[00:00:00] Henrique Cruz: The most important thing to increase retention in Churn is just to bring the right people in.

[00:00:08] VO: How do you build a habit for your product? You need to invest in it. And we saw these different… You don't just gun for revenue in the door.

[00:00:15] Andrew Michael: This is Churn.fm, the podcast for subscription economy pros. Each week, we hear how the world's fastest growing companies are tackling churn and using retention to fuel their growth.

[00:00:26] VO: How do you build a habit forming product? We crossed over that magic threshold to negative churn. You need to invest in customer success. It always comes down to retention and engagement. Completely boosts the strategy, profitable and growing.

[00:00:43] Andrew Michael: Strategies, tactics, and ideas brought together to help your business thrive in the subscription economy. I'm your host, Andrew Michael, and here's today's episode.

[00:00:53] Andrew Michael: Hey, Henrique, welcome to the show.

[00:00:54] Henrique Cruz: Thank you so much for having me, Andrew.

[00:00:57] Andrew Michael: It's great to have you. For the listeners, Henrique is the head of Growth at Rows, the spreadsheet where data comes to life. Prior to Rows, Henrique was the head of product at Veniam and prior to that, a Product Manager at Talkdesk. So my first question for you today, Henrique, is how is the homepage experiment going?

[00:01:14] Henrique Cruz: Better than we thought. The initial idea was that we wanted to test if removing the homepage reduced friction and got more people to try the product and activate and so on. I think just a couple of numbers to show what the experiment looks like now is we've tripled the conversion rate to sign up. So we used to convert 8% to 9% of people who visited the website on desktop to sign up. Now it's in the high 20s.

[00:01:40] Andrew Michael: Amazing.

[00:01:43] Henrique Cruz: 20, 24, 27 today. So that, I think, is one of the big KPIs. And I think another one is the number of net activated users per week grew four to five X because of that. The activation rate as a percentage went down a lot because just the top of the funnel is much larger now, but the actual net number of people, which is what you really care about at the end of the day, just really increased.

[00:02:07] Andrew Michael: Really increased. Amazing. Just for a little context for the listeners, I visited Rows' website earlier just to gather some context information about the company products and it was immediately put into essentially the product where you can start creating spreadsheets and you automatically generate an anonymous ID for the user. And then I'm assuming that when you sign up, you save the history as well. So if I had to leave the page and come back, I'm still the existing anonymous user.

[00:02:34] Henrique Cruz: Yes, exactly.

[00:02:35] Andrew Michael: Very nice.

[00:02:36] Henrique Cruz: We store, exactly. You have an anonymous ID. We store part of that data in the cookies so that if you come back tomorrow, you still have the same spreadsheets that you've created. All of your data is there. Some features are gated, so for it to nudge you to sign up, but the core experience that we do have to do a better job at communicating this. A lot of people have this question, will I lose my data if I come back tomorrow? I think that's part of the iterative process of the onboarding. But, yeah. That's how it works.

[00:03:00] Andrew Michael: Nice. At first I was like, Okay, where am I now? Have I actually questioned like, Did I sign up for Rows before? And now I'm just logged in automatically and then I could see the sign in button. But, yeah. It's interesting that you see sort of that conversion rates. Something similar I've actually experienced previously with a startup of mine where we did exactly the same thing was like, how could we ungate the experience so that you could actually use the product before you signed up? And similarly, we were almost like we went from around 8% to around 24% conversion rates through a few different iterations, but it definitely has a big impact on that top of funnel and be able to convert more users. The hypothesis then took us through a little bit about that because obviously this is a very big activation experiment and like, where did this come from originally?

[00:03:52] Henrique Cruz: Yeah. Taking a very small step back, we're building a spreadsheet, right? We're competing with excel, competing with google sheets, and this means two things. One thing that works in our favor and one that works in our like, this favor. The thing that works in our favor is because we're building a spreadsheet, our audience already understands the base product experience. They know what a table is, that you can add columns. They know formulas. They understand the concept of a spreadsheet and what you use a spreadsheet for.

[00:04:21] Henrique Cruz: So a lot of the explanation of what the product is, we don't need to do that. The downside is that everybody already has a free spreadsheet on a tab open in their browser, most likely. And so the idea for the experiment was, Okay, if everybody already knows what the spreadsheet is and they already have a spreadsheet opened, let's try to reduce this fraction as much as possible and get them already into the spreadsheet environment that they already know, and do a good job of showing them why they should try something that is special about roles. In our case is, importing data automatically from their sources, using AI to crunch the numbers, and just the spreadsheet is much more beautiful.

