"Our best and worst marketing experiments" Lucidchart Co-Founder and CEO Karl Sun

My name is Carl Sun. I'm the CEO and one of the co-founders at Lucid Software. Our core product, Lucidchart, lets you express your ideas visually. The systems you work with, the processes you're working on, they're all getting more and more complex with more interrelated parts. When you lay something out and you can see it visually, it really helps you understand, and it helps you convey that information to others. People will come to realize, oh, that's the kind of problem that I can solve by mapping out in a visual way. Just like people would say, oh yeah, I can do that by putting this into a spreadsheet. I was going to introduce myself. I didn't realize I was going to get an introduction. So, uh, but I'm Carl. I'm super happy to be here talking about some of our marketing experiments. And the lessons that we've learned through this culture of experimentation that we fostered. And, you know, as I was putting this together, I realized that a lot of the experiments I'm going to talk about, mostly from the marketing side, have actually mirrored our evolution as a company. So hopefully what I'm saying today will be interesting and will be at least a little bit useful for you. But before I start, I guess I should introduce a little bit more about Lucidchart. So Lucidchart's a visual workspace that combines diagramming with data visualization and collaboration so that we can accelerate understanding and drive innovation. So our idea is we want to help you visualize anything so that you can understand it. Because once you understand it, we think you can figure out where to take action to improve it. So ultimately what we want to do is help you see more so that you can then know more and do more. A little bit more about us, I think Lloyd mentioned a little bit here, but, you know, we are based in Salt Lake City, Utah. We have over 500 employees now, over 15 million users spread across over 180 countries, and we're represented in 96% of the Fortune 500 companies and adding about 700,000 users every month. So this is a photo of actually a wall in our office where we've got painted on the walls some words that come from our core values. And two of our core values are actually innovation in everything that we do, and then also this idea of individual empowerment, initiative, and ownership. But really, for our employees to truly embody and live those values, we have to give them room to fail and to even encourage that, right? So we have to empower them to be innovative, even when that means sometimes, or oftentimes, things will go wrong. So as I mentioned, we have this culture of experimentation throughout the company. And today I'd like to talk about how our marketing department has really embodied this culture and has done some pretty amazing things, some things that, like, certainly Lloyd referred to as well. So there's sort of two parts to my talk. And I mentioned they sort of mirror our growth and evolution as a company. The first part I'm going to talk and things that we did early on were sort of pretty tactical experiments. They were around, you know, driving traffic to the site and increasing registration and conversion rates and, you know, the payment flow and things like that. And then the second part, we talk about more about branding and the vision as we start to tell more of the story about who we are and what Lucidchart is out to do as a company and as a product. So the first start— first part starts with our growth team that we started probably 5+ years ago. Years ago now. And they were created as an experiment in order to, and with a mandate to, run experiments. And the reason we created this as a dedicated team is because we understood that success would not be immediate. And instead it would come through lots of trial and error and iteration. And we thought that without a dedicated team, you know, everyone would be working on borrowed time. They'd be running experiments on the side. And it's hard to do that when you've got your real job and your real goals that you're trying to shoot towards. And it's especially hard to take risks and try things that may seem like bad ideas and will probably fail if you're doing it on borrowed time. And what we've actually learned ultimately is that sometimes the bad ideas turn out to be the best ones. So when we started the growth team, we had this goal that we'd Our target was if they can be successful, that would mean they would contribute like $1 million in incremental revenue, ARR, over the course of that first year. That was sort of what we were shooting for. It turns out they hit that within the first quarter. So we figured out pretty quickly that this was something that could be a huge win for us. We think about, as I mentioned, Experiments come in all sorts of different shapes and flavors. And so I'd like to talk about some of the experiments that the growth team ran in sort of 3 different categories. And the first is what worked that we didn't actually think would work. And so here is our original payment page. And early on, we really emphasized getting traffic, getting as many people to our website and then as many people to signup and register for the product as possible. So you can see that under that start free trial, we talk about no credit card required. We wanted to, as much as possible, reduce the friction that it would take to get people into the product. What we realized over time, the implication for this approach, however, was that while we certainly would get some people in and they would like it and they would pay us within that free trial period, the ones who didn't pay us would become free users. And we had a really hard time converting those free users ultimately to paid. So we'd sort of get people who would pay us initially, and then the long tail sort of didn't necessarily come very well, right? Like, people would stay free forever. And as we thought about this, we had this hypothesis that the reason was because we were trying to get those