Mark Roberge (Stage 2): How HubSpot Became PLG

Blake: Coming up on today's episode of BUILD.

Mark Roberge: Once it works, you disrupt yourself before someone else does, and that comes in a lot of different flavors. For us, it was easy because it was a different product, so just kind of brought them together and we had to train the core sales team on selling both and the support team and there was a process there. For other folks, I know when salespeople in certain territories quit, they don't replace them. They just let the PLG take over the territory. It depends on your business context, but you essentially have to find a way to disrupt yourself.

Blake: Anyone who knows me knows that I talk about PLG a lot. I'm constantly talking PLG in one- on- one settings with founders and product leaders and growth leaders, plus I'm talking PLG for wider audiences on podcasts and live talks. Regardless of the conversation, I always get the same question. Somebody says," I work at an established SaaS company that is not PLG today. How do I transition to PLG?" It's a topic that's on virtually every non- PLG software company's collective mind today. I get asked this question weekly, and guess what? I always have the same answer. I say," Yes, it's possible, but it's a long journey." That's because PLG isn't a go to market strategy on its own and it certainly isn't merely just adding a growth team to your existing equation. Rather, PLG is a whole company strategy that touches everything in your organization from the product you build to the customer journey you see to how you design the org to how you even think about OKRs and goal setting. Yeah, it's hard and it's slow to make this transition, but it's important because the PLG wave is building and you can't resist it. When I'm answering this question, I always give the example of HubSpot. HubSpot wrote the playbook on inbound marketing and had a ton of success with a sales and marketing led go to market strategy, but if you go check out the company today, you can sign up for all of the products, all of the hubs, on a self- service basis. They've effectively transitioned the entire company at HubSpot to PLG. And guess what? The company didn't blow up. Quite to the contrary, HubSpot is stronger than ever and has become a major rival to establish incumbents like Salesforce and many others. But enough about my views and theories. Today's guest is Mark Roberge, who was CRO at HubSpot for nine years. He built both the initial and original go to market playbook at HubSpot that was so successful and then was a part of the tiger team at HubSpot that disrupted that successful playbook by pioneering PLG at the company and driving this transformation. Today on BUILD we hear the full story directly from Mark. How did HubSpot actually transition to PLG and what are their lessons learned that you can apply to your own efforts to migrate and transition? This episode is a doozy and is a direct response to the most common PLG question I received. With that super long intro, let's dive right in with Mark Roberge and hear the HubSpot PLG story directly from the source. One of the questions that I get most often about PLG is how to transform an existing company from a sales led motion to a product led motion, especially if it's an established, somewhat mature company. The question is, is that even possible? How do you do it? And if it is possible, who's an example of somebody that has done it successfully? And whenever I get that question I always say yes, it is possible, but it isn't easy. It's a long process and the best example I can point to is HubSpot. So, would love to hear the story. When was this at HubSpot? What was going on at the company? Why was this an idea and then how did it start to unfold?

Mark Roberge: So passionate about it. I think it's so important for the ecosystem these days with PLG. I think PLG is like SaaS in terms of the disruptive potential. I don't think it's as applicable as SaaS was to the breadth of categories and software, but I think it's applicable the more categories than people think.

Blake: Yep.

Mark Roberge: I think if it is applicable to your category, it's going to be hard to fend off the PLG attacker, just like it was hard for the on premise incumbents to fend off the SaaS attackers. The transition is going to be tough too, just it was tough for the on premise incumbents to transition to SaaS. They went through a phase of like," Oh, SaaS isn't going to work. No CIO's going to put their stuff in the cloud," to," Oh gosh, this thing's kind of working," to," Oh my gosh, we're losing market share really fast. Let's hire McKenzie or BCG and try to figure this thing out." Some did okay and a lot didn't and I think that's what we're up against here. Yes, HubSpot did do it. We didn't do it because of loss of market share. We did it because it was probably about a year into the business. We're probably about 2007, 2008. There were probably 12 of us. A lot of us at the founding team were out of MIT and one person a lot of us were tight with was Drew Houston who went off and started Dropbox after MIT. Obviously, they were one of the few companies that did this motion. We didn't call it PLG then. We talked it we called it freemium and Dharmesh would watch it, talked to Drew and was pretty envious, honestly, of it. He's like," We have to do this. We have to do this because this is such a less friction way to distribute the software, especially in the S and B ecosystem that we're in." We probably had about two or three million in revenue at the time, thousands of customers, and at the time we were probably selling the thing for two, 300 bucks a month and we're like,"Okay, let's roll it out. Free product and$ 20 a month upgrade." Failed miserably and pretty obvious in hindsight why that was the case. I think it sheds important abstract lessons on where PLG is applicable and where it's not. The main result there was we didn't have a low time and effort to value use case. You need to have a low time and effort to value, and ideally retainable value, to be able to pull that off. For the HubSpot marketing software, the whole premise was like if you blog for four months, your leads will go up four times. Well, blogging for four months is not a low time and effort.

