When I asked Andrew Chen about “startup growth,” I immediately realized my mistake.
“I think maybe we’ll use more precise terminology,” he kindly offered, before making a key clarification.
“I think of startup growth in two distinct stages that require very different strategies: Zero-to-one. This is where you’re trying to figure out, does the product even work at all? And one-to-n, where you ask, ‘How can we add in more channels and get very metrics-oriented to scale as quickly and efficiently as possible?’”
That formed the basis of our conversation, where he dug into:
- The key differences between early and mid-stage growth
- What startups get wrong with measurement
- How to build growth into a product at early-stage companies
- An expanded commentary on the “startup brand fallacy”
Q: What are the other key differences between those two stages?
Andrew: Zero to one is all about nailing product/market fit. You should be observing and interacting with users, while closely monitoring metrics like retention cohorts, frequency, monetization and others. There are many signals, but there’s one important question to answer: is the product experience working and providing value to users?
As you become more established and get ready to raise a Series A, the goal shifts to scaling your distribution. Many startups find that early growth channels are too small and they need to find bigger ones. The scalable channels tend to be things like Facebook, Google Ads, virality, SEO — channels that are much bigger and provide the high ceiling to continue growing over time.
An Uber example: Uber Kittens was a fun, effective early-stage growth tactic that encouraged employees to get their friends to use Uber. Not very scalable, though! Later stage Uber was much more about using paid marketing, SEO, and referrals.
It’s easy to push off questions about scaling until the future; after all, you need to nail product/market fit for it to matter. But your ultimate goal isn’t product/market fit — it’s to build a sustainable business. To get where you’re going, you need to think about later-stage growth from the early days.
Q: Does that reflect a “do things that don’t scale” mentality?
Andrew: Yeah I think that’s one of the things that resonates. When YC says do things that don’t scale, the question is: why?
And I think there is a lot of different reasons. One is you learn a lot about your users that way, if you are pre-product/market fit. Once you’ve hit product/market fit, though, you should focus on things that do scale.
That’s because your notion of growing your business changes over time. If you only have 100 users, getting 1,000 users is a huge deal, that’s a 10x improvement. But once you are at 10,000 or 100,000 users then getting 1,000 users seems small and you have to find some bigger channels.
Q: What’s the biggest mistake founders make transitioning from zero-to-one versus one-to-n?
Andrew: People have a sense that they need a theory for why their product is valuable, and it’s definitely important to have a strong thesis there. But what I also want to see is a growth thesis: a hypothesis around growth and distribution that is a deeper strategic insight about the ecosystem. This goes way beyond tricks or hacks and should be an integral part of your Series A pitch. I want to see:
- Why now?
- Why is it that this kind of product can work?
- How does it leverage some of the channels and platforms that are all of a sudden available?
An example: everyone understands that video is really big right now. More and more products are incorporating video — not just traditional video plays like YouTube, but also live streaming, Instagram stories, Snap stories.
If you’re an entrepreneur that’s working on something that touches video, it starts to get pretty exciting because if you can have a thesis that says ‘hey, this product that I’m using naturally generates video,’ then you have a thesis and overall strategy for how to grow your company.
That’s one of the fundamental reasons why eSports and gaming is working so well. When you are playing Fortnite, it naturally produces high-definition video that you can then stream to all these other services.
Q: What are some misconceptions about early-stage growth?
Andrew: If you are at the zero-to-one stage, you can’t optimize your way to success. You’re not going to go from something where a product is not retaining any of the users and then run a bunch of A/B tests and have it work. Ultimately, it’s about getting your product to a point where it’s sticky, you’re able to retain a bunch of your customers and they’re excited about it. That has much, much more to do with the value proposition in the market of what you are trying to do than anything else.
Q: How do you recommend measuring aspects of a business that are difficult – especially for companies in longer sales cycles? Brand, NPS, things like that.
Andrew: Metrics like brand and NPS have a time and a place, but they’re also lagging indicators. You can measure your NPS, but it’s not a lever because it’s an iterative metric that takes a long time to measure.
If you’re going to measure 90-day retention, are you going to measure annual renewals on your subscription contracts? Those are really hard because ultimately you don’t want to wait an entire year in order to figure that out.
