Startup School Reference Guide for Lecture #9 — Growth for Startups with Gustaf Alströmer

The lecture in its entirety. Recommend watching and then referring back to this afterwards!

While working to create the Snapchat for voice — — and attending Startup School 2019, I was inspired to create these “Cliffs Notes” for all the valuable content that YC provides to their attendees.

Table of Contents:

0. Why Growth for Startups?

1. Product Market fit & Retention

2.Growth channels & tactics

3. Making decisions with A/B testing

Why Growth for Startups?

The world is a really busy place and there aren’t lots of people waiting for you to launch your product.

They’re not standing there waiting to try it the moment you launch it.

To grow your startup at the beginning, you will need to DO THINGS THAT DON’T SCALE.

You will need to do completely unscalable things — like signing up users one at a time in person — at the beginning of your company.

  • This is a very uncomfortable and counter-intuitive thing for people who are MBAs and who have worked at big companies.

Founders are the ones who make startups take off.

Doing unconventional, scrappy things — oftentimes in person — gets your startup over the initial adoption hump. Once you’re over the initial adoption hump, then you can engage in growth tactics that scale.

AirBnB — how they did things that didn’t scale

In the early days of Airbnb, they flew undercover to New York, where most of their initial hosts were, claiming to be hired photographers for Air Bed and Breakfast. They didn’t say they were the founders because that made the company sound smaller. While one of the founders was taking the photos of the listings to make them look better, the other founder sat down with the host and asked them questions about the product. What are the challenges you’re having with the product? What are the things that are not working? Can you show me how you use the product? And by doing that, they got to meet the people that were their customers, which they really haven’t done before, and they got to see how they use their products. So they learned that this payouts thing didn’t work, or there was a big UI bug on this page, or it didn’t work on the Internet Explorer well. They went back to San Francisco. And they sent an email the morning after and it said, “Here are all the photos we took of your house. They’re now up on, and by the way, we fixed half of the bugs that you emailed us about.” That made the hosts love them, and those hosts became the reason that Airbnb eventually took off. That’s doing things that don’t scale. You can’t go and fly to meet every single one of your customers, but when you start doing that, you will learn things that you can’t learn sitting in front of your computer.

Sad, but true.Sad, but true.

Product Market Fit & Retention

Product Market Fit

Product Market fit can be defined in a lot of ways. But the best way to figure out if you have product-market fit is to use unbiased data to understand if you have made something people want.

  • Determine your top level metric — the one action that represents the value you provide to users above all else. What is the metric — the datapoint — that represents the value of your company?
  • Figure out how often users should be performing that action. Nobody is going to buy a house every day. Figure out the frequency of your top level metric.

See below for example companies, their top level metrics, and their ideal frequencies.


Retention is the best way to measure product-market fit.

Retention is how often people come back to your product. Repeat usage is the best, most unbiased way to figure out if someone is liking your product.

You can measure retention as:

  • the top level metric you chose on the Y axis (in %)
  • mapped across the X-axis (time) using the frequency (say, weekly) you determined was ideal.

But what is “good” retention?

“Bad” retention eventually drops to zero.“Bad” retention eventually drops to zero.

“Good” retention flattens out, with a large (usually >30%) of people coming back to use the product over and over and over again.“Good” retention flattens out, with a large (usually >30%) of people coming back to use the product over and over and over again.

Other (not so great) ways to measure Product-Market fit

  1. Net promoter score. Look up the best products and best companies in the world — they all have a bad Net Promoter Score!
  1. Surveys. Asking people how they feel — instead of seeing their behaviors — is going to introduce tons of bias. Surveys are great for figuring out specific ways to improve your product, but not to measure PMF.
  1. “How would you feel if you could no longer use this product?” Likely the closest you can get to asking a question to determine PMF, but still not as accurate as retention.

BAD ways to measure PMF

  1. Registered users. Says nothing about repeat usage, or if they liked your product or not.
  1. Visitors. Says nothing about whether your product is going to be valuable.
  1. “Conversion rate.” Who exactly are these people that you’re converting?
  1. “Customers that aren’t paying.” If you’re building something that’s paid, and the only way you can get people to use it is to give it away for free, that is not a good sign of PMF.

Growth Channels and Tactics

This applies when you have Product Market Fit.