[00:04:07] Henrique Cruz: And so we thought, Okay, how can we do this? We had seen a few other examples like Excalidraw, this tool that you can use to do whiteboarding that also doesn't have a homepage. We have a few friends with startups that have similar funnels and who told us their experience. And we thought, Okay, let's just put this out for a few weeks, see what the results are. If it doesn't work, this is a two-way door. You can easily revert back to having the homepage, and let's do that.

[00:05:24] Henrique Cruz: And what we found out, I think we actually now come up with what we call a login-less formula, which is if a product is a single player… If you have a good single player experience, if your business model is a freemium, meaning, you don't have to monetize everybody, if people don't need a lot of initial setup to use your product and the market is large enough, I think it's an interesting experiment for you to push. A lot of B2B products, people can get value on their own, and, therefore, this experience doesn't really work. Or there's a lot of sales engineering pre-work that needs to be done for you to get value or the market is small, you're really targeting a small vertical, so you really need these conversion rates to be really high and the three trials to convert really well. And so I don't think they translate very well, but for a product like us, I think it was a worthwhile experiment.

[00:06:10] Andrew Michael: Yeah, makes a lot of sense. I think one of the things that came to my mind was just sort of like the engineering requirements to run an experiment like this and we'll go, but it sounds like from your end, it wasn't much of a big uplift to be able to unlock this and enable usage.

[00:06:26] Henrique Cruz: It's a relatively small lift because you're basically showing the product to everybody. So what do you need technically? You need a way to automatically instantiate an account without an email. You need some logical anonymous IDs. You need a way to store some of this metadata in your browser so that you can come back to it. And then you need a few paywalls or gated experiences, like if you try to share a spreadsheet, we just show you a menu saying that you can, nudging you to upgrade. And then we need a way for when you upgrade that fires an event that says, Look, this old anonymous ID now is this workspace ID, we transfer all of the data to this new workspace, and this is it. The engineering piece wasn't terribly complex.

[00:07:12] Andrew Michael: Nice. And there was actually another question that I had then as well. So you've created this experience to essentially land you inside the app. As a first time visitor to the site, how would I then… Two questions. One go about discovering the actual features and is that expected just through usage, and then, the plans and billing. So I think looking in is quite hidden at the moment to be able to figure out how much I'll be paying cause you mentioned that as well. I already have a free spreadsheet. I wanna know what I'm going to be paying for something like this, and is it going to be worth it? And so how are those two challenges handled through the experience?

[00:07:53] Henrique Cruz: I think those are two valid questions, and the way we think about it right now is on the first one, people have a really short attention span. And so if they land on the product, we really want them to do one or two things. We know that what's really correlated with activation and retention is you importing your own data into your spreadsheet, connecting Facebook ads and pulling a report, connecting BigQuery and writing a SQL query. And so what we try to do, and we're actually iterating on this now in the next couple of months is on the right side, we have a small, mini website-like experience of Carousel that tells you what the product is. And then the first actionable screen is inviting you to connect the data source because we know that this is really correlated with retention.

[00:08:35] Henrique Cruz: And then after you create your data source, then it gets fuzzier because there are different things that you can do with the data and we still have to do a better job at that. But the core action that we want you to do right now is import data from one of your data sources. On the pricing and the rest of the features, its welcome screen has a few… I have these core actions, like connect the data source, and then, do you need inspiration? Visit our homepage, book an onboarding call, watch our tutorial video, which is an attempt of, Okay, if you need a little bit more guidance, here are three resources that will help you. And that is on the pricing side that mitigates most of the questions on the... I think where we actually get more questions is still, Okay, I landed here, what makes you special versus Google Sheets, particularly or versus Excel? I think we still need to do a better job there.

[00:09:25] Henrique Cruz: And the second thing that we're iterating on is actually building more of a set up moment. So maybe asking you one or two questions in this initial widgets to prepare for your onboarding better. Maybe recommend templates. If you tell us that you use Google Ads, we recommend the Google Ads templates. Or if you tell us that you're just browsing the spreadsheet, we maybe do like a small tour. We're still... I am skeptic that we can increase the signup conversion rate much more, honestly. But I think what we can do is of the people who sign up or of the ones that experienced the product, they are more likely to activate. That's where we can–

[00:10:00] Andrew Michael: We can focus energies. Yeah. Now I think it was definitely, high twenties, really best in class conversion rates for any kind of product, no matter the audience. So definitely, working your way down the funnel then and trying to see, Okay, which other metrics can be improved from here is important. The point you mentioned as well, I think this is probably not something for everyone. I think it's a very interesting notion as well. And it's actually part of the reason starting the podcast in the beginning is itself and you hear these best practices and then everyone's like, Oh, Rose did it, we should do it as well. But there's also a cautionary tale that you can't just take one strategy and place it in another company.

