How to design a high-impact growth org for a PLG startup — Webflow & Dropbox’s Melissa Tan

 Maybe to set the stage for the conversation, we could start by talking about kind of comparing and contrasting your experiences working on growth at first Dropbox and now Webflow. What are some of the things that you've seen are quite similar and maybe what are the differences and what have you learned from those differences?

 I would say, you know, some of the similarities are, you know, the sense that both have really strong bottoms up adoption, both Dropbox and Webflow. Once users activate, there's just a lot of strong pull, lots of word of mouth, and they've both been really conducive to, um, this self-serve motion, this bottom up self-serve motion that's become really popular over the last couple of years.

That said, there are also a lot of differences, so, Dropbox in many ways was an early pioneer in product led growth. We had a freemium product that funneled tons leads to our paid consumer and B2B businesses. But back then it wasn't even called plg. So in many ways we were kind of creating our own playbook.

We grew a ton from just product virality, you know, the fact that it's a collaboration tool. We also had a wildly successful referral program. We dabbled a lot in partnerships. So at one point Dropbox partnered with Samsung where it came pre-installed on a lot of Samsung devices. And that was also, um, how Dropbox grew.

But we didn't really have this programmatic motion until our team started in 2014, and this was already quite late stage in Dropbox's journey. We were well over a hundred million in a r r. And so our playbook looked a lot like thinking about given Dropbox's volume. building, this high volume experiment machine where we would launch hundreds of experiments in a given year.

And this landed very consistent impact across our entire funnel and helped to reac accelerate growth leading up to Dropbox's i p o Web flow's Journey has been very different in the sense that we started our growth team here much earlier. Stage PLG now is much better understood, and many companies now are thinking about how they can drive growth more programmatically.

Um, and one of the main differences is even though Webflow has a lot of pull and that strong bottom of adoption, it doesn't have the same network effects and virality. in the sense that it's not a collaboration tool, right? So you're not necessarily gonna share your Webflow, um, website with somebody else, um, outside your company.

Or let's say you're a marketer building a Webflow website, you're not necessarily gonna share it with someone on sales. It's not a wall to wall tool. Um, And it's also, you know, not everyone has a need to build a website, so it's very different in terms of use case. There's a lot less volume than what we had when we were at Dropbox and, and the product is also a lot more complex.

So whereas at Dropbox it was really building this experiment machine at Webflow, our team has gone through the journey of really understanding deeply where our growth team can influence the user journey. And as a result, we've ended up working on a lot of levers that are less traditional for a growth team to take on going after bigger swings.

One of the things that jumped to mind as you were sort of just describing that is this idea of product complexity. Um, and I guess the other part of that is activation energy. When you think about somebody setting up a Dropbox account, it seems pretty different than what it takes to get someone to maybe create their first site with Webflow, the amount of effort, education, all sorts of other things.

 So how does that express itself in the way that you've built the growth function or how you think about executing on growth at Webflow versus Dropbox? I.

 Yeah, that was definitely one of the key nuances there where if you know how to use your desktop, you'll know how to use Dropbox. Whereas at Webflow, Webflow in many ways is a visual interface for coding. And so understanding concepts like the box model and c s s definitely help in your learning journey at Webflow.

And the other key difference here is when you have a longer journey for driving activation at Webflow, activation doesn't happen in one session. It's typically you're learning the product over multiple sessions. We've taken a different approach for understanding. How to drive activation versus, um, you know, other areas of growth where you have a finite point in time.

So for example, if you think about conversion and the checkout flow at that moment, someone's gonna either convert or not convert at Webflow. That activation journey takes multiple sessions. And so one of the things we did early on is we realized that the user research we did needed to be very specific.

We invested in doing what's called a user diary study where we followed user over their activation journey over a two week period. We asked them to screen record their, each session they had and talk out loud as they were learning the product. And this was super helpful for our team to ha to identify the specific insights of what was happening within that user journey.

That's like one piece that really helped us. The other things that we've realized as a team is that when you have a really complex product, not limiting yourself to just traditional product growth experiments of just, you know, an onboarding tutorial, the stuff you do needs to be a, taking much bigger swings.

And so our team has actually even done work to think through how we can simplify the product itself. So we're actually building features in the product, which is very different from other growth teams, which is just building on top of something the product team has already shipped. Um, and so I think the two biggest differences here is, you know, for companies that have, are working on growth where the user needs multiple sessions, getting the user re research that is really specific and gives you insights.

And the other one is you might need to take bigger swings rather than just kind of locally optimizing something that already exists.