Blake: It takes a while to get to that aha moment if it takes four months.

Mark Roberge: Exactly. Dropbox is like," Yeah, click here and you back up your device in 30 seconds." Great. It's beautiful.

Blake: Yeah.

Mark Roberge: We didn't have it. The other thing that was against us, which is very relevant to your question, Blake, is we were trying to do it in the construct of having an existing install base, albeit only being a couple million box in revenue, but when you have... It's hard enough to you and I start a company, Blake, and we're like," Let's do PLG. Let's attack the marketing and software arena." Cool. We have a complete white space and we can architect what's the right use case that should be free and how do things get moved up to the... That's fine. That's white space. It's still hard, really hard, but when you add to that the complexity of putting enough in that free product where it works but doesn't cannibalize the$ 3 million that you have in revenue so that everyone doesn't downgrade from 300 bucks a month to 20 bucks a month and you have to explain it to your next investors why your revenue went down by 50%, that's next to impossible.

Blake: Plus, there's the fact that this customer journey takes a while to have the payoff and so the mix just wasn't right at the first attempt for PLG.

Mark Roberge: Yeah, wasn't the right use case and very hard. We didn't approach it in the right way even if it was the case. Okay, so then forward fast the clock to roughly 2012, 2013. Now we have 80 million in revenue, I don't know, 10, 000 customers, 1000 employees, I'm not sure. We decide to go into the CRM space, into the sales software space. We're exclusively a marketing software company and we decide to stay in SMB, go into the CRM space, feel like there's a lot of white space there. Now we can do it, and that's the first abstract lesson for people who are concerned that they're in a category that is PLG- able and they're going to be attacked. That is you have to create a safe sandbox for a small team to go and figure out PLG. It is going to take at least six months, probably more like 12, and that team, at least five or six people. Now, that could be a new market. It could be an international market or you could carve out a territory in the US. It could be a new product like was the case in our sector. Yeah, that's what we did, and we actually followed Clay Christensen's Innovator's Dilemma and Michael Tushman's Ambidextrous Organization if you want to read those frameworks. I find them to be great guides on how to do this. One of the Cliff Note versions of it is you have to set it up so that the big company doesn't slow down the project. And so what Brian and Darash said is," Okay, Mark, we're going to team you up with Christopher O'Donnell." Christopher was a very successful product leader in our organization, went off to be chief product office of the organization. He's like," The two of you. This is a critical project. We honestly feel like we might lose to Salesforce because they just bought Exact Target so they're clearly coming into our sector. We have to do this so we're going to put the two of you in charge of it and you're going to start your own company. We don't want to see you. We'll give you the bottom floor of the office. Do not build this on our tech stack." We run the entire platform of HubSpot marketing software on NetSuite. Build it on Stripe. Don't let anything slow it down. We don't have any overlap.

Blake: Truly no rules, no holds barred. You get carte blanche, startup within a startup. Go figure out what the right answer is for this problem for this market.

Mark Roberge: You have to do it that way. Yeah, because, Blake, we couldn't say," Oh, I think this is the right answer, and then does that benefit HubSpot?" Because the minute you say that, you're constricting your opportunity to perhaps be a little more guided by HubSpot's context as opposed to the true market opportunity. And so we had to be sincere, genuine, and true to the opportunity without being distracted, and also anything to be slowed down. We didn't want to have to get approval from the CFO to do something, or from the CPO on that side to do something. And so totally different tech stack. They were like," Here's budget for 20 people. You can poach 10 from the current team and you have to hire 10 from the outside." That was the compromise of using resources. That was it and that doesn't come without negativity. As we started to scale, people would be a customer of the marketing product and be using a CRM and they'd call up our support team with a question on the CRM and they'd be like," I don't know. I've never been trained on the CRM." It doesn't come without costs but those are necessary costs to figuring this out.