[00:10:42] Andrew Michael: In the example that I gave as well from my previous company, like the market was huge. Like, there was an endless number of users that tend to be also prosumers, SMBs, XOs. There was a large, large volume of them. But something like this, as you say, might not necessarily work or it probably more than likely is not going to work if the audience is small or the use cases are maybe a little more complex or it needs multiplayer to be effective.

[00:11:07] Andrew Michael: So that's really like that multiplayer, single player notion and how you think about it, Rows, I assume as well, because the power of Rows isn't really a single player. Like, you want to have that collaboration aspect too. I mean, I'm making assumptions now, but how do you think about transitioning those use cases and through the onboarding and activation? Because I see you have the invite team member here and there, but what does that work like for you?

[00:11:35] Henrique Cruz: It's an interesting tension we find because you have types of products where they are obviously a multiplayer experience for the whole organization, like Slack. If you install Slack, it's really gonna be like, it will replace whatever other messaging system you have most of the time. Or if you start using Figma, it's really for your whole design team. There are tools that are really going to be used by the whole team. But then there are products which are a hybrid of single player and multiplayer. And I think a spreadsheet is a little bit of that because you can get a lot of value from the spreadsheet on your own, like you can build a report and an analysis that is multiplayer because you share the link with someone.

[00:12:11] Henrique Cruz: But it doesn't necessarily mean that this person is working together with you. And at the same time, it's really hard to penetrate the whole team with Rows and just becoming the default spreadsheet because maybe the finance team really likes Excel and there's never gonna move away from Excel, and these other team uses app scripts and so they're not gonna move away from Google sheets. And so we might have a small team inside an organization using Rows and maybe it's three seats or five seats, but it's hard to permeate it to the whole organization.

[00:12:42] Henrique Cruz: And so what we try to do is, what are the use cases that can translate better from single to multiplayer, even if it's within a smaller team. So for example, within marketing, someone from performance marketing starts using us for building their reports, then they take their charts, they embed into Notion and someone else from social or SEO or whatever, sees this chart and really likes to look and feel and asks to be invited, and then you start getting a small motion here. You might not translate to the finance team and the CFO using it, but we can have every organization having a smaller team. And then once the product is good enough, it starts expanding from there. So that's kind of a little bit of attention there because we have hundreds of teams, like, one person having a lot of value. And then, like, how do you invite the second person who is the second person? What's the motivation there?

[00:13:30] Andrew Michael: Yeah. It's very interesting. I think as well, to speak to it is essentially you have this product that has broad utility. It could be used in many different ways. It obviously excels in specific areas. You have a definition of your ideal customer profile. But then within the customer profile, you have personas that are going to be the champions and influences and where they're at today. So how are you going about discerning this and figuring this out? So you highlighted two use cases now within marketing, what does that process look like? Because obviously there's endless ways that you can use a spreadsheet as well, and that's the beauty of spreadsheets. And how are you thinking about prioritization when it comes to not only building the features that go into that to serve these use cases, but then also, when it comes to the activation and onboarding and which direction you're pushing users into?

[00:14:15] Henrique Cruz: I think there are two layers here. One of them is the product layer. And in the product layer, we've been very intentional to keep the product very horizontal. This means that we only build features that make it as a whole, a better spreadsheet for analysis and reporting. The thing that is vertical is integrations, right? We build integrations for marketing. Someone asked us for a TikTok integration, we built a TikTok integration, which is useful for the CFO, but it's good for our ICP. So I think that's kind of one layer. But then there's the marketing and communications layer and the templates that we build. And then we surface closer to our ICP. We see that Rows is perfect for marketing agencies, performance marketing teams, growth marketing teams. And so we do a lot more effort there on communication, on content, or SEO to try to attract these people while the product keeps evolving to be a better product overall.

[00:15:09] Henrique Cruz: So the type of customers that we have, we have marketing teams, is really, really strong. Then ops teams, so people who want to connect to their own databases to do analysis, if they want to connect to MySQL databases or a custom API or something from their own environment. And then we have what we call the Notion community because Rows plays really well with Notion. And so we have a lot of people like product managers for whom Rows is the spreadsheet counterpart to Notion. Because you can embed stuff from Rows into Notion, you can connect Notion databases to Rows, so you can import the database to Rows, crunch the numbers, build pivot tables, embed it back into Notion. So it's kind of these three key types of personas that use the product.

[00:15:54] Andrew Michael: Nice. Can we talk about that last one a little bit more in detail? Because I imagine that's also potentially a good growth channel for you. How is that set up today? How are you driving growth through this channel? And then, was this a deliberate action taken as a path circuit? Is it treated as an acquisition channel versus just a good tight partnership?

