portfolio grows, manual tracking breaks down. This is where specialized platforms shine, especially when you want to track cohorts by vintage, sector, or stage with integrated dashboards and dynamic visuals.

PitchBook & Crunchbase:
Powerful for financial and deal data. While not built for cohort analysis per se, they’re excellent for benchmarking your portfolio against market-wide company cohorts.

Carta:
A comprehensive platform for equity and portfolio management. Carta enables you to track valuations, cap tables, and performance metrics across companies—helpful when segmenting investments into cohorts by stage or timeline.

Dealroom & 4Degrees:
These platforms focus on pipeline and deal flow management. They help identify when deals were sourced, how they were categorized, and how they progressed—useful data for building cohort timelines from the top of the funnel onward.

Tactyc:
A newer portfolio modeling tool. Useful for scenario planning and forward-looking analysis once your cohorts are built.

Vestberry:
Tailored for VCs, it centralizes performance tracking, LP reporting, and cohort comparison with automated data collection. Ideal if you want to scale insights across funds.

TotemVC:
A full fund operating system. Includes tools to build and analyze investment cohorts over time, track IRR by vintage, and assess strategy-level trends.

Each of these platforms brings a different strength to the table—what matters is building a workflow where cohort data flows cleanly, updates consistently, and helps you make sharper decisions.

Watch Out: Cohort Analysis Mistakes To Avoid

Cohort analysis can be a powerful decision-making tool, if it's done right. But when the inputs are flawed or the interpretation lacks context, you risk drawing the wrong conclusions. Here are the most common pitfalls VCs should watch out for:

Relying Solely on Financial Metrics

Dollars tell part of the story, but not the whole thing. If you're only tracking IRR, TVPI, or exit multiples, you're missing qualitative indicators like team resilience, founder growth, or product evolution. These signals are often the early warnings, or green lights, long before the next valuation comes in.

Ignoring Dead Companies

Survivorship bias is real. A cohort that looks healthy because it excludes the write-offs is giving you a false sense of progress. Every analysis should include shutdowns and marked-down positions. The strength of a portfolio isn't just who wins, it's how many don’t.

Drawing Big Conclusions from Small Samples

If your 2019 fintech cohort has three companies, and one hits big, that doesn’t mean you’ve nailed the thesis. Small sample sizes can skew outcomes and inflate confidence. Patterns need repetition to be meaningful. Be cautious when the “cohort” is really just a couple of bets.

Not Updating Data Regularly

Outdated cohort data is as good as guesswork. If valuations, follow-on rounds, or shutdowns haven’t been refreshed in a few quarters, you’re basing your decisions on stale numbers. Build a cadence - quarterly updates at minimum - to keep your analysis actionable.

Overcomplicating the Cohorts

Trying to cohort by five dimensions at once - sector, geography, founder type, stage, and time - can lead to analysis paralysis. Start simple. Choose one meaningful lens and layer complexity only when it adds clarity.

Final Thoughts: Building a Cohort-Driven VC Culture

Cohort analysis isn’t a reporting exercise, it’s a mindset. When used well, it becomes a core part of how a venture firm learns, adapts, and compounds insight across funds. The goal isn’t just to look backward, it’s to make sharper moves forward.

Bring cohort analysis into quarterly investment committee meetings. Not just as a dashboard, but as a discussion. How is our 2020 vintage tracking? Are our seed-stage bets in B2B SaaS following the same retention pattern as our 2021 cohort? Make it a habit to review performance not only deal by deal, but pattern by pattern.

Cohort trends should inform strategy. If certain themes consistently underperform or a stage is yielding lower-than-expected markups, that’s a signal. Adjust pacing. Recalibrate ownership targets. And when a cohort reveals breakout performance, double down on what made it work. These insights also help sharpen LP updates; you're not just sharing results, you’re showing depth of thinking.

Everyone on the team - partners, associates, platform leads - should feel empowered to spot and surface trends. Maybe it’s noticing that second-time founders in your 2021 cohort are hitting milestones faster. Or that your pre-seed investments in climate have a higher survival rate than fintech. The real value of cohort thinking comes when it becomes part of how your team sees the portfolio, not just how they report on it.

Cohort analysis gives structure to intuition, and direction to experience. Build it into your firm’s DNA, and it’ll sharpen everything from deal selection to fund strategy.

‍