When should a startup build a growth team?

Only after you've proven strong retention. The most common mistake is investing in growth acquisition before fixing a 'leaky bucket.' Most companies hired their first growth PM when they had about 15 engineers — making the Growth PM typically the 3rd or 4th PM on the team. If your retention metrics are below median benchmarks for your vertical (e.g., 55% for social networks, 33% for subscription, 16% for e-commerce at Month 12), focus on retention first.

What does a Year 1 growth team look like?

The standard formula across Facebook, Airbnb, Uber, and other top startups: 1 Growth PM + 2-3 Growth Engineers + 1-2 Data Scientists. The Growth PM should be intensely data-driven, have prior growth experience at a competitive company (not Google/Apple), and ideally be a former founder (60% of growth experts were). The first hires are 'magnets' — they establish the experimentation culture and attract future talent.

What should a growth team's goal look like?

Always an absolute number, never a percentage change. Instead of 'improve conversion by 10%,' set a goal like 'achieve 5M first-time room nights this year.' Break the absolute goal into sub-goals using a growth accounting equation. Facebook's: [x] MAU = New MAU + Retained MAU + Resurrected MAU. Set the target halfway between 'sandbagging' and 'too hard' — and get the CEO fully aligned before communicating it company-wide.

What are the best growth channels for startups?

It depends on how users naturally behave with your product. If users need to connect with others to use it (Slack, PayPal), the product itself is the channel. If users talk about the product organically (Uber, Dropbox), invest in referrals. If users search for solutions (Airbnb), invest in SEO/SEM. About 70% of growth experts said referrals were the #1 channel in Year 1. Most products find only 1-2 channels that really work.

How many growth experiments should a startup run?

At scale, the benchmark is 1 experiment per growth engineer per week, meaning a team of 10 engineers runs ~500 experiments per year. Only 20-30% of experiments succeed — and that's expected. The point is to encourage risk-taking and learn quickly. You need an experiment dashboard to track results and a biweekly peer review process to discuss findings. Teams start with tools like Mixpanel and Optimizely, but 100% of experts eventually built their own internal tools.

How to Build a Startup Growth Team: The YC Framework Used by Facebook, Airbnb, and Uber (2025)

Growth Hacking Is Dead. Growth Teams Are Not.

"Growth hacking" sounds clever, but the experts who actually built the growth programs at Facebook, Airbnb, and Uber hate the term. "Hacking" implies a haphazard, gut-driven approach — and the reality is the opposite. The companies that achieved massive growth did it through structured experimentation, rigorous data analysis, and dedicated cross-functional teams.

Y Combinator's Continuity team interviewed 25 growth experts from Facebook, Airbnb, Uber, Stitch Fix, Square, Slack, Pinterest, and Instagram to identify the best practices for building a growth program. This guide distills those interviews into an actionable framework — from knowing when you're ready to invest in growth, to hiring your first growth PM, to what your growth team should accomplish in Year 1.

The Growth Framework

Growth Experts Interviewed

Year 1 Initiatives

70%

Used Referrals First

100+

Experiments/Year at Scale

Step 1: Don't Invest in Growth Until You've Fixed Retention

Here's the most counterintuitive advice from the experts: the fastest way to waste money is to invest in growth before you've proven you can retain customers. If your product is a leaky bucket, pouring more users in just means more users leaking out.

Before building a growth team, run through this retention checklist — the same framework used by Airbnb and Uber:

📋 The Retention Checklist

Pick the right retention metric

Choose a leading indicator of revenue and repeat behavior — not a vanity metric like app downloads. For marketplaces, you need metrics for both supply and demand.

Airbnb: "Rebook rate" — % of customers who rebook after first booking, plus "nights booked per year" for existing customers

Uber: Weekly rides per rider, average rides per week

Pick the right cohort period

Daily, weekly, or monthly — depending on your product's usage frequency. Shorter periods for high-velocity products, longer for low-velocity.

