Growth AMA with YC Partner Gustaf Alströmer

by Y Combinator10/29/2018

YC Partner Gustaf Alströmer recently did a Growth AMA with Startup School founders. It was so good that we wanted to share it here.

And here’s Gustaf’s Startup School lecture.


For two-sided marketplace, at which point do you recommend to start focusing on demand side? Do you recommend to grow the supply side first or focus on both at the same time?

Most marketplaces start with supply-side first. If you don’t have demand in the early days you should probably create an environment where you can mimic what it would look like if you had a lot of demand. For example, when Uber and Lyft would launch new cities they would promise drivers a certain amount of compensation if they drove during the weekend. Once they had critical mass of supply they would go out and get demand (riders).

Do you have any tips for leadgen / growth, for B2B SaaS startups? Right now all the available leadgen tools out there, costs thousands of dollars, which is expensive for startups!

This is what we recommend most YC companies who is starting to do sales:

  1. Figure out who is the actual buyer of your product. Who is the decision-maker that will pull out the credit card / sign the bill?
  2. Create a Google Spreadsheet, put 100 linkedin profiles in the spreadsheet that fit the profile of the potential buyings, then use a tool like Hunter to figure out their email-addresses. You can check the email format on their website if there is no response.
  3. Email them a plain-text email saying you are the founder of your company and you are reaching out a small group of people who you think would be perfect or the product you are building and you ask them to try it. You could also include a short gif/video (like really short) of how it works.
  4. Use an email software that allows you to check open-rates and click-through rates. If your email is opened many times it means its being forwarded around/getting attention. Email again if no-one responds.
  5. Expect about 40% to open your email. 5-10% to click on whatever link you included and less than 5% to try your product. If that works, then you just have do the same thing 20x and you have 100 users. The biggest mistake people use when doing this is sending too few emails.

I recommend using Mixmax, PersistIQ, and Streak but there are so many.

What are the 3 things you recommend startups to do NOW to get users when starting out?

This depends a lot of what kind of company you are. If you are a consumer company one good way is to try to figure out who have this behaviour today? For example, if you would start Lyft or Uber today and needed to get riders you would think about who takes taxis today Well, bar-goers and people arriving at airports take a lot of Taxis so going to a bar and handing out coupons or doing ads in the airport would be a good place to start.

If your users will be online I would actually recommend Reddit. You’d be surprised that nearly every single topic in the world have a Reddit community and they are often your most passionate users/harshest in the feedback. Another good places to start is Product Hunt.

Do you have any particular ideas on how to break into this community which can be notoriously ‘anti-commercial’?

That is a good question that many struggle with. You have to find a way to approach the community in a way where you are helping. Reddit committees are notoriously sensitive to commercial messages. Make yourself “one of them” and ask for feedback on your “project”. Language will be very important. Spend a lot of time in the committees to understand the unwritten rules.

What are the most effective ways companies can target and acquire early users? Are there certain marketing tools you recommend specifically for B2C companies?

There are only a few really large scalable sources of growth. For example word-of-mouth, SEO, Paid Growth and Referrals are good ones but often when you are small you’ll go after something non-scaling like asking people in a Reddit forum, meeting people in person and asking them to try your product. The most important thing is to try a lot of different non-scaling growth mechanics. Paul Graham wrote an excellent point on this topic.

What are your top questions to ask during a user/customer interview?

  1. What do you think this product does?
  2. Try to do x (for example buy something) and then say all the things you are thinking when you are going through your flow (say it loud if if just a thought you’re having)

What is your take on whether to try to get big corporations as your first customers or do you recommend to try to get mid size companies first and use them as a selling point to get the big ones?

Nearly always easier to start with smaller companies first. Start with companies that move fast and need your product but don’t have a “procurement team” yet, i.e. a company where the person who like/use your product actually can decide to pay for it themselves. You want to go after companies who can afford your product yet have the lower amount of internal friction to get it done.

Any clever strategies for solving the old chicken and egg problem?