Blake: Just a quick break in today's conversation to make sure that you are getting all the latest in PLG content from OpenView. First things first, if you haven't subscribed to BUILD in your favorite podcast app, make sure you do that now. We drop four episodes per month and subscribing is the best way to stay in the loop. While you're at it, drop us a rating and review for the show so that others can find it as well. Secondly, did you know that I'm a YouTuber? I put out weekly videos on the latest and greatest in PLG with my show called the PLG 123. Every video is two minutes or less and features VC perspectives from yours truly on the latest in VC, SaaS, and, of course, product led growth. Find me on YouTube by searching Blake Bartlett and make sure to subscribe to my channel so that you don't miss a single video. Now, let's dive back into today's conversation. You want to go after sales and build something that eventually will become the sales hub within HubSpot, but you didn't necessarily, at least from my understanding, kind of build a better Salesforce on day one. What did you end up building and what was that progression to ultimately get to full CRM platform? I guess with this PLG lens, what was the starting point and how did you make self- service work within the CRM space?

Mark Roberge: Okay, so first off it was getting the right team, which a lot of people don't do well. A lot of people, whether you're starting from PLG from scratch or you're starting as a project like this, a lot of PLG or a growth team in charge of PLG in B2B reports to the marketing team. That's the first mistake because PLG and growth is all about rapid experimentation on the entire growth funnel. That's what the growth team owns is they own driving users into the funnel, getting them to activate, getting them become DAU, WAUs and MAUs, getting them to trip the monetization wire and getting them to retain and maybe drive virality if you're doing virality. That's what they own. The best growth teams just run rapid experiments in a data- driven environment to drive the conversion rate day after day after day. We're talking about multiple experiments a week, not like three a month. We're talking about rapid experimentation. If you just put that team in marketing, you only have marketers, you don't have any engineers. Engineers don't want to work in marketing and so when you put that team in marketing and only have marketers, you dramatically restrict the footprint of experiments you can do. And so that's the first thing is we have to make this a product driven team. It's a lot of engineers, a lot of designers. It's engineers, designers, data scientists. That's the driver here. We went out and Christopher inaudible, we call him C. Todd. I think you call him C. Todd, Blake. He was like," The best growth leader I know is Brian Balfour." And I'm like," All right. I remember Balfour and he was an impressive guy and he's a founder type." He's like," Yeah, anytime these larger companies come, they don't make it interesting enough to compare to me going off and doing my own thing." And I'm like," You have a lot of fans at the company, Dharmesh, me, C. Todd, David Cancel who was our head of product at the time and inaudible." And so we got it done, we moved his wife over to Boston and we spent three years doing it. That was a big thing, the right team, the right mindset around the stuff, and he taught us a ton of stuff about... He taught us a lot of the stuff, and then, of course, it was getting back to our point here in your question, Blake, is, well, what is the product? Because you can't start off with a CRM. That's not low time and effort to value. Adopting a CRM is putting a bunch of stuff in there, setting up your stages. We need something, ideally, more like Dropbox. I think it was C. Todd and his team created the Sidekick app, which was more about initially understanding when and who is opening up your sales emails and they brought everything through a Chrome extension into your email. That was a starting point. It was like, yeah, just download this thing and you'll know when people are opening their emails. There was competitive apps out there then but because we did this in a very growth oriented, rapid experimentation way and because we had the advantages of the HubSpot brand still, we were able to accumulate lots of users very quickly. That allowed us to leverage that opening use case to expand the footprint of the use cases into what we all know as a broader CRM. You'd be surprised. If you get someone in this particular case, you get someone to install a Chrome extension into their email and start tracking all their emails, guess what? The CRM is just populating behind the scenes. It was kind of this self-built CRM.

