March 17, 2021
This week we’re having tea with Jamie Quint, former Head of Growth at Notion. Before joining Notion, Jamie was the Senior Director of Product - Monetization at Reddit, building out the revenue product org from scratch and leading it to 3x revenue growth. He's also a growth advisor and has consulted for companies including Twitch, Everlane, and Credit Sesame.
This interview has been lightly edited for length and clarity.
Mistakes founders make when thinking about growth
The best process for experimenting with a new channel
How to approach growth for products with a high learning curve
Attributes to look for in a good growth hire
To some extent, having a good product does solve for growth itself. But good product, in that sense, basically means a product that’s already been optimized to grow. In the case of Notion, it's a product that's inherently viral because one of the primary use cases is using it with other people. Facebook, for example, would have this. It's better if you don't use it by yourself. Notion's different in that you can use it by yourself, which helps because there's a single user use case. Then you can move into multi-user over time.
Having an inherently viral product does a lot of it for you. You can always do better than the base case of having a good product, which is never fully optimized.
People focus on growth too early. Growth can't save you from a product that doesn't have product-market fit. You don’t solve product-market fit with growth tactics and hacking. Some products have done that successfully, but some products did that and failed. Viddy optimized viral growth for this video product, by hacking growth in the Facebook feed. But they had no retention. Their growth graph went up and then straight down because as soon as the Facebook growth engine stopped, none of the users were users. They were just passive passing through. People make that mistake too often by not creating an innovative product that users love, and just doing growth hacking.
Rahul’s thing originated from Sean Ellis, which was that 40% of people would say they'd be disappointed if they couldn’t use your product anymore. Rahul took it a step farther. He was trying to debug the difference between the people who were in that 40% and the people who weren't based on a feature set: how do we work on particular features to move into a product-market fit?
A lot of times people build and launch a product and, suddenly, 40% of people are like, "I love this thing." You just have product-market fit from the outset. And then other times, you struggle to get there. The most successful companies are usually the ones who end up coming out of the gate with some product-market fit.
First, the product-focused levers for growth: What can you do in onboarding? What can you do with invites? What can you do with notifications? How do you integrate all those things into the product? I think of growth as how do you debug the problems that are preventing users from achieving their goals, and remove the obstacles, rather than trick people into doing things.
Performance marketing falls into a growth bucket where you're doing quantitative marketing and thinking about how to acquire users for a particular dollar amount. Another thing that I've seen fall under growth is a growth platform. You have all these software tools, from analytics to email to push notifications. Whether it be outreach on the sales side, or a Customer.io on the email and push notification side. You're doing marketing in parallel and have to coordinate audiences and own the underlying software stack to make sure all that stuff works with the data and data team that you have.
The base is just analytics. Everything can be built on top of that. For small companies, using something like Amplitude is great. As you start to get bigger, then you think about, how do I get a data warehouse in place? How do I organize all the data there? How do I do more complex queries than I can do in Amplitude? That comes past 15 people once you are getting to the edges of what something like Amplitude can do. For a SaaS company, have a good handle on revenues. Profitwell is an amazing tool and it's free.
The first thing is to ensure that the growth is legitimate: determine the retention. It's one of the most important metrics and you want your retention to be reasonably high, but stable. If that number is just going down perpetually, then you just have to keep acquiring users forever, which is not good.
Assuming you have that in place, then it depends on the product. For consumer e-commerce, start thinking about how to find more scalable marketing channels, rather than using word of mouth. With word of mouth, you only have so much control over scaling it as an e-commerce product. If you're a product like Facebook, then start thinking about what growth loops exist inherently within the product.
For example, Reddit's growth loop is people come to the site, read all this interesting content. Some percentage of them sign up and become engaged. Then some percentage creates content. And then people find that content via search or someone sharing a link. That drives users in and the cycle continues. You're either trying to find performance channels that work, or you're trying to find and optimize natural growth loops, within your product.
I educate myself by talking to people who have run the channel before. This is an underused tactic. A lot of people just jump in. If you talk to five people who've done this at scale before, they can just either tell you, "I wouldn't use that if I were you." Or, "this is how I would think about it."