people to go from sort of a commitment level of zero, where they were free, all the way to a commitment level of 100% where they were paying us. And that was just a huge span and gap to bridge. And so we had this idea, "Well, what if we could get some of that commitment level early on?" Like, those of you who, you know, maybe in sales, like, are familiar with this approach. I don't come from that background. But this idea of you're not just giving things to customers or users, you're asking for something in return. So we wanted to raise their commitment level a little bit call it take them from 0 to 40 first. And then later on, if we were going to ask them to pay, then taking them from 40 to 100 would be an easier gap than trying to take them from here all the way. So to do this, there was this hypothesis that maybe we should require a credit card to start a trial. To be clear, there was obviously a trade-off here. When you ask for people to enter a credit card, a lot of people will get there and they say, "No, I'm not going to do that." I myself do that when I get to a lot of places, you know, there's that level of, well, am I going to forget to, you know, cancel or all that kind of stuff. So there's a definite trade-off and we realized that. We would lose people who had sort of low or no commitment. But at the same time, what we would get is we would get people who were sort of higher quality level. So to do that, this was the next iteration of the page and we ran this experiment where we required a credit card on trials, but we did it in stages. So first we only ran it with our Basic and Pro subscriptions, which are basically our lower-risk individual subscriptions. And then when we saw that had some success, we then shifted it over to the team subscription level as well. And we also originally started only on this pricing page in this flow. There are all sorts of different flows on our website where you can get into the product and, and pay us. But ultimately we ended up rolling it out across the entire site. And from there, we even continued to iterate on the exact wording of how we phrase things and even the number of days to offer as part of the free trial. There was like the 7-day trial or the 15-day trial or the 14-day trial. These changes ultimately on the pricing page ended up back in the day for us contributing about $500,000 of incremental revenue in just 12 months, which was a huge win for us back at the time. And overall, it increased payments by 25%, which is obviously, you know, great. And it came with a lot of original skepticism. Like, a lot of people thought that by requiring that credit card, we'd turn people off. I was actually one of those people. I was against this approach. I said, well, let's try it, but I don't like it. And— but by doing it, we realized that this was actually a huge win not only from a revenue perspective, but I think actually from a user experience perspective as well, because we could figure out which customers were actually really interested in our product, and then we could tailor our efforts to serve them a little bit better. Next category, I want to talk about some things that we expected to work but actually didn't work. And so I'll go back to sort of the pricing example here. And I mentioned that we had started the company with this idea of just getting people in. So getting as many people to the site and then getting as many people registering. And then after they registered, we'd work on getting them to convert over time. And so we had this big emphasis on free. So you'll see what's circled here, a start of free trial. And even where we talk about payment, the amount they would pay us is in really small font. We're trying to deemphasize that so people don't think they have to pay us a lot of money. And our thought obviously was that free would lower the barrier for people to enter. And it's true that it increased a lot of registrations, right? But what happened is we brought in a lot of interest, but because they were anchored on free, it really decreased the conversions. And so again, here's another shot of the pricing page that we ended up with. And as you can see, now we're emphasizing more this idea that Hey, this is a valuable product and we hope that you'll pay us for it, you'll want to pay us for it and you'll get value out of it. And so even the pricing is much more prominent. We're not trying to hide the pricing. We're being very upfront with you. And instead of saying start a free trial, we're now saying try or buy. And so that's actually a meaningful difference, like just de-emphasizing that idea of free. And so what we found is obviously we've We lost some traffic, lost some registrations, but the quality of the people that we're getting has more than made up for it. So this has been, you know, a really successful approach for us. And then finally, I wanted to talk about some experiments we ran, things that failed initially, but sort of with repeated effort, they ultimately ended up working. So this is a shot of our original homepage. And it was our homepage for the longest time. And we actually didn't really like it for all sorts of reasons. But every time we tried to change something on the page, things would break. Like the experiment we ran would fail. We'd, like— traffic would go down. Registrations would go down. So every time we made a change, we'd, like, hurry up and change it back to this. For example, like that orange navigation bar, we didn't really like that color. But we tried to change it to blue, and that failed. We changed it back. And you'll notice that on this page, the messaging is very tactical. And in fact, there's this term flowchart maker that occurs multiple times. When we first hired our director of content marketing, she almost threw up over this because she hated flowchart maker. She's like, who wrote this page? But it turns out flowchart maker is actually a really important word for us in terms of driving traffic from an SEO perspective. But this tactical messaging, the problem we ran into was as