Lessons from Gorgias: How to Close your First 1000 Customers Based Solely on Data

Hello everybody, I'm Roman, I'm the founder of Gorgias, and so today we're going to talk about how to get 1,000 customers using data. So just as a way to start, I'd like to tell you a little bit about Gorgias to set the context. So we are a customer service platform for e-commerce websites, and one of the things we try to do is to automate customer support. When you want to automate something, you need to do two things. The first one is to gather data, and the second one is to build some workflows to provide this automation. And so in this room, lots of you are early-stage founders, and I'm sure you have lots of conversations with other early-stage founders about your problems, about how to grow the company, and so forth. And so as a founder myself, I've been having those conversations with a few friends. And every time I was talking about that, people were saying, hey, it's pretty interesting how you built a growth machine. We'd love to learn a little bit about that. So I was sharing tips around these topics. And so when I got to prepare this conference, I was like, OK, get to speak with 1,000 other founders. So let's talk about this topic with 1,000 people. So here we are. So 18 months ago, we had 15 customers. And so we were like this little blue tile at the beginning of the curve. And as any SaaS founder, you ask yourself, all right, where am I going to find the next 950 customers? And so we were thinking about that, talking with other people, reading the different literature. So like two books that were great to us were Predictable Revenue by Aaron Ross, and another one which is From Impossible to Inevitable, which was written by Jason M. Lemkin, the founder of Saastr. And what these books talk about is how to build a growth engine for your company that can repetitively help you grow. So I'm gonna talk about that today. So here we are, 18 months ago, trying to figure out how to build this growth machine for our company. So let's take a look about what the growth machine looks like. First of all, you have some inputs on one side, which are all the things you do. So for instance, they can be trade shows like this one, They can be email campaigns, people you hire, content you write, and so forth. So everything you do goes into your growth machine and then produces outputs. And so these outputs are, of course, value that is delivered by your product. They can also be money, so in this case, annual recurring revenue. And so what you want to be doing as a founder is to find a sustainable way to take the inputs to generate outputs in a repeatable fashion. And so what's really interesting about this topic is that when you are early stage, you kind of get customers a little bit from everywhere. Like, you talk to other people, you do things that don't scale, but you don't really know how you get those customers and how to get them in a sustainable way. And so there are different points in time when you can decide to build this machine to try to grow and accelerate. Most companies, when they get to $1 million, $2 million, $3 million ARR, they don't really have the machine yet. So they typically end up hitting a wall for a few months because they don't really know where the growth comes from. So they need to figure out where it comes from and then to figure out where to double down. So I would argue that since now we have access to, we have lots of data points about like trade shows we go to, the payment information, the sales data in our CRM, we can combine all these data points and start building the machine early on, as soon as you have 50 customers. So let's talk about that, and let's jump into how you can build the different parts of the machine in marketing, sales, success, and so forth. Let's start with the top of the pipeline, which is marketing. So here, what we want to do is build a growth machine that is based on data. But remember, we have 50 customers so far, so we don't really have data. So the question is, how do you build a machine that uses data when you don't have it? Simple answer. You buy it. So here are two platforms. Datanyze, that is a platform where you can buy data about a list of domains, like what tools they are using, how many followers they have on Twitter, like a different kind of— a whole bunch of information about these companies. And the other one is the Alexa Rank, which is the ranking of the websites on the internet. So in this case, like, Google would be 1, Saastr would be like 10,000, Gorgeous would be 100,000, hopefully less over time. And so we have this data about a target market. So in our case with Gorgeous, we sell to stores on Shopify. For those of you who are not familiar with Shopify, it's an e-commerce platform that is used by 600,000 merchants. So here we are, 18 months ago, trying to figure out who we want to sell to, to find our next, next 950 customers. Turns out Shopify is a publicly traded company, so you have access to a lot of information about them. So we looked at the list of all the stores, and what was interesting is that the yearly churn is 80% for Shopify. It's pretty frightening because at the end of the year, there's only gonna be 20% of the stores that were here at the beginning of the year that will still be around and still be your customers. So we don't really want to go after the people that are gonna churn at the end of the year. So we got this data. Now we can basically like build some correlations between our customers that we think are sustainable businesses to find among the 600,000 companies who are the ones that we want to focus on. So first use of data, find like qualified people. Another thing you can do is let's say in this room all of you are the 60,000 companies. If I go to you for instance and say, "Hey, are you looking for a customer service solution?" You could just tell me, "Hey, we looked at it like 3 months ago. We made a decision to go with a specific vendor, so right now I'm not interested anymore." So it's not the good timing. So we know who is qualified, but we don't really know yet who's ready to buy. And the thing is, you can reach out to everybody in this room and expect that only the ones that are ready to buy are going to respond to you, but it's not very— you're going to have like a pretty low conversion rate. So what you want to do is to find signals that mean