[00:16:15] Henrique Cruz: I think our strategy with embeds is similar to how YouTube grew on top of MySpace a bit. I always make this analogy. If you think back to the early 2000s, YouTube explored it as an embed to MySpace pages because the community was already on MySpace. They wanted to customize their pages. People came across the YouTube embed, and then they would sign up, et cetera. And the way we think about it is, okay, the spreadsheet is only part of your toolkit. And very often, the spreadsheet is a middleware. It ends up somewhere. It's a screenshot that you take to PowerPoint. It's an analysis that you put on an Ocean docker conference.

[00:16:50] Henrique Cruz: And so we've built this beautiful embed element that you can take a spreadsheet, and you can put it on a homepage for a calculator landing page, or a lot in your wiki. And Ocean is kind of the primary wiki there. And so this works really well as an expansion mechanism more than it doesn't work extremely well to acquire new workspaces or new accounts, but to expand the accounts because you're a marketer. For example, you do an analysis, it has a bunch of charts, you want to embed them into a notion doc because this is where a team is. You use Rows, there's a watermark. If you're on the free plan, people come across it, they can comment on the embedding it like it starts this whole dynamic inside the embedded elements, which is very good for dissemination of the Rows brand inside the team.

[00:17:36] Andrew Michael: Yeah. Interesting.

[00:17:39] Henrique Cruz: And it's, yeah. It was really one of the breakout features. I think over the past 40 weeks, it was growing 8% week over week. Then now it's tall over the summer. But it was really a big channel for us.

[00:17:50] Andrew Michael: Amazing. Yeah. And it's like, it makes sense that maybe it's not an acquisition channel as such because in my mind, I was thinking like, How would you onboard new Notion users onto the product service? They would need to have discovered it somehow outside of the Notion ecosystem, unless they're really going into searching through the available add-ons and things like that, which is probably a much smaller audience of the total.

[00:18:14] Henrique Cruz: Yeah, we do that, but a bit more organically. For example, someone does a video on how to create pivot tables with Notion and we are the tool to help you do that, and we get some top of funnel there or we try to do like some articles on our blog perform well. For example, how to build a social media tracker in Notion, and we show you how to do it in Rows, and then embed that into Notion. But I think it's marginal. I think it works better when you share something inside the doc, your team sees that they like it. If the product is good enough, then they will convert.

[00:18:45] Andrew Michael: It becomes expansion.

[00:18:46] Henrique Cruz: Yeah.

[00:18:47] Andrew Michael: Nice. And you mentioned a few different roles as well. And I'm keen as well to understand like... So the ICP then sounds a little bit marketing, marketing agencies, so forth. Where does the data team, data analysts, fall into this? Do you potentially maybe see tension between the two? Because often it's like, another place for analysis and another place where the source of truth now no longer exists. So is this something that you see as well? Did you deliberately not mention analysts?

[00:19:17] Henrique Cruz: It depends more on the size of the company, honestly. For large corporations that use ROS, it's really hard to expand within those teams because of IT. Essentially, just getting a purchase order in, and we also don't focus a lot on that. For other smaller teams, let's say, if you're a tech company and there are 200 people or 250 people, maybe you were already using an add-on on top of a spreadsheet and you weren't really happy with it or you had a bunch of CSVs laying around and you're moving that to Rows. And so I think, maybe, there's a bit of shadow IT, and the IT, sometimes, doesn't like this very much.

[00:19:54] Henrique Cruz: But on the other hand, if you're connecting, for example, directly to your MySQL or BigQuery database, at least the IT team or the infrastructure team knows that you're querying the right place. And you're not fetching some CSV that someone sent you over Slack. And that's where you're making your analysis on top of it. I think it's a bit of a hybrid. With these things, it's always a bit like that. But at the end of the day, what we find that's really hard to fight against is the spreadsheet is really the only natural data environment that any business person can understand. And so it's hard to fight against that.

[00:20:30] Andrew Michael: No, because I think that's definitely where everybody gravitates back to, even you can have the most sophisticated data and set up. Like, most people feel natural and comfortable in the spreadsheets and they're gonna go back to it, say, ultimately for the quick ad hoc analysis that they do.

[00:20:45] Henrique Cruz: I'll tell you one interesting statistic that we have to... Part of our goal also till the end of the quarter is to improve on this, which is even though we have these integrations, let's say 10%, 12%, 13% of new users connect an integration. But, 14%, 15% of new users, the first thing that they do is they upload a CSV. But the CSV is not exactly... You cannot refresh a CSV. It's not as beautiful as an integration. And so we don't do a great job on helping you on the next step with what you do once you have a CSV in.