we continued to grow and evolve as a company and started to sell more into businesses and enterprise, this message was not resonating. This message was more about what you could do as an individual but really didn't help talk to companies about how we could help solve their problems. So in order to— as I mentioned, every time we ran the experiment, it failed. And so the team got pretty frustrated and dejected because even though we hated the homepage or really didn't like it, there was not a lot, it seemed, that we could do without sort of jeopardizing major parts of the business. So we decided to get the marketing team together, and we thought about how we could think about this not as piecemeal experiments on different parts of the page but more holistically. And to do that, we realized we had to do 3 things. So first, we had to de-risk this experiment. And we did that by moving the SEO value of this page to a separate page. Now remember, this was our home page, so we got a massive amount of traffic here. So we moved that SEO value so we wouldn't lose the traffic. It would still come in. And then once we did that, we could start to experiment and understand what different parts of the page contributed to the user experience and why people were looking at different parts of the page. And then finally, once we did that, we could then test and iterate on the messaging. So this is an example of our homepage today. And here are some things that we learned from that iteration and experimentation level. So first, we learned that above the fold, we still do need to focus on registrations. If we don't do that, we'll get a massive drop in registrations. But if we do that, then below the fold, we have a lot more freedom to experiment with our messaging and then to have a message that really resonates with more sort of our B2B crowd. So we can talk about what we're offering for the enterprise and what kinds of additional value we're providing. Back to that navigation bar I mentioned, we had changed it from orange to blue, that didn't work, but when we thought about the whole page in context, it turns out that changing it to white would work, maybe just because it was the contrast between the different parts of the page, right? So these are some things. And actually, that diagram that you see that's on the homepage above the fold, we experimented with all sorts of different kinds of diagrams given, you know, our— you know, we have a good clue on who our users are. It turns out that this org chart was the most effective. And not only was the org chart more effective, but we realized that when you put people's faces and pictures, it resonated more with people. And even furthermore, it worked best when these faces were sort of grayed out or desaturated as opposed to being like full color. So those are things that we never would have gotten a chance to figure out unless we thought about this whole experiment in a more holistic way. All right. So that was talking about— the first part was talking about some of the more tactical things that we did. And as I mentioned, I want to talk a little bit more on the creative side of what the marketing has has done. You know, we have a product that we believe can benefit everyone, but not a lot of people would consider themselves diagrammers. So our challenge is to get our product in front of those users who are looking for a diagramming application, but also get in front of those who don't even know yet what it is we do for them or why they need us. So our creative team first started experimenting with ways that they could appeal to a broader set of people. And out of a summer hackathon, there was this idea for what we call the Songtacular Flowdown. Where we would take popular songs and use it to display the ease of use and functionality of our product. And so here's an example of one of those. Hopefully we have it. <LYRICS>Bounce with it, girl. Dance with it, girl. Get with it, girl. Ba-da-bang-bang. Come on, come on. Turn the radio on. It's Friday night and I won't be long. Gotta do my hair, put my makeup on. It's Friday night and I won't be long till I hit the dance floor.</LYRICS> Ba-da-bang-bang. All right. So thanks. So that was one example, but we actually also did, you know, "Hey Jude" by the Beatles and "Beat It" from Michael Jackson and Taylor Swift, "Shake It Off" and things like that. So it inspired us to think about, hey, every summer we should kick off a creative campaign and try to do something big. So the next summer, we decided to take pop culture topics and try to turn them into flowcharts like this one. So we actually did a bunch of these and actually saw a lot of great traction. It got over 3 million views. And this one you see is Star Wars focused. And let's see, what else we got? We did one coinciding with the American football season about how your team could make sure that they have a winning season. We did one about leadership style based on the characters from Hamilton the musical. And we did one about would you get expelled from Hogwarts depending on your personality characteristics, things like that. So these were great, and like I said, they were fun and they got picked up a lot and people wrote about them, but we realized It actually didn't do a great job of introducing you to the product, even less so than the song ones, and telling you how to use our product or what it would be useful for. So the next summer, we decided we were going to go big. We wanted to go all in and try to create something viral. We were going to try to create the largest diagram ever created and go for like a Guinness World Record. So this was sort of our plan for how to do that.

Walter goes for a walk, and you decide what happens next. Maybe Walter strolls across Mars, and maybe he runs into Albert Einstein. So add your sentence to Walter's story and help set a world record. The world's largest crowdsource diagram. A giant choose-your-own-adventure story written by the internet. And you could even win $1,000. So click to add your sentence to Walter's story.