that a group of people are actually ready to buy and in the purchasing decision right now. So in our example, we basically looked at what are these companies installing on their websites, like if they install like a chat app, and then we know that when you install a chat app on your website, it's likely that you are on a free trial for a customer service app. So now we have our 3 segments, the ones we don't want to talk to, the ones that are qualified but not ready yet, and the ones that are ready right now that we absolutely want to talk to. So we know who we think should buy from us, but we need to find a way to talk to them. So here is how we talk to them. For the ones that are ready, pretty straightforward, we just do outbound with emails, and we say, "Hey, I'm seeing that you have a, you're currently exploring different helpdesk options. Can we be part of this conversation?" Easy. The other one that's a bit more tricky is when we want to get the yellow circle, the 60,000 people, to be aware of our product even though they are not ready to buy right now. So what we did is that we just spoke with the 50 customers we have and just asked them, like, what are the things you do? Why do you talk to other merchants? Where do you learn about how to pick the projects you're going to focus on this year? And they told us, we go to trade shows. And we work with agency partners. Great. Okay. So what we did, we went to the trade shows. The first one was ShopTalk in Vegas. It's an e-commerce conference that happens in the spring. And so here we are walking the floor and asking all the vendors, a little bit like today at Saastr, like, what is the ROI of this event? Like, how can we make this a repeatable process? And the answers we got were a bit disappointing. We came to some people and they said, "Hey, we are kind of profitable with this show, but we don't really know. But don't worry about it too much. Down the line, it's going to help you build the brand, and you'll see some results in like 2, 3, 4 years." Great. The thing is, when you are an early-stage company, you cannot really afford to plan your growth based on something that's going to happen in 3 years, because you only have money for like 18 months. So you're like, okay, thanks for your tips, guys. We're gonna have to find a way to figure this out. So the approach we took was when we went to trade shows, every time we had a conversation with somebody who was interested to buy, we basically just sent them a follow-up email. Pretty simple. And I did like a BCC address that would log the conversation in the CRM. And so the cool thing is that let's say like a month after the trade show, you can see among all the people we had conversation with, who closed. So now you know, okay, I spent $10K on this trade show, I made $40K. That's already helpful. So we have like a little part of our machine, we know that if we go to similar trade shows, we're gonna make a 4x return on investment. What's even more interesting here is that after this month, we basically know who has closed, but at 6 months after, We know who's still here, we know what is the churn, we know how they like our product, and so we can make better decisions on who we want to talk to and where we should go to find customers. So let's take a step back and look at what comes out of it. So basically after we did that, we had like a sustainable way to generate revenue every month. So if we produce the same actions, the same inputs, we know that at the end of the month, we're gonna have the same outputs. All right, of course there is a bit of variation, like 10 to 20%, but that gives you a rough idea. And of course, as an early-stage founder, it's also like super comfortable to know that next month, there's still gonna be revenue coming in because the things that you do are predictable and you measure the impact that they have. So in this case, we knew that we would make like $250K AR out of it. So here are a few tips for those of you who are trying to build this growth machine. So one of them is that try to understand who your target customers are and try to think of them so that you can identify the signals that they send that mean that they are ready to buy. In our case, it was installing competitor apps. It could be talking on Facebook groups. It could be making a new hire. All that kind of things. And so when you know that, you know that your buyer is ready, and so you can double down on your marketing efforts. The second thing is that we are here to build a growth engine. So there is no real value in just doing something that's not gonna be able to be measured and not going to be able to be repeated. So everything we do, we want to measure the inputs and the outputs. Great. So we have leads that come in a sustainable way every month to our pipeline. Now, what do these leads do? They go to sales. And so what's great is that we have this data, remember, from DataNice, the AlexaRank, we have lots of data about our leads. And so what we want to do as early as the seed stage is that we want to have like the most customer-aware sales process that is possible. So what we do here is that we predict what the deal size is based on characters we have about the customer. So in this case, we assume that this nice company Sunski, which was, by the way, one of our first customers, they sell like nice sunglasses in San Francisco. And so these guys, we estimate that they're gonna pay us about like $20,000 a year. So we get this data from like Clearbit, which is an enrichment software, and Datanyze from the step before. And so now we know they're gonna pay us 20K. So what we can do first is like start to specialize our team. Say you have two AEs, one of them could take like the big customers, the other one could take like the small customers that need less attention and that needs like a faster pace. So we give the big deals to the best AE. Second is that we know what tool they are using. So instead of giving them the same pitch all the time, we can already tailor it to the company. Another thing is, in our case, we have dozens of integrations with all the tools. So we don't really want to do a demo that's like, "Hey, like, what are the tools you use?" Trying to understand a lot. Of course you want to do a bit of discovery, but you have this data already. So what you can do is like hook them up right away with the apps that they use, because