In my experience there aren’t that many products and markets where the chicken and the egg are actually equally important. Usually there is one side of a marketplace that is more important than the other in the beginning and you just have to go out o your way to find growth of that side. In Airbnb’s case you need hosts to be able to have guests but you don’t need guests to be able to have hosts so you start with hosts.

How do you find them? PR was important in the early days of Airbnb but PR are hard to measure and hard to scale. It’s always a good thing in the grand scheme of things but not a very good source of new users for most companies. It’s better to build a direct channel with your users than rely on press. Then you can be more targeted in your communication

Paul Graham wrote a great article based on these early days of Airbnb.

Startup growth is actually a lot less theory than you think and more “doing”. There was a founder in the most recent batch that decided to visit a bunch of doctors offices to see what they staff thought of his product – that is actually a better idea than trying to answer that question online for some customers.

What is the single biggest mistake early stage (MVP) startups make in year one around growth?

PG covered a lot of these here.

The biggest misunderstanding about growth is that is somehow magically happen once you have a good product. It’s not the case. You have to do the intentional work in the early days (as well as in the later stage but different type of work).

How do you balance getting an early product in the hands of a nontrivial number of users without leaving a bad taste in their mouth if it is under-baked? If you turn them off the first time around, do you get a second chance once you address their complaints?

There is a lot of founders who fear that if people hate version 1 the game is over and as a founder I should just get a job. Reality couldn’t be further from the truth. Airbnb “launched” 3 times because the first 2 no-one cared. Turns out listening to customers if they don’t like your first version is very powerful way to iterate yourself to a great product. You can’t build a product in isolation so a “baked” product is basically a lie founders tell themselves. You should just launch. Launch early is a great sign of success.

How can we get our target users at a certain city to find and download the app when we are on a budget? Our advisor suggested Facebook ads and App Store ads, so we are planning to start with those.

Before you do anything related to ads I would try to do things that don’t scale. There is probably a way to test this hypothesis without buying ads. Start with your friends. Ask them how important a dish or ingredient is compared to other attributes of restaurants. You probably don’t need more than like 50-100 of friends and friends-of-friends to test this hypothesis. See if they use the app, do they use it again? Growth here is a function of building something people want – not trying to get users with ads. I worry in this case that you are not solving a 10x problem.

Do you recommend any specific strategies for asking users for a favor, e.g to share/review your app? When to do it and how to do it?

I think the best times to ask users to do something for you is when they are happy, i.e just accomplished something they found valuable in your product. Many of the “Please review my app” schemes are tied to a certain number of sessions to make sure only people who like your product will actually be asked to review it. When it comes to sharing, try to figure out the motivation why someone would share. That is often at the core when sharing is working really well vs. the actual sharing flow. For Airbnb, sharing worked well on sharing listings you might want to stay in with other people you were planning the trip with and Referral codes.

What is your advice on displaying streaks “you used the app for 3 days in a row, don’t screw it!”. Should we use it or do you find it more harmful in the long run?

I don’t think this actually is good for the world and don’t think it builds the right long-term incentives to use the product. You will fool yourself/your users to drive a metric.

How would you help users make something like journaling or exercising a habit?

If you figure out the exercise part you have probably found a big business (Peloton did). Figuring out how to correctly motivate humans to exercise seems still largely unsolved but with a big price – so people continue trying. I would try to help you understand the incremental improvements you are making. This requires a lot of real world testing with real users. You can’t solve this in abstract in any way.

Instagram put the label “You’ve seen everything below this line” to make people spend less time in the app, but feel more “full”. If the feeling of fullness is the key to make users happier than should we try to maximize user time in the app? What is your opinion on limiting access to apps?

I love this topic and many will be surprised to learn this is actually tightly coupled with growth. To me, anyone working on growth should understand the power of dark patterns and then build principles that allows the team to know when you are there or not. The great user experiences here are actually quite hard to measure. Someone feeling a social product is meaningful are harder to measure than how many days/hours they spend active on the product.

Any tools/products you would recommend to growth hack at the initial start?

Some good ones are: Mixpanel and Amplitude to measure your data and Fullstory or Inspectlet to record users sessions to learn from their behaviour.

One growth channel we are trying is conferences where we can meet multiple leads at once. What do you think?