 Can you talk more about the process of, of user diary study, like how it was set up, and then maybe some of the key insights that came out of it.

 Yeah. Um, the way we set up the user diary study at Webflow, we noticed that it takes, mult takes multiple sessions for users. So what we did is we recruited users that had the need to build a website and that fit our core I c p. We, um, asked them over a two week period to essentially screen record every time they were using the product.

We didn't force people to keep coming back, so you would also see natural drop off if users after one or two sessions gave up. So we did actually have some drop off in, in our user diary study. Um, and then we also interviewed users. Um, this time for our specific one, we did it at the end. You also could interview users throughout if you'd like, and that's how we conducted our study.

We found a ton of insights of just, so I would say the reason the dire study is really helpful is the learning journey at Webflow is non-linear. It's not like you sequentially need to learn certain steps to understand the product. And so we actually found what of the different pathways users learn, where do they reach hiccups?

Um, people talk a lot about aha moments. We were, we talked about Naha moments. It was like death by a thousand paper cuts at Webflow. David Hong on the team came up with that term and from the diary study we were like, okay, wow. There's a lot of areas where users are just getting stuck. We found non-intuitive insights that made sense in H sense in hindsight.

So for example, users would get even more frustrated if they started off of a template and then weren't able to figure out how to edit the template because they thought they were closer to figuring it out. And that was even more frustrating for them to be like, okay, I found this template. Oh wait, I paid for it.

I don't know how to edit it. So we got very specific insights, um, that we otherwise, you know, we actually did conduct user research at Webflow previously, but when you ask users to recall a previous experience, it's really challenging because, you know, you're really relying on them to remember every single session and all these things.

And so we were able to directly see in these user diary studies what's happening. Um, and as a result of that, we ended up prioritizing a few things. One of them was, there's one area in the product called layout where all users were consistently kind of getting hung up on. And so we decided we need to figure out how to teach users how to use this.

And actually we're working on simplifying layout. Um, as a result of the diary study, uh, we also realized, okay, um, we need to provide a bit more structure. Even though it's a non-linear journey, let's provide a little bit more structure in how we teach users the product. So we've also launched checklists within the onboarding journey that are different by starting from a blank site versus starting from a template.

 What was your pro. Process of digesting all of the feedback that came out of that and then turning it into some sort of roadmap. What, how did you figure out what to prioritize in what order?

 Yeah, we essentially, um, went through this synthesis because there were multiple users we had recruited. Not everyone had the same user journey. So we started to compile all the major themes. Here are the themes of where users are getting hung up. Um, these are the problems that we need to solve. These are just consistent problems.

And we started it from a very, you know, this very PM framework, but what are the problems we need to solve within the product? Quantifying and or I guess prioritizing the impact, prioritizing these problems by, The volume of users we saw and like the impact we thought it would have, right? And impact can be measured in many ways.

Here we looked at how frustrated users were getting, um, and then also looking at the converse. If a user figures something out and once they find something out, if it's an unlock, how can we help guide users to those unlocks? So we took a very, you know, standard PM approach of just looking at the different problems we wanted to solve, starting to prioritize those.

And then while we were running the diary study, because that took time, we actually were also building experiments in the product itself. These smaller experiments to test various hypotheses. And the things you can test for here are like what mode of education do users prefer? Checklist, interactive tutorial, do videos in the product help.

And so as you're also doing user research, I also think another great way to learn is just. You know, shipping low lift experiments that are easy to get out of the door, that are proving outta hypothesis, hypothesis. Um, so the, that's essentially how we took the synthesis. We also shared it broadly across the organization and that also ended up being, um, really helpful for other people on product to just even understand, oh, this is where users are getting hung up.

Just like building that user empathy too, in terms of realizing, wow, our product can be really complex or, and just seeing users through the videos. I think there's something just really real about that, that is very different from reading a synthesis.

 So maybe kind of going up a click. If you think about other companies who have a more complex product, who have more friction to get to that activation moment, are there a handful of things given what you figured out that you would recommend they think about or action on as it relates to growth given that specific shape of product?

 The higher level takeaways I'd have for other companies trying to solve this journey is, you know, really, I mean this is like maybe growth 1 0 1, but really looking at the data first. Where are there clear drop off points? And for us, we've realized that activation is just an early engagement within the product, is the biggest opportunity.

Something that we did try at Webflow early on was we also ran monetization experiments. That were kind of these typical no-brainer experiments that have worked at many other companies that all of us had worked at. They were wildly successful. They didn't work at Webflow, and the reason they didn't is we realized that's not the point of influence within web flow's journey.

