Marvin Liao on growth best practices

“Working on the right problem is more important than working hard.” – Caterina Fake

What is the most important thing for startup success? Is it having a good product? Solving a general problem?

The most important thing for startups, is to start with a vision.

Assuming that you’re making things people actually want, then beyond that, the best companies had a very clear vision of what they wanted to accomplish from the very beginning.

Marketing is the differentiator

It’s now expected that you have a great product, and it delivers a great user experience. The differentiator is now how you are found and how you acquire customers.

Don’t look at a successful company and study what they did in year five, look at what they did in the very first year. Your tactics will be defined by the scale of growth you are in. There is a wealth of great information out there (books, blogs, videos), but you need to filter based the content to fit with where you are at as a company.


In the early phases of a startup, growth hacking is really important. As your company becomes bigger, marketing becomes more and more important. With marketing, you can invest a dollar and know what you’ll get out – this is important when you need more predictability around growth.


Growth is a process. You have a hypothesis, and you run a bunch of tests, over and over again.

Story from Facebook in 2007: they were spending so much time and effort on bringing new users in, but those users were dropping out. They stopped focusing on acquiring new users, and spent an entire year looking at how to keep those users they already had. This ultimately fuelled growth, as those users who stayed invited more friends to join.

Figure out Activation, then figure out Retention. These two are the most important things you can do, even more than acquiring new users. This is contrary to what you commonly read.

Keep in mind: 9 out of 10 tests you run will be inconclusive. But that’s OK. You’re learning something small each time. You have to do this over and over again, even when tests are not providing conclusive results. This is hard, and it’s why most companies don’t succeed at it.