[00:21:17] Henrique Cruz: Now, this week, next week, we're releasing a big update to our AI analysts. I think we will solve a lot of this. But I think to your point, we've built a really strong, simple data layer that you can pull data from. But, people naturally just put files there, because they have hundreds of CSV files laying around. And if you do a better job just analyzing your data this way, I think it's also an interesting market for us. Because even today, even though we push it to connect and integration, just put your files in and try to make sense of it.

[00:21:46] Andrew Michael: Yeah. No, I think it's also interesting is in the notion of the sort of PDF space. And I think small PDF comes to mind in this, where they make multiple different products all around PDFs, all around individual use cases for PDF. And all of them were effectively powerful individually, and then collectively as well. It feels a lot like, the spreadsheet is like this little compartments of superpowers that like you can just isolate and just use it for uploading and analyzing a CSV or it can be something a little bit more complex, which is connecting a data layer and a data source as well. Nice. What has been something that hasn't worked well with this experiments of changing the homepage sec, something that you maybe perhaps got surprised by when you rolled it out.

[00:22:42] Henrique Cruz: One thing that happened was a lot of our existing users got a bit confused because when we rolled out the experiment, let's say that you hadn't signed in in a week or two weeks, when you went back to Rows, your accession had expired and so you landed on a new account, but you didn't understand it was a new account. And so we had people telling us that they started building spreadsheets, but they couldn't find their old spreadsheets. So then they signed up to a new account and it wasn't reconciling with their old account. So that happened a little bit, especially the first few weeks after we rolled out the experiment. And we were not expecting that, or we didn't think, I think, hard enough about that.

[00:23:23] Henrique Cruz: I think the second thing that is negative about this is you have to be comfortable with the fact that you will know less about your users when they first come into the product. Right now we're doing the discussions of the Q4 quarter roadmap, and we're gonna do another iteration of the onboarding. And there are some things that we would like to do to know about our users that we just don't know in this experience. We don't know their email. We don't know if it's a big company or a small company that we should nudge them to an enterprise or if we should show them the pricing at hand. We don't want to ask too many questions, like too many set of questions, because that detracts from the log in of the experience of it. So there's attention there. It is possible that for some companies we're just leaving money on the table because this experience is freer than otherwise they're used to. That's tension.

[00:24:17] Andrew Michael: Yeah, for sure. And I could see that as well, like for large companies, especially like if you need to build a case and need to put together a wire rows and it's not easy to access and find information to be able to share with your team and like if you share the link, like if I go and share it with the CFO in the land or the like CTO, whoever it is and they land in this place, like, what is this thing? And it will take them time to figure things out as well. So I could see that.

[00:24:42] Henrique Cruz: That's why we offer visiting the homepage because it still exists at rows.com/product  But of course, when you type rows, you land on the product, which I think it's a beautiful experience. It should be, gives you a value right away, but it's a bit off putting cause it's not what you expect.

[00:24:57] Andrew Michael: I didn't see if you're running another experiment, I didn't see on the rows.com place to visit the homepage.

[00:25:04] Henrique Cruz: So you didn't see it?

[00:25:07] Andrew Michael: [unclear] Nice. Ok. So that makes sense. How are you thinking about that then, though, as well from a context? Obviously, there's services where you can do reverse IP lookup and you can get a sense of where the traffic's coming from and so forth. Are you doing any of this under the hood just to gain a little bit more deep insight into who is visiting? Or are you thinking more from first party dates and just asking questions?

[00:25:29] Henrique Cruz: I think the high level thing of geography is where are they coming from on top of G4. Not, nothing of these invasive lookup reverse IP type of tools. No, we don't do any of that. We try to look where the geographies come from. And honestly, like if 20 something percent of people sign up after that, we just have to do a good job with the onboarding and giving you value there. I mean, these are thousands of people per week. And so I think what this has been telling us is big improvements on the product have a big direct impact on, our North Star metric is power workspaces. So how many people use the account three days or more in the last seven?

[00:26:10] Henrique Cruz: And that's really what we have to worry about. When we push a big update, when we improve the experience, how does that impact the power workspaces? One thing that works well for us is, because this is the North Star metric, every week we send an email to the people who wore power workspaces two weeks ago but were not at power workspace last week. They had a decline in activity. Maybe they were using four days, they went to one day, or four days, they went to zero days.

[00:26:36] Henrique Cruz: And we just asked them, you know, we noticed your workspace came up in a recent analysis. We noticed that you're not using the product as much. Do you mind sharing in a couple of minutes what happened? What could we do to make you come back? And we have a good response rate from that. And these are all valuable data points because people use several days a week, they stopped using. We really have to know why. Is it a product issue? Did you find a better alternative? Was this a one-off project? Did you go on holidays? Is there any signal from this?

[00:27:07] Andrew Michael: Can you talk to that a little bit more detail? So how did you go about setting this automation up? What are the criteria that you've set as well to be able to collect feedback? Just a little bit more specifically.