If you haven't learned the product, you are either gonna be ready to monetize or not. And so really identifying where your points of influence are are really important. So if you have a more complex product finding, what are those moments in the product where people are making key decisions? That's one.

The second one are what are key unlocks during the onboarding journey and how can you help point people to those key unlocks early on? And the third one is really thinking about are there points in your product where if you encourage users to do something, it gets people more invested?

One interesting insight from us is we actually found that. If users add their domain early on, they were more likely to activate. And our hypothesis here is that they already were thinking about, okay, I have this website I've already thought through my domain. And so there are actual psychological things you can do in your onboarding flow if you have a very complex product.

You also need to think about what is a trade off for like, uh, getting a user to an insight. It's almost like being really methodical about the journey. How do you get users to an insight? How do you get them to a benefit and kind of encourage them to continue the steps of activating onto your product.

 So you, you just mentioned this a little bit, but if you think about, uh, points of influence these unlocks or getting someone unblocked or getting someone more invested, can you share some stories or things that have worked for you sort of along those three? Uh,

 Yeah. One that was interesting that I refer to is users adding a domain. So something that was really interesting is we launched a checklist for how to start from building from a blank site. And in the checklist, one of the actions is to actually, uh, add a domain and publish. And what we found is that that drove activation as well as monetization.

And so, I mean, it, it made a lot of sense. Hindsight's 2020, but at the time we were like, oh wow. People, once they add a domain and they monetize, they feel this urgency. That they need to figure out the product in a way that, Hey, you didn't pay webflow anything. You didn't add a domain, you don't feel as invested in the website.

So that was definitely, uh, one very specific example for us. Um, another thing that we then did, and this is something many growth teams do, is we took that learning, we said, okay, how do we use a checklist on templates? So learning a template is also very different.

Let's, um, also add a, a checklist to the template. Um, adding videos within the onboarding journey, which like we've invested a lot in videos, uh, our education team has where y we teach users Webflow. And so being able to embed videos has worked really well, especially because it's nice to be able to see those videos side by side.

Um, and then I would say there's a couple of bigger bets that. The verdict is still out on how they'll perform, but I'm quite bullish on them and one of them is this investment in layout and just making the product easier to use. We are ging that in a couple of months. And then the other one is actually leveraging ai.

So you know, we, a lot of companies now are thinking about how to incorporate ai. Our team is definitely because of our context on the user journey. We've been very passionate about thinking through where AI can help in the learning journey, whether that's, you know, through assistive AI or search or et cetera.

That's the other one where I'm quite bullish on that is also not something a growth team might traditionally take on. And so our team is actually very actively involved in web flow's strategy for ai.

 You mentioned this a little bit as we've been talking, but I'd be really interested to hear what you figured out in terms of goal setting or setting metrics in the context of growth. Maybe your kind of broader philosophy or things that have worked or not, or maybe areas you've changed your mind about.

 my overall thinking here on setting metrics is the growth team should be set up so that each PM or each pod has clear metrics they're driving and these metrics need to be. Uh, essentially leading metrics that are movable. And so a great example here is when I joined Webflow, we did have an activation metric, but it took users about four weeks to get to.

And it was, um, this concept of you've built a website and you've got 10 views. That was a clear success metric that meant you are more likely to retain. The problem is because the metric takes four weeks to get to a team can't move quickly to move that metric. Let's say we ship an experiment in early January.

If it takes four weeks to get to fruition, let's say we need to run the experiment for a month, that means we can't make a call till two months later. And so my overall philosophy here is the metrics the team is driving needs to be something that can move. It needs to ideally be something that you can measure pretty quickly.

And so one of the first things we did is we tried to break down that activation metric into something we could make a call on within one week. and this is something I've seen an evolution on too, where people will break out activation to an aha moment and then a habit moment and really thinking about micro conversions within the journey.

that is definitely one thing that was very relevant at Webflow. The other thing that I. feel pretty passionate about is in terms of a growth team owning metrics. There does need to be a regular cadence, I believe, for all growth teams to be syncing on the weekly metrics every week. So, um, you know, each team is driving certain KPIs as you are starting the week.

I, we always start the week by looking at the last week's metrics and the different trends, aligning on different initiatives that the growth teams are working on. Learnings from those and making sure we're just thinking about the user journey very cohesively. And so that's the other thing that we did at Dropbox that was really successful and we started to do here at Webflow, which is just looking at the metrics very closely.

In general, I think growth teams need to be very results oriented, so it's the, you know, you, you can't hide behind the metrics. It's important to be looking at those metrics very regularly. Um, and that should be driving, that should be how you measure success ultimately.