Airbnb: Annual cohorts (people don't travel often)

Uber: Monthly cohorts (high-frequency usage)

Identify the initial user action (Period 1)

What does 100% of your install base do that indicates revenue potential?

Airbnb: Booking at least one night

Uber: Completing their first ride

Calculate follow-on retention (Period 2+)

What % of the install base still engages in the next period? This is your retention curve.

What Does "Good Retention" Look Like?

The 3 tests for good retention:

1. Stable long-term retention. Your retention curve should flatten and run parallel to the x-axis. A dip after the first period is normal — but the curve must stabilize, not keep declining.

2. At or above median benchmarks for your vertical. Stable retention of 10% is great for e-commerce but terrible for a social network. You must benchmark against your specific category.

3. Newer cohorts perform better than older ones. If your Q4 cohort retains better than your Q1 cohort, it means you're improving your product and value proposition. This is the strongest signal of product-market fit evolution.

📊 Retention Benchmarks by Vertical

Vertical	Period	Long-Term Target	Median
Social Network	Month 12	45% - 65%	55%
Subscription	Month 12	25% - 35%	33%
Travel	Year 2	20% - 35%	29%
On-Demand	Month 12	20% - 30%	22%
E-commerce	Month 12	10% - 25%	16%

Source: YC Continuity team analysis, Second Measure credit card data

💡 What If Retention Is Bad?

Don't hire a growth team to acquire more users. Instead, apply a growth approach to retention. Stitch Fix hired a retention-focused PM to run experiments improving retention before investing in acquisition. Fix the bucket before you fill it.

Step 2: Build Your Growth Team

Once retention is solid, it's time to build a dedicated growth team. The experts converged on a remarkably consistent formula for the initial team:

🏗️ Year 1 Growth Team Formula

Growth PM

2-3

Growth Engineers

1-2

Data Scientists

~15

Total Engineers (Trigger)

Timing: Most companies made their first growth hire when they had about 15 engineers. The growth PM is typically the 3rd or 4th PM on the team. The most common CEO mistake is waiting too long.

The Ideal Growth PM

Your growth PM is the most important hire. 100% of experts agreed on the must-have traits — and the interview revealed some surprising patterns.

Data-Oriented (Must-Have — 100% of experts)

Intensely data-driven and inquisitive. As one expert put it: "The scariest day is when numbers are down. The second scariest day is when numbers are up and you don't know why."

Prior Growth Experience (90%+ of experts)

Experience at a company focused on growth in a competitive space (e-commerce, dating, gaming, social). This means you won't recruit from Google or Apple — those teams didn't scale through competitive growth strategies.

Former Startup Founder (Bonus — 60% of experts were founders)

Former founders think independently, take risks, and persevere through failure — critical when most experiments won't work.

Existing PM (Bonus — 40% were promoted internally)

Someone who's already built social capital internally can accelerate cross-functional collaboration. Facebook and Slack promoted existing PMs. Airbnb and Uber hired externally.

The Ideal Growth Engineer

Self-Starter

Proactively generates hypotheses and runs experiments. Infinite curiosity — constantly asking "Why?" to uncover hidden insights.

Doesn't Cry Over Lost Code

Comfortable knowing that a large amount of work won't make it into the final product. Experimentation means most code is throwaway.

OK Doing Things That Don't Scale

Engineers with 2-4 years of experience often fit better — less attached to rigid requirements and roadmaps than senior engineers.

Great Communicator

Growth engineers work with design, copywriting, data, marketing, and product. Cross-functional comfort is essential.

Step 3: Your Growth Team's First Year (5 Initiatives)

The first year sets the foundation. Get these 5 things right and you'll build a growth engine that compounds.

1. Set an Absolute Goal (With the CEO)

The most important word is absolute. Your goal cannot be a percentage change ("improve conversion by 10%"). It must be a concrete number ("achieve 5M first-time room nights this year").