[00:27:19] Henrique Cruz: So every Monday, again, the key, the North Star metric is how many people use the product three days or more in the last seven. From that, we just have a query on BigQuery that we export to rows with the IDs of the people that we lost as power workspaces. Then I just send in our email marketing tool, we use customer I.O. I send an email from my account asking them this thing that I just described, an email to, let's say, the 100 people that we lost this week as power workspaces, plain text email, asking them for a plain text response. And then I answer to every email, essentially. That's it. It's not crazily automated because it's not that much work.

[00:28:03] Andrew Michael: It sounded in my mind as well, like the volumes and you mentioned thousands of signups a week that potentially could be thousands of people you're sending these emails to.

[00:28:13] Henrique Cruz: Well, if we were losing thousands of power workspaces every week, it could be both a good sign and a bad sign. It meant maybe we had like hundreds of thousands, but no, maybe we lose, let's say 50, 100, 70, and that's it.

[00:28:27] Andrew Micheal: Yeah, because I think that like the criteria is quite strict in the sense like three out of the last seven days, like there's a lot of reasons why you would maybe potentially stop doing an analysis or using a spreadsheet, but still be like a power user and come back to it as well.

[00:28:42] Henrique Cruz: Exactly. And what we see is these, these people are much more likely to reply to your email because they've had some connection with your product, they use it for several days, they have good feedback to share. When we try to send surveys to people who use used it once and churned. We essentially have a 0% response rate. Like a few weeks ago, I was trying to do that. I was just doing a deep dive on everybody who used the marketing integration, because I wanted to understand where to improve that experience. And I was doing this on the day after, like today, I was looking at the people who used it yesterday for the first time and contacting a few of these people. Essentially nobody responds. Just have no vested interest in this and little connection to your brand.

[00:29:26] Andrew Michael: Yeah. Absolutely. I mentioned this a few times on the show. I think like, you're far better off asking people who've just renewed if they're on a subscription, like what almost stopped you from renewing rather than asking like, why have you turned? Why have you stopped using? Because like, as you say, there's no vested interest and also like, you're not even sure if they were a good fit to begin with. So collecting their feedback from potentially bad fit customers, like, it can also send you off in the wrong direction and starting to like prioritize.

[00:29:52] Henrique Cruz: Yeah. And I think it's even more true in the freemium model. If it's a free trial where, for example, people have to put up a credit card to start the trial, then they are more interested. If it's a freemium, someone just saw us on Twitter, signed up, started using. Doesn't even mean that they want to use a spreadsheet. Maybe they were just curious and they saw a tweet about us.

[00:30:12] Andrew Michael: Exactly. And especially, I think with the product Rows, like you do tend to get quite a bit of attention lately as well through like Twitter and other places. And they can, a lot of can just be like, it piqued my interest and I want to come check it out.

[00:30:25] Henrique Cruz: I'll tell you a short story. A few weeks ago, we had a content creator doing a TikTok about us in Latin America that went absolutely viral, as in, I don't know, five million views. And so that week, we had 30,000 signups from Latin America.

[00:30:44] Henrique Cruz: The percentage of these people who were actually interested in a spreadsheet like ours was like exactly almost zero. They saw this crazy tweet about, you know, Rows has this AI tool to analyze spreadsheets, which is partly true and very valuable, partly a little bit of exaggeration from the creator and people were just very curious. Maybe some of them will reactivate, but, you know, we've got 30,000 new people and very few of them converted in the end.

[00:31:09] Andrew Michael: And I think that's also like an interesting thing for freemium to work. You really need to have all the sort of little upsells and triggers and whatnot in place. Like you can't just say, okay, we need to do freemium, switch it on and expect it to work. There's a lot of work and craft that goes into understanding what should be the triggers and the points. And so how do you think about it from your perspective at Rows? Like, do you feel you're in a good space or do you think there's a lot of work and optimization around monetization to encourage conversion to paid?

[00:31:40] Henrique Cruz: I think most of the work is ahead of us. What we say sometimes internally is, I think it's much easier to build a 5-10 million ARR business on a free trial than it is on freemium. You're more narrow, people upgrade faster, etc., etc. And so I think a lot of the work that we have to do is on the natural conversion triggers, getting the right audience in, having a good experience, letting them establish the habit. And we've been iterating on the price maybe every six weeks or so, because there's always something that we're learning and we're like, we're still really far away from the best in class freemium conversion, KPIs. We launched our pricing in October of last year, so it's relatively new and kind of a tweaking experiment.