I run some small-scale tests and see, specifically on performance, if I'm within 50% of the numbers I want to see. I'll assume that if we invest a lot more effort into that channel, we can get from something that's performing at 50% to what we want, 150%.
On the less paid side, I debug where the highest leverage problem is. In a given growth funnel, I map out the steps. For example, if you have a three-step funnel and you're losing a third of the people at the first step, and then 10% of the people at the second step, and then 1% of the people in the third step, most of the time, in the last step, you can only go from 99% to 100%. There's not that much improvement you can get. You can get a 1% improvement. On the middle step, it's much higher and on the first step, it's much higher. Here's where the most leverage is. How do we try and move this first? Because that's what's going to produce the largest gains.
They can. I did some math in terms of what matters more: retention or virality. A lot of times, people are like, "Well, virality is important because I want to grow exponentially." But you can grow exponentially faster if you improve your retention first, then focus on virality.
If you're improving a conversion funnel from paid advertising, at the end of the day, you're spending some money for a click, and then that can't convert at a higher than a hundred percent rate. Whereas, if you're driving a viral loop, you could get three users for every user, and that snowballs pretty quickly. Not everyone has a product where that makes sense.
New channels are where a lot of opportunity is. If you figure out a new channel first, it can be a sustainable advantage for you. A lot of successful e-commerce brands, like Everlane, figured out Facebook before others. It was a big part of how a lot of these direct-to-consumer e-commerce companies were able to grow early on. Back in 2013, Facebook released look-alike audiences, and then that became the way that, and still is the way, a lot of these e-commerce brands pay to acquire customers. It was 10 times cheaper six or seven years ago. If you can find that channel where it's not saturated yet and do it well, it can be a sustainable advantage.
It depends on your business. If you're selling toothpaste or something, it's just not going to pay off the first time, but it's a product that people are going to buy a lot. If you're selling mattresses, then you should probably pay off the first time, because it's just not a product that people are going to buy a lot. So, it just depends on the recurrence rate. It's good to aim for payback within a year.
A lot of people trick them themselves too. They're like, "Oh, I spent a hundred dollars to drive a hundred dollars of revenue." They're a clothing brand and the clothes cost $25 of revenue, and there are other costs too. They'll spend all this money acquiring customers, which works as long as they continue to raise money. But it's not a profitable acquisition when you take everything into account. Folks should keep more of a close eye on that. But about a year to 18 months is reasonable.
The standard playbook for that is you have SDRs in your outbound email, and they drive leads using tools like Outreach and SalesLoft. You drive them to talk to an account executive and you're trying to optimize this entire flow. The marketing becomes very enterprise-oriented. You're doing webinars and you have these one-pagers that are about specific teachers.
Every tactic gets overused. People do it because it works, then it stops working at some point, because it's overused, and then people migrate to something else. Usually, it comes back around, too. For example, a number of companies were successfully doing direct mail, over the last four to eight years. That was probably an oversubscribed channel at one point. And then, it stopped being an oversubscribed channel and started working again. It goes in cycles.
People are trying to do content marketing strategies on LinkedIn. In Enterprise SaaS, the marketing is harder than for someone like Lattice. Lattice targets HR professionals in the workplace. They can just do all this content marketing and events around things that HR professionals are interested in. It's easy to create a target profile for those people. If your target is just people at startups, it's a bit harder. The more precise your target is, the easier that will be for you.
Build a product that people love. There are some things that you can do early on in the course of your business, if you have some amount of product-market fit and believe things are working, to over-invest in the success of your customers. Notion provides support for everybody, regardless of if you're paying or not, and it's very high-quality support.
I don't know if that's sustainable all the way up until you're a public company. But in the early days, support is a feedback loop. You're learning from those people while you're providing them support. They're providing value back to you, and you can integrate that into your product development and be responsive to users. That is underrated.
I was talking to Merci, who ran growth at Slack, about this. Slack is a complex product. The base thing is easy, you just chat with each other. But then there are threads and private rooms and shared channels. You want to have emerging complexity: make your product as simple as possible for the base use case, and expose people to the complexity later on. If you have a product that's complex and people have to understand it before they can use it, it's just hard to make it that successful, unless you have people coming in who know they need all the complexity.