Conferences isn’t very scalable. I have rarely seen it work. The way to think about sales is to find a scalable way to identify your audience and then a scalable way to contact them. Turns out email works for this in both small scale and large-scale but you have to send at least 100 emails before you know if it works. So just find 100 people who you think is in your target audience and then email them.

When using a referral program for growth, how do you sustainably incentivise a b2b referral program where your product gives the business using it a competitive advantage from other businesses?

It’s quite tricky to get B2B referral programs right. With consumer Referrals program it’s the audience is large and you have many more connections to other people like you. With B2B referrals you know far fewer people who would want to use that product. Solving that problem is the hardest problem. Gusto have the most impressive version of a B2B referral program that I’ve seen so far.

Do some products need some hard growth effort before it goes viral (eg. Facebook just spammed user’s address book)? What are some of the latest growth hacking strategies today’s startups can use?

Growth isn’t a set of hacking tactics at its core. Growth is understanding the value of a product and removing as much much friction as possible between users and people getting to that value.

The value of Facebook is that you can reconnect with your friends. Once you were able to see the names and faces of your friends you understood what facebook was all about. The role of the growth team was to get as many people as possible to see the names and faces of their friends as fast as possible. Using an existing graph like your email or phone-book was a natural way to start.

What are the key skills you would look for, as a startup, in making your first growth hire?

You should be your first growth hire. The founders should be the ones doing growth in the early days – it’s not something you can outsource since it’s so core to the company similarly that you can’t outsource product. Once you have a bit bigger team here are two good resources how to staff a growth team:

How do you demonstrate stickiness for an app (DAU/MAU or any other way) in a travel vertical, where you know that once a user books, they are not likely to repeat for 4-5 months?

DAU/MAU isn’t the best metric for all products. It works well for things you’re expected to use every day/month but for many products that doesn’t make sense – travel is one of those cases. Airbnb didn’t really measure DAU/MAU at all. We measured retention within a much longer retention period. What is the goal of your product? To get someone to book? Couldn’t you use bookings as a metric then?

When Airbnb introduced referrals, you would have acquired a lot of users who weren’t looking to travel at that time, but sometime in the future. What are the strategies you suggest to reactivate them or market to them, or any secret sauce you could share from your airbnb days?

If you sign up for Airbnb and got $40 on your account to travel for in the next 12 months we would email / push message you on a regular basis making sure you knew this. This made a difference and is worth doing.

Is growth rate in % is more important than absolute numbers? Does it really make sense to focus on for the early-stage companies?

The two metrics I think make sense for early stage startups are Weekly growth and Retention. Weekly growth can be averaged over a bunch of weeks if you have spiky growth but maintaining 10% weekly growth over time is very hard and a great motivator.

Retention is a great metric because it helps you understand if your users love your product and if they come back to use it/buy.

What’s your view on how to speak to business owners of small local businesses? For example hairdressers, restaurants and dentist. Owners are often busy and not always present.

In my experience this audience are better at answering the phone than they are answering email. Selling to them is hard though because it takes a lot of energy to make it work but the ROI is relatively low since these are often low-margin businesses. For dentists, email is a good strategy I would pursue.

As many of the startups are now planning for different sort of growth hack ideas, is there a moment where we need to stop doing growth hack and focus on organic growth?

The way I think about it is that organic growth is a result of a great product but you can always help more people discover the value of the product. For example if you have the best possible travel recommendation service that people love using but signing up is really hard then your growth team should prob spend a lot of time optimizing sign-up so that people who haven’t tried your product yet get to the product as quickly as possible. In that scenario you are still optimizing organic growth with growth engineering.

When does a startup need to have its own dedicated growth team? What positions or expertises are needed in this growth team and what do they do on daily basis?

Anu Hariharan wrote an excellent post here about this

Any book/article related to growth you would recommend us to read?

Here are some good ones:

Are there any thought out processes to balance working on new features vs acquiring users?

This sounds easier than you think to answer – but we struggled with it at Airbnb. I think the answer depends on the stage your company is in. In the early days you really need both and primarily build features that makes the product/usage growth. It’s easy for companies to think that just building more features will solve their growth problems but that’s not the case at all. Andrew Chen wrote a great post about this.