 Do you run that metrics meeting in any specific way?

 We essentially, um, the way we run the metrics meeting here, and it looks similar at Dropbox, is everyone working on growth and in this case, self-serve at Webflow meets at the start of the week. Um, and that is like growth marketing. They might be driving, you know, whatever your channels are, SEO, paid, et cetera.

It's, you know, lifecycle marketing, which is like email maybe if you do anything in app. And then product growth. If you have a product growth team, it starts with just looking at the overall metrics of the funnel, right? So looking at signups, you know, new customers brought in, looking at by segment. So you might have priority segments that you are paying closer attention to, looking at activation rates.

Um, monetization, expansion, whatever the team's driving retention. And that is the basis point to then discuss certain trends that we're seeing as a team. Not all trends are explainable. There's times when, you know, there might be in the summer, there tends to be seasonality or there's some weekly fluctuation, but usually there's action items coming outta that meeting to better understand certain trends.

For example, if you're seeing that a certain segment that is a priority segment is going up, trying to see is it, is it tied to some campaigns we ran on growth marketing. What comes out of that too is teams will end up talking about the initiatives they're driving, and that's like the, the second part of the meeting, which is initiative reviews.

We usually do async, everyone fills out bullets of the main things that are happening with links to specs or more detailed write-ups. People are also sharing learnings like, Hey, this messaging worked really well, or We just did a homepage test and this was a key learning. And that helps, you know, the team identify, oh, like let's try to test that learning as well.

Um, or like, let's say, you know, growth marketing is trying a new initiative for certain segment, we might identify, we should be collaborating closely on that and certain teams might start meeting more regularly to collaborate. Um, and so that's the, essentially the how it's structured first looking at metrics and then the second part of that is talking about initiatives.

And then there's often follow ups that come outta that meeting.

 When you think about goal setting or metrics, do you think about long-term versus short-term, and is there some optimal percentage? Or maybe put slightly differently when you think about some of the unintended consequences of metrics. One of them might be that you are only short-term oriented as a growth team.

Um, and so how do you combat that or think about that?

 Yeah, it's definitely a, a good question. We, I found that for that weekly sync, looking at an eight week period is a nice balance to not get so overly myopic and just looking at the week to week trends. And then as a team, we always do a step back for the quarter, for example, um, there's also a different meeting that we do with the leaders of the growth team.

And essentially our exec team at Webflow, where we do look at year over year growth trends. We look at that every week where we're looking at more long, both the short term and the long term. And sometimes in those meetings we look at longer term trends and identify, Hey, this is a longer term trend as a team, we need to keep a pulse on that.

Or maybe we wanna it, it also helps inform, you know, essentially what we might wanna work on for the quarter, or, you know, an annual planning. And so that's how I found there to be a good balance, which is the working group looks at a shorter, timeline. they also are doing like quarterly planning, quarterly look backs.

But then there is that other meeting, which is just a deeper dive that's like run by finance, essentially a deeper dive into our metrics, um, to make sure we have all our bases covered.

 Can you talk more about that meeting structure, kind of the team lead and e staff meeting?

 Yeah, so this one at Webflow, um, essentially is we do the team meeting essentially Tuesday morning, and it almost is like prep for that other meeting because it's a very detailed review of what we see on the team. Then Tuesday afternoon or Wednesday, we typically meet with, at Webflow here we meet with our exec stakeholder on Selfer, which is our cmo, Shane.

Um, and we look, finance runs this meeting. We look at all the, the metrics. Um, there's also like stakeholders for biz ops, uh, finance, et cetera. And that meeting essentially is a, more detailed readout. That's like maybe what you would expect to see in like a board deck, right? Um, it includes metrics that team can't necessarily control as levers.

Uh, but you should keep a track of, I'll give an example. One of them is a s p average selling price week to week. The team isn't necessarily running experiments to drive a s p, but it's a metric you wanna monitor. Um, and so this is like a more detailed financial readout of what's happening in the business.

Also, looking at how we're tracking to target, we do that in the team meeting, but this one is just looking more holistically as a company. Um, and we did something similar at Dropbox as well, where we did the team meeting. And then at Dropbox we did that review with our coo. Um, and then we did a monthly, we did this at Dropbox and well flow, we do a monthly readout.

Also. The monthly readout is with e-staff. Um, and that is like also a very detailed financial deep dive.

