How to Get Users and Grow - Gustaf Alstromer

Nov 11, 2018 00:00 · 1274 words · 6 minute read Startup School 2018

These are my notes from “How to Find Users and Grow” by Gustaf Alstromer. An entry of my Startup School series where I follow the curriculum in Y-Combinator’s Startup School

Gustaf is the Product Lead of the growth team at AirBnB. He grew from 3 people to > 100 people today. These are my notes on what I thought were the interesting parts of his Startup School Lecture

Why is growth Important?

Some people think that if you make a product, and put it out there, people will find it. In my experience, that’s not how it works.

Story from Facebook

  • 2006 - datascience team at facebook predicted ~400 Million users by 2015.
  • 2009 - translations pushed growth to > 500 Million users by 2009
  • 2011 - mobile pushed growhth to > 1 Billion by 2014
  • 2014 - to get more people online to push growth > 2 billion

The forecast was wrong because it didn’t take into account what the growth team would do. Including adding translations, making Facebook mobile friendly, creating

Who is this talk really for?

A lot of this knowledge was gained from consumer companies. But these should be applicable to B2B companies, any company that sells things online.

This advice doesn’t apply to everyone yet. Don’t have a growth team if you haven’t found product market fit!

If you apply these things we’re going to talk about here before product market fit. Really bad things will happen. You may grow like crazy in the beginning, but the fuel you had isn’t fueling anything, you’re gonna fall right back down.

Product Market Fit

  1. Identify the metrics that represents the value my users get from my product
  2. measure the repeat usage of that metric

Examples of company - metric - ideal frequency

airbnb - bookings/stays - annual facebook - active users - daily/monthly gusto - running employee payroll - by-weekly/monthly

  • Now graph how many people used your product for that metric
  • If the line goes down you don’t have product market fit
  • If the line goes flat, you have product market fit

The blue line is a product with NO market fit, because almost 100% of people who signed up in week 1, are gone by week 15. The green line is a product with market fit. 50% of users who signed up in week 1 are still using the product. And no one else is really leaving

Here is real life example of a retention graph. 70% of users are still using it after 12 months, 30% are still using 7 years after signing up! This is Netflix. They have product market fit. they should now spend everything they have on growth

Growth and Marketing?

Traditionally, you had product team who made the product, marketing team who tried to go sell it. No longer the case. These people should be mixed, sitting right next to each other. maybe even be the same person.

Three ways to grow at scale

1. product growth

growth engineering - engineers changing the product to encourage growth.

Each feature of your product is a funnel. Try to increase the conversion rate optimization for each part of the funnel. For example, here is the funnel for your referrals system:

Here are some examples of how you can increase conversion rate for some parts of this funnel:

  1. internationalization - get more people at the first step of the funnel
  2. authentication - you’d be surprised how hard signing up and logging in is. go to airbnb whatever they have is the most optimized, these companies spend so much time experimenting and optimizing.

2. Growth channels

If you have < 50-500 users, you shouldn’t be looking at this yet. do things that don’t scale still

If what you’re trying to do, people go to google for. For example How to buy a house, you should focus on SEO. If what you’re doing is used everyday, you’re not going to go to google for it anymore. For example facebook.

  • it still matters in my experience
  • there’s a difference between what you see and what the search engine sees.
  • run your website through a seo browser, if you could only see this, would you still know what your product does
  • on-page optimization - things you can change on your website.
  • start with keyword research. figure out what people are searching for, try to match your site to high volume, low competition
  • seo experimentation - once you cover the basics, all you can do is experiment. Change the title of a page, see how it affects traffic after a week or so, etc
  • off-page optimization - other people linking to your website
    • use online tools to find who’s linking to you
    • whenever you get written up by someone, ask them to link to you

Do existing users already share your product through word of mouth? Incentives with referrals. other strategies for viral. A referral system for example!

Referral invite email example:

  • Airbnb referral system is the result of dozens of experiments
  • subject line has senders real name, people are more likely to open it
  • body title line has clear value, you get $40 on your first trip
  • body has urgency, sign up by may 35th to get the money
  • “accept invitation” makes it exclusive
  • has name, address, how long I’ve been using the product

can you make a list of people you want to sell your product to - you have a product to sell to doctors, you can make a list of all the doctors offices - you should go out and do sales.

  • create the list of your potential users, go and sell it to them
  • you should do this before SEO, or other channels if you’re able to get this list.

Do my users have high Life Time Value? - you should do paid acquisitions. performance marketing - facebook/google marketing.

  • you shouldn’t do it until you have revenue
  • CAC customer acquisition cost
  • ltv (life time value) & payback time
  • if cac is lower than ltv, then you’re good
  • attribution - you need to be able to attribute a new user to a dollar you spent

brand marketing - hard to measure, startups shouldn’t be doing this until they’ve hit the limits on the first two.

Making Decisions

Experiment. Most of the world will make decisions using their gut, they will either be lucky or wrong - mix panel

Use experimentation to validate decisions. a/b test. you’ll know the true difference between making this decision and not making the decision. otherwise you wouldn’t know if the result was from your change, or one of these other possible causes:

  • something that happened in the world
  • like the world cup, or the start of school

every-2-3 weeks sit in a room, and guess the experiment outcome. it’s fun, and it drives home the point that product decisions are very hard. if you don’t use experiments for these decisions, it’s the loudest person in the room and you don’t want that

iOS sharing sheet

  • control -
  • experiment - +40% more shares when using the experiment layout

airbnb trip invitation email

  • control - button says “join joes trip” -
  • experiment - button says “accept joes trip invitation” - +14% increase in trip joins with experiment

sharing icons

  • control -
  • round buttons - +18%
  • square buttons - +52% more shares with the square buttons

In the beginning, when you’re small. the decisions you’re making might be pretty obvious, but the moment decisions start getting hard, you need to start experimenting.


tweet Share