you know what type of value they are going to get out of this demo with your sales team. The last thing that's actually pretty interesting is that we know where Sunski came from. They went to ShopTalk a year ago in Vegas. So the source is pre-populated here on the slides. But remember, we are building a growth machine. It's not only marketing, it's also sales, success, and product. So we want the data that is collected by sales to flow back to marketing. So in this case, say for instance Sunski didn't hear about us from an event, but they actually heard about us from an email they got or a conversation with an existing customer, we want the sales team to change that in the CRM so that marketing can attribute the credit to the right channel. Because at the end of the day, we want to know how much money is generating every month on each of the channels. And so what's the result of doing this? In our case, the close rate went from around 15% to 50% in the course of 12 months. So we tripled the close rate by having this hyper-targeted approach, giving the right pitch to the right people and talking to the people that are ready to buy. So a few tips here that you can use. Again, as early as the seed stage, you want to understand why your customers are buying from you. You want to enrich the leads before the demo. You want to assign the leads to the right person. And the last one as well is that We are trying to connect all the data points that your business has. So for example, if the customer starts paying on Stripe, we don't want the sales team to come in the morning and update the CRM. In our case, we have like the ACV is pretty low, so we need the deals to close fast. So we need the AEs, the account executives, to focus on the customers and not updating the CRM. So again, it's pretty straightforward to be able to automate the deals that move in your CRM that will save your team like about like 20% of their time. All right. So the Sunski is closed. They've decided to use Gorgias. Fantastic. So now let's talk about the experience inside the product. I'm sure a lot of you in this room have signed up for tools like Slack, for instance. And what's interesting is that when you sign up for Slack, you can be like a startup in Silicon Valley, you can be a school, a university. Like, profiles can be very different, but the thing is the onboarding experience is always the same. So why do we give the same onboarding experience to everybody given that we know who they are, what tools they use? So something we did is that we started personalizing the experience, meaning that if they use, say, Shopify and Recharge, we're gonna suggest to add those apps inside the product instead of giving them the the regular onboarding. If they haven't booked a call with our customer success manager or our sales team, if they are trialing, we want to put our sales and customer success team inside the product. Because of course you can send them emails, you can send them like chat messages that would pop up, but the best time to engage a conversation with a customer or a potential customer is inside the product when they have decided to invest time on your— on your product. So we built like a— there are like about like 20 combinations of different onboardings and we try to adapt it to the customer. So for instance, if they stop somewhere in the funnel, we're not gonna give them like the regular emails like, "Hey, like go add some users." We will know like where they were stuck and then like give them like a— send them like a specific email to bring them back that is contextual to where they are in the product. So again, what are the results here? After we did that, we went to 95% customer activation after they paid, meaning that people made the decision to buy and 95% of them start using the product and getting value. And so here are again some tips that you can apply to your business. First of all is, once you have this data, tailor the onboarding to your customers. You can use tools like Appcues. I saw that they were exhibiting in the main hall, um, that allow you to put some models or like specific buttons in your app depending on the data that you have from Datanyze or Clearbit. Second is that you want to put your team inside the app. You don't want to be sending emails like, hey, book a call with me. If they are inside the app, the conversion's gonna be higher. And the last one that, uh, to me is like extremely interesting is we now have the usage data of the application. So we know basically who's giving you a high net promoter score. We know who's like getting a lot of value out of the app. And we can propagate this data from product all the way back to marketing. So let's talk about how to build a stack like that and how it works behind the scenes. So first of all, what we want to do is to connect all the tools that we use to make sure that they talk to one another. A simple example is like we're gonna connect our application, our calendar, our payment app, which is Stripe, Datanyze that enriches data, and we're gonna send all this data to a tool called Segment. And what Segment does is that it repeats all the pieces of information it gets to all the tools. Simple example is the customer started paying. Senske made their first payment. So Segment is gonna repeat this piece of information to the CRM. And so in our case, the CRM is HubSpot. So when the account executive gets in HubSpot, they see that the deal has been closed and they see the revenue that has been generated. We use a tool called Hull to propagate the data from HubSpot and have a two-way sync between the two. And the value is that we get a unified view of the customer. So we can access the SunSky record and see like, when was the last time they booked a call with our sales team, when they paid, what are the events they attended, the full picture. Remember the example I was showing you earlier, which is sales changes the source of this deal. So say Sunski didn't come from a trade show, but from outbound, for instance, outbound emails. So what the stack is going to do is propagate this data all the way back to the database, and then we're gonna have dashboards built on top of this database that contains all the data points that our business has, And so this way we can have dashboards in the office that say, this month, this is how much money came from trade shows. There is nothing to