 What about the number of metrics? You know, you talked about a, a couple things there. One of the ideas, leading metrics and things that you can take action on very quickly is like one kind of idea. What about the number of metrics for a given individual team, broad function, and what you figured out there?

 really what I would say is when you're thinking about metrics for a team, ideally you have one North star metric, maybe two, you, you really can't be trying to optimize for too many things and it sometimes it's at the team level, sometimes it's at the, and often like every individual needs to know what their KPIs are.

That's for sure. I'll give an example of how we thought about it at Webflow. So, you know, We have an activation PM and this person is basically accountable for the activation rates at Webflow. We have also a monetization PM that's looking at conversion rates. Once someone signs up for free, what is that conversion rate to paid?

Um, and then there's also times when you have more, I would call it like quarterly goals. So we also work on pricing and packaging here at Webflow, and we'll have a goal around not owning a kpi, but saying, Hey, one of our goals is to drive. Um, you know, we did a pricing change last year, so we're trying to think through maintaining a certain, uh, ARPA or a s p, um, and or reaching a certain milestone by the end of the quarter.

That's, um, an example of how we think about more initiative based goals. And then growth marketing will have people on different channels and they have goals both on driving specific customer growth by channel and then also essentially staying within a certain CAC boundary. Um, And so I would say each person needs to have like, I would say max two goals.

At the, as a company though, you're looking at a broader set of metrics, um, that might even be cut by segment or are things just to make sure you're not kind of missing the bigger picture.

 So, and, and you mentioned this a little bit for Webflow, but if you go back to Dropbox, what, when you think about the growth teams broad, either North Star metric or few metrics that you always came back to, what were those and what was sort of the rationale behind them?

 Yeah, at Dropbox they were slightly different. One of them was, and I ran the B2B growth team, so every most purchases at Drop for Dropbox business starts with a free trial start. That's where we point many users to start a free trial. And so one of the North Star metrics was just free trials, and we did just volume of free trials.

We didn't say percent free to paid. That's because Dropbox had such a massive free user base, hundreds of millions. So if the team is trying to drive certain conversion rates, it would be really small conversion rates. So we found that our goal should be free trials. That was one. And there were multiple PMs working on driving free trials.

Second metric was what we called trial to purchase. This is of the free trials, what percent convert to paid, and that is what the onboarding pm or the activation PM was trying to drive with it, which is within that 30 day period. how can we help users increase their livelihood of Ping for Dropbox business at the end of it?

Um, we also had a team focus on expansions. So at Dropbox, most expansions come from seat growth, and so this is a team thinking through how to drive seat growth and owning a specific expansion rate. Um, and then there was also retention. So thinking through, at the end of a monthly or annual cycle, um, you know, the percent retained, um, of, of the user base.

And so we broke it down very specifically by funnel into what each, uh, either PM or each team could drive.

 In the case of Dropbox, when you think about those metrics, what's the process to sort of land on those and was it intuitively obvious or was there a lot of iteration to sort of land on these are the things that are gonna govern the business set priorities, that type of stuff.

 The metrics by the time we started, I think part of it was because Dropbox was already pretty mature, where some of them were very obvious. Free trials, for example, that was obvious. Trial to purchase was also obvious, just because if everyone starts a free trial, we wanna make sure they convert at the end of 30 days.

The ones that were more nuanced, I would say were expansion, which is, um, and even as I was leaving the company, I don't think we fully had a handle. It was harder for us to drive sea growth than we realized. And so figuring out what is something that. You know, obviously we know what's better for the business, but what is something the team can actually drive and what is a good quarterly goal for the team was something that we went back and forth on.

and then the other thing that we set that I hadn't mentioned is we broke down retention further by saying we also had a goal on driving a percent mix of annual conversions. That helps with retention. That just increases L T V. And so we also would break down and have specific quarterly goals on things that maybe we were trying to focus on as a company.

 Why was it harder to drive seat growth than you thought it was gonna be? That's pretty interesting.

 it is. The reason why is maybe similar to that nugget at Webflow, realizing where the points of influence are. What we found at Dropbox was we had this hypothesis that, oh, if we. Make it, make, make you aware of other people in your organization that are a freezer. Um, she share the same domain. Are you more likely to add that person to your team or not?

And what we found is that people either have the budget and we also would look at like account sharing for example. We'd be, oh, that team looks like they're sharing an account. Maybe they both wanna purchase. What we found is either you have the budget and you're already adding who you need to the team, or you don't have the budget, you don't wanna add more people.

6481080b8cb32a5077e7c876

Melissa Tan is the GM of Self-Service and Head of Growth at Webflow and formerly the Head of Growth and Monetization for Dropbox Business. She unpacks the nuances between the two PLG businesses, including the growth strategy, approach to pricing and packaging, and goal setting techniques. 