If most people come to your product and never come back, Growth Channels and Tactics don’t matter.

There are two ways to Grow at Scale:

  1. Product Growth/Conversion Rate Optimization
  1. Growth Channels

Product Growth/Conversion Rate Optimization is working on improving specific parts of your product that you know are awesome, and getting more people through that funnel.

Growth Channels are marketing off of your product — say, Google or IG ads — to get people to use your product.

Product Growth/Conversion Rate Optimization

So, what is Conversion Rate Optimization? Every single step of your product experience is a funnel that can be measured. (It’s called a funnel because it’s wide at the top and skinny at the bottom.)

At every step, some people will drop off. If you count how many people drop out every step of the way, you will know exactly how many people make it from the start of the product experience to the end!

Example product experience funnel.Example product experience funnel.

What are some specific examples of Conversion Rate Optimization?

  • Internationalization
  • Authentication (sign-up flow)
  • Onboarding (new-user experience)
  • Purchase Conversion

Growth Channels

Again, there is no need to invest in Growth Channels until you are sure you’ve hit PMF.

Handy chart for the type of Growth Channel to explore for your product.Handy chart for the type of Growth Channel to explore for your product.

What if my product falls under all of the above? Should I focus on all of these channels?

Probably not.

Most companies grow huge using only 1 or 2 of these channels!

Specific Tactical Advice for Growth Channels

Referrals & Virality

Referrals are defined as a financial incentive to tell your friends about the product.

If word of mouth is a strong driver of your product, then referrals are going to be one way that you can amplify that word of mouth.

You can use a few optimizations to drive this virality, including:

  • Social Proof
  • Clear Value of the Referral
  • Urgency (time-bounding)
  • Exclusivity (Accepting an invitation oooh)

AirBnB’s referral email uses all of the above tactics.AirBnB’s referral email uses all of the above tactics.

Paid Growth

  • Don’t do it! (Unless you have revenue)
  • Figure out your Cost to Acquire a User (CAC). Most online advertising platforms (FB, Google) do this for you.
  • CAC / Payback time is the most important metric in online marketing! You can’t take all your money and spend it on marketing that you’re not certain is going to pay you back in the future.

Search Engine Optimization

This has changed a lot in the last few years. There used to be millions of websites ranking for tens of millions of keywords.

Now, really big companies started getting really good at ranking for all those keywords, so a TripAdvisor for example might rank for every single travel keyword that you can imagine. That’s hard for small companies.

Key Points for SEO

SEO is a zero sum game. You’re competing against other results for the #1 Search Result spot.

Top Keywords change all the time. For example, ASMR recently became popular. If you’re building something that has new language, you can rank highly.

How does SEO work from the tech side?

Same page, but the robot sees it much differently. Only text!Same page, but the robot sees it much differently. Only text!

Google only sees the text of your website.

  • To be good at SEO, you need to understand what text you are showing to Google, so Google can understand what the site is about!
  • If Google can’t understand what your site is about, it is not going to rank it.

Main lever for SEO #1: Things you do on your page.

  • What’s the title of the page?
  • Can Google read the page?
  • Does the page throw errors?
  • Have I done research on the keywords that people (not me) use to search for my content? Is that text on my site?

Main Lever for SEO #2: Off Page Optimization (What other big websites link to you, and how often)

Making Decisions using A/B testing

Most companies don’t have to worry about A/B test at all early on. That said, it’s a great decision making tool as you achieve greater scale.

**Example scenario: **“I want to launch a new design of our Home Page. I did, then the numbers went down. What happened?!?”

Typical testing of new features at a startup.Typical testing of new features at a startup.

🚨 DON’T. DO. THIS. 🚨 (There is a better way)

The only way to know if your change actually made the metric go up, or down, or have no effect is to have two separate parallel “universes” where the ONLY change is what you tested. That’s the definition of A/B testing.

Example of a truly scientific A/B test.Example of a truly scientific A/B test.

A/B tests help you make decisions in your company at scale.

  • The only way to really know what happened in an experiment is to run an A/B test.
  • This is hard to internalize because most people think of themself as good product thinkers!

As always, lmk if you have any questions or feedback!


Dave Goldblatt
CEO @wavechatme | Fb 2007-2017 | Crypto is cool | I like AI too | Knicks, Jets, Napoli, Yankees |