Blake: There was all this value to the person who was using it, which was seeing when somebody opens your email and being able to call them right at that moment and stuff that sales reps care about and SDRs care about. But then on the back end it's creating exhaust and you're capturing that exhaust and that's starting to build the foundations of what in the future can become the CRM.

Mark Roberge: Exactly. That's a very specific case of an abstract point here, which is some well thought out packaging design about what is that opening use case that's low time and effort to value? Single player mode is a big advantage. If it is single player mode, meaning you can get value by yourself, that's much easier and lots of value for the end user as opposed to the executive buyer. I've looked at other folks who are like," Hey, I've got this forecasting tool and I want to do PLG." The forecasting tool is valuable for the executive, but the person that actually has to do it is the salesperson and they don't care that much about the forecast.

Blake: Yeah, it's not solving their problem. It's solving their boss's problem. It's like you've built the wrong product for the wrong persona. There's a mismatch there.

Mark Roberge: Right. Exactly. Those are some of the abstract lessons we can pull away from the selection of that free use case. And then there's a whole nother set of decisions around, well, what should the pricing tier be based on? And that's tricky because I'm still working on... Maybe you have guidance, Blake, because you've been thinking about this space a lot. But I know if you can choose something that doesn't restrict more usage by the end user, it's good because it just gets more entrenched. If it doesn't restrict the network effect of this thing to spread across the organization and maybe even other organizations, that's really good. I like to kind of try to charge for things that the executives really value but doesn't restrict security or user roles or integration with the bigger tech stack. You know what I mean? These are things that the end user doesn't care about as much, but when I find out as a CIO or something that I can't do this stuff and I'm torn between turning off this app that all my people love or just paying up and locking it down, making it secure, most CIOs want to do that today. Because another pattern I've been seeing lately is if you are going to try to monetize the single user to drop them into the paid account as a free trial for the first 30 days and then downgrade them to the free version to try to get them addicted. We did a little like that at Sidekick. I don't remember exactly what the rules were. I think there was a limitation on how many emails you could actually track, but we gave them unlimited in the first month just to get them going and then dropped them down a tier later. Once we had that all going, all that set up, and that right packaging model, two things that Balfour did and then one as we thought about layering the sales team, but I do think that are abstract points and a lot of people mess up. The first part was what are the north star goals? I do find that most PLG teams and B2B growth teams move to monetization way too quickly. The better growth teams that I've looked at don't do that and Balfour certainly didn't say that. Usually it progresses from step one, just get enough users to just experiment with. I don't know what the right number is, maybe 100 a week. That's where paid marketing is really useful because you can just force a certain number of users in and at this point we don't really care about CAC. We don't care how much these users cost. Let's get some experiments. And then once you have that, okay, fine, we figured out how to get 100 users in a week. The next one is free user attention. We usually talk MAU, WAU, DAU, monthly active user, weekly active user, daily active user. What we did, I think what Balfour professes to is you just need to track your WAUs in weekly cohorts and just make sure that it levels off at some point. It doesn't go to zero, because if it just goes to zero, if you acquire 100 users and then eight weeks later they all go to zero, there's nothing there. It has to level off and whether it lands... Obviously, the higher the better, but if it levels off at 10% or 30% or 60%, we got something and that's the second north star metric. Then the third was, okay, cool. Now I'm proving that I can throw 100 users in this every week and I'll retain 30% of them and now I've got this beautiful thing growing. Now, prove to me that you can acquire those free users economically. How can we acquire those? That's where now we start thinking about content marketing and virality, and I'll tell you that we walked into one month having acquired users through paid acquisition and it was$ 100 per free user. The team ran 27 experiments that month on that metric. After the month, it was$14. I was just sick, just all the learning that came from that. Okay, so that was third. The third one is like, holy cow, now I can put 100 users into this thing for free users at a price that I know I can make a business at and they retain. Now it's time for modernization. Now it's like, okay, can we get the ACV to work such that the overall payback and LTV CAC to work? But all too often people are like," Oh cool, we have 100 users coming in. Now let's figure out the pricing tier." And the final one was how we integrated the sales team, which was kind of a little more on my area, because a lot of people... You've got this cool PLG motion and then you start adding sales people and those sales people have a$700, 000 quota and all of a sudden they start missing months and they're behind, and guess what they do? They ruin the PLG motion because they end up calling those free users before the magic has happened and they do what sales people do. They give a demo. They tailor the pitch. They lie. They don't set expectations right and now your churn goes up and now you're a sales led company, not a product led company and your CAC goes up. Everything's ruined. And so we had to do two things there. The first one we started with a very small team and I didn't put them on quotas. I had three reps and I was like," It's more about learning than it is quotas." We had each of them do a recording every day and we invited the whole team to listen to the recording of the discovery call to learn what's happening. This is more about learning. And then once we moved through the funnel, we had it going and it was time to accelerate B2B monetization that only they could do. Then, I comp them by paying them more for the expansion revenue than the first revenue, which is very unusual. Most comp plans are you just get paid more for the first foot in the door and then the expansion's the easy part. When you do that, you put the sales rep behavior and motivation contradictory to the buyer's ideal adoption process. The buyer's like," Oh yeah, I'm loving this thing. I'll definitely upgrade to the CRM and I'll try it for this team of five people." And the salesperson's like," No, no, you have to roll it out to all 500 right away." And they don't say this, but it's because that's how I get paid. But if you say you get paid 20% more for the expansion revenue, now they're aligned. And I had my rep being like,"Oh sweet, I got a 500 seed account on the line and I'm going to sign up three accounts this quarter and then get the other 497 next quarter." I was like," Sweet."