[00:32:25] Henrique Cruz: Just last two weeks ago, we added a new paywall based on the usage of the API. We were not, that was not part of the plan. We already have a few upgrades because of that. We're just learning what triggers the habit, how long does it take. I think freemium models maybe see this a lot, and we certainly see this, which is there's actually an inverse correlation between churn and how long it took for you to upgrade the account. People who upgrade on the first day of usage churn much more likely than the people who upgrade after 20 days.

[00:32:58] Henrique Cruz: Because if you upgrade in day one, it's normally because you either want to give the full product a try, like, let me just spend $20 and I just have access to the whole features, or you have a very specific one-off project that you need one of these paid features for.  Whereas if you upgrade after 10 or 15 days, you've already kind of established a natural habit of using the product. You're much less likely to churn. So premium has these things.

[00:33:24] Andrew Michael: Yeah. I don't know. That's a very interesting observation as one. I wonder if it does play true into other solos, but it makes a lot of sense in this case for Rows as well, like where, and I've definitely done it as well in the past of like where there's premium offerings, I just want like specific job I need to get done now and upgrade specifically for that. And then almost immediately like churn, once I've resolved what needed to get done. So…

[00:33:49] Andrew Michael:  I think that's also an interesting angle to think a little bit about when it comes to pricing and packaging and, perhaps there's like a way to package the product differently that enables these one-off use cases or project-based nature of usage and it's something similar. We like rappled with a lot at hot jar, because, hot jars will had two buckets of users, like one would be like these SMBs maybe where you're not launching features every single day. You're not doing like rapid experimentation and iteration. So you don't really need it continuously. You need it for these sporadic moments in time and project nature basis.

[00:34:23] Andrew Michael: Whereas like we saw like a lot of churn and reactivation as a result of that. And that's probably what you'll end up seeing as well within rows as you start to mature and a lot of people just, I use rows last time for this. That was fantastic. Let me go back another 20 and you start, and I can't remember who the guest was, but the, the, what they started to do is also like, trying to calculate the LTV of what they call these like project nature, like use cases.

[00:34:48] Andrew Michael: Because what they ended up did using is like when you started a certain level of volume that this there was value in this audience and it wasn't just like a one off like the LTVs $20. But actually ended up being like 120 or 150 as opposed to like maybe the 1500. There was a continuous use cases. So and then that became a big enough bucket that like reactivation became a thing and like pausing and stalling and like. So, yeah. Very interesting.

[00:35:13] Henrique Cruz: Yup. Every week we see a few reactivated accounts every week, people who used to be paid, churn, and then came back. Not really because we do a lot of effort on reactivating. I'm always very skeptical of emails for reactivation, but it's because naturally a project came back or they saw something about us or there wasn't the time to use the product again.

[00:35:35] Andrew Michael: Yeah, 100%. I think I'm also, not also, I am very skeptic of emails and they're working for reactivation. I think like an underused channel potentially though is like retargeting and advertising.  And especially like if your product has matured and it's been a few months since users came or if you know the reasons for churn was specific features and thing. I think you can be really targeted of like, trying to win back customers in that route. But like just sending emails and hoping that they, somebody's actually gonna bother reading it and then coming back is, it's a channel, I think is all just as much as the message.

[00:36:09] Andrew Michael:  What would you say is like the thing that's keeping you up most at night now when it comes to Growth at Rows?

[00:36:15] Henrique Cruz: There are two things. One of them is our ICP are these marketers. We really have to increase the top of funnel of people who want to connect the marketing integration. And that keeps me up at night because we're doing different experiments on content and SEO on influencers on just improving the base product. And you don't really know what's gonna work very well, right?

[00:36:37] Henrique Cruz: Because it's not a big natural environment where people like, it's not like a marketing agency is gonna go online and say how great Rows is to use for marketing agencies. It's not the type of virality. Henrique Cruz: Maybe it's a lot more on the templates or on the blog or content creators. And so that's one thing that keeps me up. And I have, how can we 2X or 3X or 5X the top of funnel specifically for markers because we do have a very healthy top of funnel but people who are looking for different things.

[00:37:03] Henrique Cruz: And then the second thing is, I think it's now very clear that LLMs and AI are perfect for data analysis tasks. Might be coding, might be spreadsheets. And so the world's just gonna be really different for spreadsheets like Rows and Excel and Google Sheets. And part of my team is just 100% focused on improving our AI analysts. And we have the feeling internally that if we ship something that's good enough. It's gonna be so transformative that a lot of these other works of the data, etc., might become less relevant overall. And so that also really keeps me up at night, which is, is this gonna be good enough? How, how can we iterate on this? How can we make this solve another capability that people, we see, we've been building this for six, seven years, 95% of people don't know how to use a V look up.

[00:37:55] Henrique Cruz: They, this is not how to look up, but they are all on spreadsheets. And so, if there's this AI layer that helps you do simple analysis much, much faster with confidence, you can really bring a lot more people into giving them more confidence to have to work with, with numbers. These are the two things.