You want to look at your users who are the most successful and understand them, and then create more of those users. Do that qualitatively, more than quantitatively. It's one thing to look at a bunch of numbers and be like, "Oh, these users are great and do this feature a lot, let's try and get more people to do that feature." What has a better result is to talk to those users about their experience of the product. Then understand the gaps between those users and your other users and solve the problem more intuitively.
Andy Johns, who did early growth at Facebook, Twitter, and Quora and is now a VC, has this PDF that people should read. It talks about the difference between innovation and optimization and how to structure teams within a company. Everything at a company shouldn’t be optimized for moving some revenue number over the next six months. There should be resources at the company dedicated to product innovation where the output goal is improving NPS or other customer satisfaction metrics.
There should be another team which is growth, focused on optimizing rather than innovating, and just driving the existing product forward by tweaking the knobs and smoothing out the input flows.
These delight features would fall under the innovation side. The prioritization there would be, "Okay, we're going to invest this much engineering effort. Is this either going to open up a new class of users for us or make our existing users substantially more delighted with the product?"
He talks about not having a growth team until your company has 15 engineers and you can feasibly split up into three or four different teams.
This works best if you have a more specific target audience. Figure out the persona you're trying to go after and the content that will engage them. Figure out the channels that are going to be most efficient for distributing that content to them.
There are two strategies. The first is to hire someone who is senior and has the experience and could eventually build a team around them. That's the harder strategy. The easier strategy is to figure out early channels that you think will work for you and hire someone who's mid-career to run and optimize those channels. A lot of times people do not invest enough effort into trying to figure out channels before they're like, "I need to hire a growth person." Usually, the founders or someone on the leadership team has to suck it up and be the person to figure out and experiment with the channels early on.
It's really hard to outsource that. You can if you find the right person, but those people are really hard to find.
Some companies have performance marketing live under marketing. Performance marketing does better living under growth and collaborating with marketing. Marketing's role is to set the overall tone in messaging for the company as a whole, and when you talk about a SaaS company, you probably have product marketing responsible for the communication of product launches. You would have marketing that's responsible for building out landing pages and ensuring that they're on brand and consistent. It has to be a very collaborative relationship. In the best situation, you have a marketing team who's very responsive to the growth team and vice versa.
I used to think that these pure growth approaches would trump marketing and branding and who cares if it's on-brand. Maybe that’s true if the performance is 5X better. But there are a lot of times where you want to give up a little bit of performance to have a consistent brand and voice that you stand for as a company.
In my experience, I work with a marketing team to generate ad and landing page assets, it's very collaborative. We'll go back and forth and we'll come up with tests. The tests are of very high quality, not just hacking stuff together. If I was working at a smaller company, we'd probably be hacking assets together and seeing what we can do to be scrappy.
Everyone focuses on the initial launch too much and almost everybody over-indexes on a small number of examples, like Superhuman or Gmail.
If you have an email product, you should focus on the initial launch, and then if it's any other product, you should just get it out there and talk to users and figure it out from there.
There is an argument for people who are experienced to build a product, get it in front of a core set of users, make it good, and then launch it to a broader market. That's a strategy where you have to know what you're doing, or it's very easy to fail. If you're going to take that approach, you have to be intentional about ensuring you're talking to users on a regular basis and they're giving you real feedback that's accurate and not sugar-coated.
You want ambitious, but achievable goals, generally not even related to growth. I like breaking down the goal into more discrete monthly or weekly numbers and trying to hit those numbers. Some of the growth projects are longer-term, a month or six weeks to implement something. Ideally, you ship lots of experiments or growth features on a regular cadence to continuously make progress.
The first thing I look for is the ability to break down a growth problem into discrete components. That's important because you want to figure out where you have the most leverage, and to do that, you've got to break the entire problem down into discrete components, analyze how you're performing at each of those, and then think about where you got the most leverage.
The second one is prioritization. That sounds easy, but a lot of people are not good at prioritizing things by just basic effort and impact matrix. To be able to do either of those things you have to be pretty analytically minded.
For product hiring, SQL and data analysis should be a prerequisite for everything, except for maybe the most entry-level of candidates. It's so important to be able to do your own analysis as a PM and be able to think in data.
You can also send us a note at [email protected] We’d love to hear what you thought of the episode, or who you’d like us to have tea with next.
The Mercury Team