Blake: Yeah.

Mark Roberge: Because you're going to set beautiful expectations for those three accounts because that's not when you get paid.

Blake: Yep.

Mark Roberge: You get paid when the three accounts work really well and then they upgrade, so beautiful magic happens when you kind of rethink that rollout.

Blake: Yeah, it's a beautiful way to get it all to connect.

Mark Roberge: And the other tweak you can use instead or also that a lot of people don't do is you could have some of the compensation be based on the aha moment, right? Let's take this, the Sidekick, the HubSpot serum that we're talking about here. Clearly, when a user installed the Chrome extension and started tracking emails, that was a very sticky moment and so if you were running something like this, obviously we had people doing that humanlessly, but if you were running sales team who were getting involved there, you can pay them half when the customer signs the contract and half when they set it up and use it, whatever your aha moment is. I've done that with lots and lots of companies and what ends up happening is the salesperson gets the aha moment set up during the sales process as opposed to waiting till the purchase happens and that just tees up your CSM team. It sets really good expectations. It really lowers churn, just a lot of beautiful things happen there.

Blake: Taking a step all the way back, there's some things here that you've outlined in terms of the process, again, for a company that might want to look to do this. I think those process pieces are incredibly important and valuable, things like creating a sandbox. You gave a couple examples as to how you could create that sandbox. We're going to launch a new product, we're going to go into a new GO, something like that. And then once the sandbox is created, it's like give them freedom. The goal is not synergies on day one. The goal is figure out the right answer for this market, give the operating leeway, all those kinds of things. But then once you've done those things from an infrastructure standpoint, you've got the right team in place, the principles that are coming through to me is start small and be patient. The starting small piece is you didn't try to build a full featured CRM replacement right out of the gates. You said," What's the thing that can add value to an individual user?" And you started there. Let's get them hooked. Let's get them loving it. Let's get them telling their colleagues and friends and peers about this thing and let's get retention and value and aha moment and activation out of something that might just be more of a feature. It's not even a product yet. It just does one thing but then we can use that as the starting point to then build more. We started small. We didn't bite off more than we could chew. Then, on the be patient piece, there's so many elements of being patient that you pointed to, whether it's not focusing on CAC on day one because beautiful SaaS metrics isn't the goal. Product market fit and building the right thing that can take off is the goal. Being patient on when do you monetize and how early do you expect to see revenue results? When do you put sales in there. All of those things. Again, it's don't get too excited to early and screw up all the progress you just made because it's delicate. The wheels are still wobbling a little bit and so you need to give it the patience, give it the time to mature and then once it is working, then you can... Once you've nailed it, then you can scale it sort of thing.