There will be a core set of features that drives your growth and nothing else really matters and your goal in the early days are to figure out what those essential features are.

Is there a difference between how to approach users to pay for a subscription vs a one-off product?

I think you are more likely to build a better business if you build a subscription service versus charging once. Charging once for your product is a relic from a world where software was only installed once. If you are building a subscription business then you can do a lot to optimize how long people stay subscribers but the most important thing is to make your product so great that people want to stay subscribers.

Do you have some tips on how to attract users from many different countries at the same time (on a low budget)?

For scalable growth channels most countries use many of the same platforms like Instagram, Youtube, Google and Facebook. You can reach people on all these platforms in most countries in the world. Only exceptions are China, Korea and Russia where the basic platforms are different. There is an over-emphasis on localization as a solution in my experience. Most countries are more similar than they are different and it’s that you have re-invent everything for every country.

What is the single best advice you can give in terms of Growth for a start-up that is transitioning from an idea to working prototype and then full blown launch?

Single best advice are different for different companies/stages 🙂

The most important advice is to read this article and internalize it

Would you sacrifice revenue for growth?

It depends. Hard to say without more context. I prefer doing a trial or giving away a small portion of the product for free than making the entire product free if you are selling to businesses. Then you have less of a risk of giving away valuable features that people actually would’ve paid for.

What metrics would you suggest as key focus for higher ticket B2C products?

Revenue and repeat purchase.

Any relevant thoughts for early stage from your time at Voxer, observing Mapbox etc.?

Voxer was very similar to Facebook in many ways. Messaging app that grew through the social graph.

Is building hype ahead of your first release important? If so, do you have tips for getting people excited and ready to signup?

Short answer: No, hype is not important. Don’t focus on building a hype around your product. Start small. Launches are overrated.

How might growth tactics differ when targeting not “the Internet” but very specific professions (e.g. plumbers, teachers, surgeons)?

In my experience Teachers and Surgeons are certainly spending a lot of time online and online is the best way to reach them. Agree that plumbers are harder to get – I would probably try to call them but like mention above people that are not online are a hard group to sell too. If it’s hard to sell to a group that might be a challenge for your business in general.

Re: Airbnb Referral. Does the value of a referral for a person receiving an invite, diminish when they learn that the person sending a referral code is receiving a monetary incentive for doing so?

Not in my experience. You want to design the language in a way that makes the receiver feel like it’s a gift. Play around with the Airbnb referral program. It’s one of the most optimized referral programs in the world and nothing is random.

Re: Airbnb Referral. How did referral user growth compound over time? What was your viral coefficient and how did you optimize it?

We experimented our way to double-digit % of all growth coming from the referral program. We ran many many A/B test on the program itself to make improvements. It compounds over times since the input in a referral program are your existing active users which also kept growing.

What should an early stage startup do about growth if unit economics is negative, so we lose money for each customer who finds and uses us? Should we seek funding, fix our unit economics, or pause marketing (or artificially restrict users via quota)? In which order would you take action?

You can’t scale a business with negative unit economics and you shouldn’t try to. You should fix you unit economics first, then do things that don’t scale to get early growth and then decide what to do next. But those are the most important things.

What key metric should be tracked and worked on for consumer/ social product that leads to long term success? Do VCs look at the same metrics?

Retention is the most important. Good investors all look at retention and cohort metrics these days.

As Airbnb growth hacked Craigslist, what are your thoughts on pushing the boundaries with other platforms T&C’s? What do you think about bots and scrapers to hack growth?

I’ve gotten the questions about the Craigslist hack many times but this was a very early and very non-important part of Airbnb growth – I joined early but I could never find any evidence of this being important part of growth. I would emphasize that there is a huge difference with doing that don’t scale vs things that scale. Breaking T&C certainly doesn’t scale.


  • Y Combinator

    Y Combinator created a new model for funding early stage startups. Twice a year we invest a small amount of money ($150k) in a large number of startups (recently 200). The startups move to Silicon