[FOREIGN LANGUAGE] All right, so we wanted to create the world's largest crowdsourced flowchart. Basically, we were turning over to the internet what adventures Walter could go on. What could possibly go wrong with this approach? I should mention that we weren't totally naive. We realized that Walter would probably get himself in some trouble, and so We had some automatic filters in place. We had several people on hand to manually curate and filter. But what we learned is that the internet is a nasty, dark place. So it did not take us long that we had to pull this. And that was really painful because of all the effort. I mean, the team had spent a lot of effort. We had to do some things in the product to make sure that you could have as many people collaborating as possible, and it could be published, and everyone could see it and contribute. So there was technical work that went into it. They had reached out to the Guinness Book of World Records to see if they would certify this, you know, but I think this was up for, like, maybe a day or two before it was pulled. And actually, out of morbid curiosity, I've asked several times what it was that Walter did that got him into trouble, and To this day, our CMO and everyone who worked on this will not tell me. The only thing they will assure me is that Walter is serving several consecutive life sentences for what he has done. But around this time, an engineer who was really into internet memes came to our creative team with a different idea. He wanted to use diagrams in our product to help explain different internet memes. In particular, all the different names for dogs or doggos. His idea was to use our product to create a video that could help explain this. So he went to a couple of people on the creative team, they sketched out ideas, they realized that they needed to do something low fidelity to stay true to this community, and then they spent a couple days pulling it together. And after they created the video, they stopped our CMO in the hallway, who was still recovering from Walter's episode. And said, "Hey, we have this thing, we want to try it out, what do you think? We think it could be fun." And this is sort of what they played for him.

This is a doggo. A small doggo is a pupper. Here's a sad pupper. But a big ol' pupper, that's actually a doggo. Not to be confused with a big ol' doggo, 'cause that's a woofer. But if it's just a small woofer, then it's actually a doggo. And do not pet a snip snap doggo. Sometimes a pupper is a real small pupper, and that's a yapper. But then it grows up to be a big ol' yapper, and that's a pupper. Now a big ol' woofer is no longer a woofer, it's a floofer. If it eats a humans, then it's a Griswold. Stay away. A Floofer won't eat a humans, and a small Floofer is just a Woofur. Duh. There are many, many doggos species, like Common Doggo, also known as Doge. Special doggos rare and very special. Other doggos could be a Wrinkler, or a Corgo, or a Shoog, or a Long Doggo, or a Puggo, or a Party Puggo. Doggos do things. A doggo in the water is an Aqua Doggo. A Woofur in the water is a Subwoofer. Lots of bork bork snarl makes for a heckin' angry Woofur. A doggo in a costume is Bamboozled. Sometimes they're doing me a frighten. This is Trash Boy. This is doggos making puppers. This is a very fast doggo running at incredible high speed. Diagram your doggos and anything else with Lucidchart.

All right. Thanks. So, you know, this was sort of a random idea and we decided to put it out there and we're like, okay, if it gets 50,000 views, that would be a huge success. Within a couple of days, it got 33 million views and we had not done a single thing to push it. So we were flabbergasted, and so of course we thought, wow, how can we recreate this? What can we do? So we continued down the animal vein and saw quite a bit of success, not quite to the doggo level, but really, really well. So we did birbs and sneks, and honestly, I'm not an internet meme guy, but anyways, they performed really well. And then, you know, we branched away beyond animals as well, and we did one on Fortnite. Which worked really well. It was really popular. And we did one on Star Wars, which actually didn't work. But we continued to iterate on this. And this campaign is called Lucidchart Explains the Internet. And it now has over 280 million views. So— and actually, it was recognized by Adweek last year as, quote, "the only ad campaign that truly mattered in 2018." So the team is really proud of that. And we know this campaign obviously isn't going to last forever. But we're continually iterating on it and looking for new ideas. But none of this would have happened if our CMO hadn't had an open mind and said, "Sure, why not? Go for it." All right. So that's basically a summary. And so what are the takeaways from here? I think for me and for us, there are 4. So the first is obviously encouraging trial and error. This idea that if you're experimenting, then You know that some, or maybe even most, things that you try are going to fail. But that's okay. And we like to say 100% of our experiments are successful. That doesn't mean they actually move the business in the right direction. But if they failed from a results perspective, we learned, and hopefully learned a little bit about why things were working and why things weren't working. The second one is don't blindly trust the data, or trust your gut for that matter. My gut was telling me that that credit card trial required thing we shouldn't do, but I was wrong about that. If we had trusted, just purely trusted the data on all of those homepage experiments, we would still be stuck with that really crappy homepage, as opposed to something that I think resonates more as we evolve as a company. Third, try to fail fast and fail often, but learn along the way. You know, with Walter, I think our problem there was, you know, maybe we should have anticipated that that was a bad idea, but at least we should have tried something a little bit earlier on. That one hurt because we put so much effort into that campaign and ultimately had to pull it within 2 days. And so that was a case where we didn't fail fast enough. And then finally, this one about good ideas can come from anyone. And here I think about James, our engineer, who is into the internet memes, and it was great to see him actually at the award ceremony when we got awarded one of these awards for the marketing campaign, and just seeing him along with the creative team and the rest of the marketing team accepting that award, because it really wouldn't have happened without him. And I think that's been true in our company, and as you build your companies, think about that idea that great ideas for experiments can come from anywhere. So— Thank you very much for your time.