Mark Roberge: That's a great summary. That's a great summary. I think that last part is a key point that we didn't really talk about. A lot of these practices, putting the team, product oriented, the way we measure, the way we read experiments, how you overlay, that's applicable whether you're trying to move to this or you're building it from scratch. But if we just talk about how you move to this, that sandbox, carving out that sandbox, separating out that team, setting up that team in the right way, and then your point, Blake, is once it works, you disrupt yourself before someone else does. That comes in a lot of different flavors. For us, it was easy because it was a different product, so just kind of brought them together and we had to train the core sales team on selling both and the support team and there was a process there. For other folks, I know when salespeople on certain territories quit, they don't replace them. They just let the PLG take over the territory. It depends on your business context, but you essentially have to find a way to disrupt yourself and that's why you might have to cannibalize your revenue a little bit. For growth startups, I typically like them to do that disruption process when they're capitalized because you might have a quarter or two of some wonky numbers and so whether that's just.... Obviously, this is a very transparent conversation with your investor, but it's just maybe a new investor coming in and be like,"Listen, this is what we're going to do. It might create some weird revenue for a quarter or two, but this is the business that comes out." Or your existing investors. You might have some wonky a quarter or two and that takes some guts, but it's better than taking your medicine in a year when you lose market share from someone else. This is the definition of leadership. It's not easy. You've got to get through it. Having done it with a couple of folks, get ready for this. Your VP of sales is going to come to you, being like," You are killing me."

Blake: On this point of leadership, there's a lot of founders listening right now, there's a lot of leaders listening right now, who are in this exact position. I want to go to PLG. I have the full conviction, but we have a going concern. We have a company that's been around for 10 years. It's going to be a process. I want to do it. This story is incredibly helpful. These frameworks are incredibly helpful, but it's going to be a journey of 1000 miles and I need to take the first step. What does that first step look like and how long should they expect for that first step to start showing some results?

Mark Roberge: Yeah, I love it when I'm... Actually, just in the last six months, I'm seeing this issue come up 30 out of 30 times. And when they listen and fix it's like," Oh my gosh." It helps a lot. Here's what happens is I'll talk about this and be like,"All right, is your category PLG- able? And we'll decide, oh my gosh, it is. It's PLG-able. We're already seeing some attackers get funded and we're at 30 million now and they're only at half a million, but we're really worried and we have to do something. And they're like," Okay, we're going to do it." And I'm like," Okay, cool. Carve out the sandbox, build the team." They got it. And then they're like," Okay, we're going to start the first experiment in three months." And I'm like," Wait, wait. Dude, what do you mean three months? What are you doing?" He's like," Oh, well, we have to build this whole self- service onboarding process." And I'm like," No, no. You're going to get it wrong." Here's what you're going to do. Fine, build some sort of MVP or use the existing product to hack it and just put up a landing page that says what your product does. Try now. Direct some paid traffic to it and then do the person behind the curtain Wizard of Oz technique. Have someone call them right away and just be like," Oh cool. Thanks for signing up to our product. Let me get you set up and go to work." Don't spend three months building the free user flow. Just hack it and have the engineers or PM sit with the person that's taking those calls and watch where people get hung off. Give the user the control to get things set up and tell them what to do and just now you start building up dozens and dozens of users next week. Use the Wizard of Oz person behind the curtain initially. Don't wait 90 days to build your free user flow and build it wrong.

Blake: No, this is great. Yeah, start small, be patient, but also orient towards action and don't over- complicate things. If you need to do the person behind the curtain strategy in order to get some initial signal, if you need to use paid marketing to juice signups, just to see if there's any value here at all, all the things we've talked about embody these three things. Start small, be patient, but also orient towards action and get things done and get them done week in, week out. Whether it's experimentation, it's about volume, whether it's about product features that you're shipping, it's about volume, and it's about speed. And so this is a perfect summary of both the story that I point to so often, and guess what? I'm going to point a lot of people to this podcast because this question's going to keep coming. And I'm like," Go check out this long version of it with Mark who is there on the front lines." And so I think the story is going to continue to go out there and is going to have a lot of value, but especially when it's combined with these principles of how do you do it operationally and then keeping in mind that you need to start small, be patient, orient towards action. And then once you disrupt yourself, it's going to be disruptive. Lean into it. I think this is perfect. It's going to be valuable to so many people. Thank you Mark so much for joining us here in the BUILD podcast and walking us through your thoughts here.

Mark Roberge: Yeah, it was so great to get so deep.