update. Basically, as soon as the information changes, the dashboard updates, and we know how much money is coming from each of the channels. So let's talk about the benefits of doing that. One is when we go to the Sunsky record, thanks to Hull, we can see all the events that this company has performed. So we can see we met them at ShopTalk, they installed the competitor app, created a Gorgias account, started paying. And so anybody at the company who's going to talk to Sunsky, let's say they have a call with them, they have the full picture. And of course it's a much better experience for a customer to talk to a customer success manager that knows everything about you ahead of time so you don't have to repeat yourself. The second benefit that to me is extremely interesting is remember this graph we had from the marketing. So now because of the fact that we've connected everything together, we can look at the different cohorts. So let's look at the one from outbound. 50K came from outbound. The average deal size is $2,000. NPS is okay and churn is not that good. Like we, we have some customers leaving. So maybe this cohort looks good from the marketing side, but when do we look at how they use the product? It's not that great. The contrary is trade shows. So let's look at them. They spend $5,000 on average. So the deal is bigger. NPS is better. So which means they like the product better, and they're going to tell their friends about it. And we have negative churn, which means like if they spend like $1,000 today, in 1 year they are likely to spend more. So this is actually really interesting because now we know that we need to double down on this trade show channel because they love our product more than the customers that come from all the channels. So that's it. That's the story of how we got to 1,000 customers. And so here is the curve like 18 months after that. And what's interesting here is that I put in yellow, as you can see, the trade show cohort. And you can see that it expands over time. And basically because we know that these people like the product better, we can make decisions at the top of the funnel to invest more on these folks. I'd like to leave you with a few tips to implement that inside your company. One is the more you wait, the harder it's gonna get. Because as you grow, your company gets bigger, there is more complexity, and so if you want to unify everything together, you have more pieces to connect together, the problem is more complex. So start today. The second one is that you want to have all the company data in one place. What you want to avoid at all costs is having like marketing doing their thing on their side, sales doing their thing on their side, and they don't really talk to one another. Sometimes somebody pulls a spreadsheet, but there is no way for you to like see what's the connection between the two. So build this company-wide data hub. Third is that you want to empower your team to make decisions using data. And to me, that's the most important one. Because it's great to build all that, but if you cannot really— if it doesn't translate to the way you make decisions with your business, you're not going to go anywhere. Like one of the other founders, I think it was the CEO of Glassdoor yesterday, was saying that the danger between 1 million AR and 10 million AR is that you start making decisions based on the intuitions that you have, Sometimes they are good, sometimes they are bad. So what you want to do is minimize the bad intuitions you have by checking this data and having your team make decisions using it. So a few things on this point. First, you need to empower everybody in your company to use analytics tools. It's not very complicated. Anyone can pick it up within a week. And as soon as they have this reflex of going to the analytics tools to understand what's going on, then they're going to start making decisions using data. So it's super important that you start with this culture. And I think as founders, you need to be the person embodying this because otherwise nobody's going to do it. Second, so one of the learnings of Saastr is that the customer success person should be a single digit hire. What does it mean? It means that you should have among the first 9 people you hire, a customer success person. I think you should do the same thing for somebody who manages all this growth stack. They should be a single digit hire as well, because it's gonna help you get faster to 1,000 customers. Another thing is we all love spreadsheets. We use them all the time. It's super convenient to like get some data really fast with spreadsheets. But the downside is that I'm pretty sure to bet that if you did a product analysis like back in October of last year, And now you want to refresh the data and know what's going on, see the change. You can't really see it. You need to rebuild the whole spreadsheet. And so you don't really want to do that. So as much as possible, try to build dashboards learning SQL. All the people I've seen learning SQL, like, it took them, like, 4 or 5 days to get, like, a beginner level that's enough to start understanding things. And so this way, everybody on your team can have access to refresh data at any time. One more thing that's super interesting is that typically when you get after 1 million AR, you start structuring your team a little bit. So like each team is gonna have their own KPIs, like success is gonna have their KPIs, marketing is gonna have their KPIs, sales gonna have their KPIs. And so what you want to do is that instead of putting some Excel spreadsheets all the time and trying to like get the data for last month, but you're not really sure, what you want to do is build dashboards based on this data warehouse that automatically refresh with the latest information. The benefit of this is that everybody in the company knows what's going on live, and also it creates a sense of transparency because everybody needs to align on the reality of the data that is collected. So it aligns all the company in the same direction. So yeah, that's it. Hope it was helpful. I'd love to talk to you if you have questions on this topic. I'll stay a little bit after the, the, the, the talk. And yeah, here is my email address if you want to contact me. We're hiring a few people in all the— to build this growth data machine. So if you're interested, please reach out to me. Thank you.