An alternative method is to find predictors – metrics that correlate to later revenue. For example, before somebody decides that they’re going to not renew the next quarter, their engagement in the product might drop off. You can then use that engagement metric as a proxy for predicting churn and ensuring your customers are realizing value come renewal time.
NPS is useful, but it’s also a lagging, noisy indicator of defect that you could track more proactively. For example, at Uber we looked into one-star trips to figure out the cause. Is it coming from really bad routes, is it coming from the driver being late? Identifying and fixing the root causes of your biggest problems essentially fulfills the role of NPS, with more insight.
Q: What about measuring marketing campaigns that are difficult to attribute? Basically, anything that could be considered ‘brand.’
Andrew: I’ll be controversial – I think zero to one companies should not think about brand at all.
It’s what I call “the startup brand fallacy,” which I’ve written and tweeted about. In summary: brand marketing is mostly useless for consumer startups because they build a great brand by finding product-market fit and scaling traction. Brand isn’t a lever for early-stage startups — it’s something that happens after you’re successful.
Instead, they should just focus on getting the core metrics up and growing. Brand has a long investment payback; it’s only something that you should start working on once you’re sure the business is going to be around for a long time. Growing your products successfully causes a brand to emerge, and not the other way around. You have to make sure that you’re not confusing correlation with causation.
When you look at the very, very best consumer products in recent years, they all had years of obscurity. Every time someone thinks about overnight success of Instagram or Airbnb, there were actually years of hard work and many product pivots before it got to that point.
The counter point I often give, that I think is an interesting one, is the companies that are executing against brand marketing the best are probably the eCommerce companies. But when you really dig into them, you actually see that they spend most of their money on paid marketing. And so if their brand is so good, why do they have to give all of their money to Facebook and Google?
Ultimately, the goal is to build a great brand. But branded marketing activities are often ineffectual and you do them later on once you’re better established and can withstand lower-efficiency activities. Startups should focus on growing through product and initiatives that are highly accountable.
Q: How do you think about building growth into a company, and into a product?
Andrew: There’s a couple of things I would say.
I tend to think of most growth strategies more as loops than funnels. For a growing company, how do 1,000 new users who joined the product end up generating an additional 1,000 new users?
Think about Linkedin, Facebook, or any product that involves invitations. Their growth is based on getting people to come in and invite other people, which builds a viral loop. Products like Yelp or Wikipedia are operated by people coming in and contributing to (or editing) the content, which then gets indexed by Google and brings in more users to continue the growth. There’s also paid marketing loops and loops for other channels.
That’s a core way that I think about acquisition: loops or flywheels that can amplify over time, as opposed to something that has a one-time efficiency to it.
Q: Any other ways that startups can better adopt a growth mindset?
Andrew: Everyone understands how to think about their product as a funnel —signup funnels, payment funnels, etc. These are definitely the first priority. But early-stage companies are often surprised that there’s also a number of secondary funnels around the edges that are also really important.
For example, once you have a lot of users, you quickly find out that a lot of them can often be locked out of their accounts. Funnels like the “forgot password” flow can often be really important. Or if a user’s credit card is going to expire on your subscription product, you need to alert them in a funnel that successfully alerts them so payments don’t stop. It’s really important to think about all the edge cases because these are often the places where the growth is going to happen, but are commonly ignored by product-focused people.
Q: You’ve been on the side of startups fundraising, and now on the other side with a16z. What are the most important things you look for when considering an investment?
Andrew: I think the big thing that we’re trying to do is to really make sure that we ask the questions and we look at the metrics that validate that users really love the product.
At a16z, we often look at metrics like frequency of the product, intensity of the usage; we’ll read all app store reviews, look at DAU, MAU, funnels, cohort retention curve. We’ll look at what people are talking about on sites like Reddit and Twitter. The more conversant founders are about these topics, the more that we can get to the same understanding of the truth.
Q: Do you think maybe there are some founders that actually have a good hand but they don’t know how to play it well because they’re not as conversant in those terms?
Andrew: Yeah, I think so. Founders should know how they compare to other companies because there’s always a number or curve that will just blow your mind. When I joined Uber, it was the reverse-churn curve: new users would join and then use the product more and more and more. When I saw this, I said “holy s—, no wonder this is a multi-tens of billions of dollars company.”
Understand what sets you apart and highlights the strength of your product, and that story will be the catalyst for your next stage of growth.
Photo courtesy of Andreessen Horowitz