[00:38:12] Andrew Michael: I see that from my own usage as well. I was like, now I do find myself sometimes just going straight to chat GPT, throwing in a CSV and asking for analysis back in there. I'll just go and validate it myself to make sure that things are correct. But like, the amount of time it saves is incredible.

[00:38:28] Henrique Cruz: And it's impressive. What we see is that because these tools are essentially code first, it's hard for you to then deconstruct or debug this analysis. And then you really want it in a spreadsheet environment where you're more comfortable with. But clearly, this is turning around, right? People uploading files, uploading data, and we just have to win this race.

[00:38:49] Andrew Michael: Nice. I could definitely see that being a very powerful experience is that like having what I do within chat GPT now. But actually built out in a spreadsheet so that I can go back and then trace back the formulas and see how everything connects the dots as well. And there's a few different like AI startups now doing this with other areas, like where, is it perplexity or one of the others with coding, Claude? I think Claude does this interesting way, like chat GPT outputs the code, whereas Claude actually outputs the output. So you can actually see the component of the sites. And I think like-

[00:39:22] Henrique Cruz: That [inaudible] cursor are really going full on agents. So I do back to the area.

[00:39:30] Andrew Michael: Yeah. At some point you think like everything's meaningless. It's all just gonna become a little limbs. And then you think, you know, like creating like really magical experiences around them is what's I think going to be like this next wave and figuring out how to distribute. Something like for me, distribution is probably the number one thing now because anyone can build anything now and you need to figure out how to actually distribute it to make an impact.

[00:39:55] Andrew Michael: What's one thing that you know today about churn and retention that you wish you knew when you got started with your career?

[00:40:01] Henrique Cruz: I think that the most important thing to increase retention in churn is just to bring the right people in. We try to fight a lot on once people are in tweaking the onboarding, showing them templates, we try to take or doing churn prevention. But really bringing the right people in is the first most important step to increase retention and activation. Most of the experiments that we've done on, for example, onboarding checklists, recommending templates, the effect on activation was very often negligible because we found that the people who were going to activate, they are most of the time going to activate anyway. A job of bringing people in. And then I think the second thing that we've learned over time that I wish I knew before it would have changed how we had built this product is, people are often really churn or leave because of details in the UX experience.

[00:41:07] Henrique Cruz: For example, you ship charts, but the charts aren't great. And therefore, there's something missing, and then you churn. But you're already working on the next feature. And so less product space, solve in a better way is better than trying to solve a bigger problem in the worst way, in these two things.

[00:41:27] Andrew Michael: Nice. Yeah. It's the irony as well. I just had a conversation this morning, busy building a house and I was talking to the carpenter and he just spoke to the second point that you mentioned is like, he started using a notes tool for a very specific thing and they changed the way it worked. And he churned because it wasn't what he expected to begin with. And as the product like seemingly got bigger and better, it became more and more difficult and less interesting for their use case to use. And, yeah, they're nice. And, it's been an absolute pleasure chatting with you today. Is there any final thoughts you want to leave the listeners with? Like anything they should be aware of, or how can they keep up to speed with your work?

[00:42:06] Henrique Cruz: We're very open on Twitter and LinkedIn with our experiences and experiments. A lot of these numbers that I've shared are mostly public. And so I would invite people just to go to rows.com on desktop because the loginless experience only works on desktop. Go there and really tell us what you think. Did you find what the product was about? Did you miss anything? Was it obvious? We're really constantly iterating on both the AI experience and the loginless onboarding. So just tell us what you think. Help us reduce churn. This is what I would like.

[00:42:37] Andrew Michael: Amazing. Yeah. And if you do churn, leave some feedback.

[00:42:41] Henrique Cruz: Exactly.

[00:42:42] Andrew Michael: Very nice. Thank you so much for joining today and I wish you the best of luck going forward.

[00:42:47] Henrique Cruz: Thank you, Andrew, it was a pleasure.

[00:42:49] Andrew Michael: Cheers.

[00:42:57] Andrew Michael: And that's a wrap for the show today with me, Andrew Michael. I really hope you enjoyed it and you were able to pull out something valuable for your business. To keep up to date with churn.fm and be notified about new episodes, blog posts and more, subscribe to our mailing list by visiting churn.fm. Also don't forget to subscribe to our show on iTunes, Google Play or wherever you listen to your podcasts. If you have any feedback, good or bad, I would love to hear from you. And you can provide your blunt, direct feedback by sending it to Andrew@churn.fm. Lastly, but most importantly, if you enjoyed this episode, please share it and leave a review as it really helps get the word out and grow the community. Thanks again for listening. See you again next week.

‍