Facebook's growth equation:

[x] MAU = [A] New MAU + [B] Retained MAU + [C] Resurrected MAU

Airbnb's growth equation:

[x] Room Nights = [A] New users + [B] Existing users

Critical: 100% of experts said the CEO must be aligned on the goal. The goal must be communicated company-wide. CEOs who wait too long or don't fully endorse the goal severely hinder the growth team's first year.

2. Identify Growth Channels

Start with two questions about existing user behavior:

How do customers find solutions to this problem today?
How do your best users use your product? Can you get more people to discover it the same way?

User Behavior	Channel	Examples
Need to connect with others to use	Product virality	Facebook, Slack, PayPal
Users talk about the product	Referrals, Community	Uber, Airbnb, Dropbox
Search for a solution	SEO, SEM	Airbnb
Look for expert inspiration	Content, Affiliates	Stitch Fix, Glossier, Intercom
High LTV users	Paid acquisition	Airbnb, Expedia, Uber

Key insight: ~70% of experts said referrals were the top channel in the first year. Most products find only 1-2 channels that really work early on.

3. Establish Systems & Tools

Four elements are critical from day one:

Clean data set — Track key metrics and goals with no ambiguity

Segmentation tools — Understand customers at a granular level

Experiment dashboard — Track experiments, control groups, and statistical significance

Peer review process — Biweekly reviews to discuss and challenge findings

At scale, a company runs ~1 experiment per growth engineer per week. 100% of experts built their own internal tools eventually, but you can start with Mixpanel, Optimizely, and Chartio.

4. Establish User Research

Data alone can't answer everything. Your first 100M users will look very different from the second 100M. The experts recommend:

• Solicit real-time feedback from users
• Use session recording tools (Inspectlet, Hotjar) to track UX
• Meet users outside your core market — SF is not the world
• Pay attention to international cultural nuances (e.g., photo-sharing norms in Japan differ from the US)
• Document every use case — what's normal for one group may be alien to another

5. Continue to Iterate

Growth teams that run 100+ experiments per year report that only a third turn out positive. The success rate is 20-30%. This isn't failure — it's the scientific method. The point is to encourage engineers to take more risks, knowing most bets won't pay off but the ones that do will be massive.

The golden heuristic: "Don't test things you wouldn't ship to everybody."

Where Should the Growth Team Sit?

This was the biggest source of debate among the 25 experts. Three models exist, each with trade-offs:

Separate Department

Head of Growth reports to CEO. Growth is its own org, like engineering or marketing.

Pioneer: Facebook

Pro: Clear ownership. No ambiguity about who owns MAU growth.

Con: Can create friction with product team over feature changes.

Within Product

Head of Growth reports to Head of Product. Growth and product are in the same org.

Used by: Uber, Airbnb, Slack

Pro: Growth experiments can directly influence the product. Better alignment.

Con: Growth priorities may compete with product roadmap.

Within Marketing

Head of Growth reports to Head of Marketing. Often evolves from performance marketing.

Used by: Traditional companies

Pro: Natural extension of existing acquisition teams.

Con: Most experts cited this as the least favorable option — rooted in the past.

📊 Growth Team Composition at Scale (100+ People)

Engineers

50%

Data Science

15%

Design

15%

PMs

10%

Mkt/Research

10%

The End Goal: Growth in Your DNA

The best growth programs don't stay siloed in one team forever. They permeate the entire organization, making an evidence-based mindset part of the company's DNA. When every team — from product to marketing to customer success — thinks in terms of experiments, metrics, and retention, that's when compounding growth truly kicks in.

As the YC Continuity team puts it: "This is the newest frontier in the cross-section of marketing and product, so it's still evolving. When done right, an amazing growth program will permeate the entire organization."

🚀 Find Investors Who Understand Growth

The best growth-stage investors don't just fund — they connect you with growth talent from their portfolio. Search 100K+ verified VC and angel profiles on Datapile to find investors with experience backing high-growth companies.

Search Investors →