The Crucial Role of Customer Feedback when Building Developer Products

Michael Grinich (00:02):

Welcome to Crossing the Enterprise Chasm, a podcast about software startups and their journey moving upmarket to serving enterprise customers. I'm your host, Michael Grinich. I'm the founder of WorkOS, which is a platform that helps developers quickly ship common enterprise features like single sign-on. On this podcast, you'll hear directly from founders, product leaders, and early-stage operators who have navigated building great products for enterprise customers. In every episode, you'll find strategies, tactics, and real-world advice for ways to make your app enterprise-ready and take your business to the next level.

Today, I'm joined by Edith Harbaugh, co-founder of LaunchDarkly. LaunchDarkly was founded in 2014 and is a feature flag management platform beloved by developers. Over nine years, the company has grown to 500 employees and over 4,000 customers, including large enterprises such as NBC, Atlassian, and Intuit. We're going to explore how LaunchDarkly started selling to the enterprise and more. Edith, welcome to the podcast.

‍

Edith Harbaugh (01:03):

Thanks for having me.

‍

Michael Grinich (01:04):

So give us a quick update on where LaunchDarkly is right now. Where's the business and team today and what's the company's current focus?

‍

Edith Harbaugh (01:11):

Yeah, so LaunchDarkly was started in 2014 in a very different era. I realize now it's 2023. When we started the company, I thought that feature flagging was a known technology, and it turned out that we basically had to evangelize the field of, "There's a better way to build software where you can do soft launches, controlled rollouts, controlled fallbacks, which is the launch aspect, and then also long-term control of features." So enabling a customer to turn features on and off for different regions or different of their own customers if they're SaaS. And we've also been investing heavily lately in more experimentation, which was always part of our vision in the early days, but the idea that if you have a feature and you've launched it, that you could also see how it's actually doing in the real world, which to any product manager is Nirvana, being able to launch, measure and control a feature long-term. So that's what we've been up to and we are putting a lot of focus right now in both the core feature management platform as well as more robust experimentation features.

‍

Michael Grinich (02:21):

I know a lot of developers when they think about LaunchDarkly, they signed up through that bottom-up motion. You can just go to LaunchDarkly, sign up, get an API key, and integrate it with your app. Can you tell me about how that evolved as you started commercializing the business and just sort of how the business grew from there?

‍

Edith Harbaugh (02:36):

So I'll say that's a common myth about LaunchDarkly. We have always had a bottoms-up motion. You can just start a free trial at any time. What we discovered is that developers would go and try something out and push and prod at it because developers like to try something, but to get sales, we needed to appeal to a manager or director, somebody who could actually force or entice an entire team to use it. So the majority of our sales are actually sales assisted, but we allow anybody to try it anytime.

‍

Michael Grinich (03:15):

How does that process work? Walk us through the timeline for maybe one of those developers in poking and prodding and then how your team knows how to reach out and what they do.

‍

Edith Harbaugh (03:23):

So we don't. So sorry. The ugly truth is that we'll find this out later. Anecdotally, developers are crafty and wily and they'll use a Gmail address, they'll use a throwaway account, and then when actual leads come in sometimes months or years later through an official channel and they're like, "Yes, we tried it out, but now we're actually ready to buy with our real account." We, of course, do all the obvious stuff. We have cadences of emails, we'll try to match emails with companies, but it's not as clean as you might think.

‍

Michael Grinich (04:01):

It seems like that could be a pretty frustrating way to build a sales organization. Pretty loss-y and hard to measure. How have you reasoned about that or maybe talk about the early sales team that you put together, how you navigated that in those first few days?

‍

Edith Harbaugh (04:15):

Yeah, so I was the first salesperson, so I had been a product manager and I managed salespeople before. We had this vision like everybody else in 2014, that we would be this bottoms-up, touchless, self-serve motion. And we built this incredible, for the time, system that would do that without any human interaction. The really frustrating part was people kept wanting to talk to us and we would push them towards self-serve and say you could just sign up and they would push back and say, "No, we actually want you to come visit our office. We want a demo." And what I finally realized was that we were a new technology and people wanted to build conviction that there was an actual company and people behind it. So literally I was really good friends with Paul Biggar, who was the founder of CircleCI. I actually did a podcast with him at the time and I still had to go to his office and talk to the engineering team, which made no sense to me.

I was like, "They could just sign up." And the rationale was, "Well, they wanted to actually see you in person." So I'd say we were kind of dragged into enterprise sales and it just compounded where I realized that, and I'll give my co-founder John all the credit for this too if one of us would show up at somebody's office and give a demo or just have that personal interaction, people were so much more likely to buy. Even to this day, I remember I went in demoed to AppDirect, which was I think customer number five or six for... I think at the time it was like $199 a month, something very low. And they still remember that. They still remember that, "Hey, the founder came and gave us this demo and that was part of why we buy it." So it's not the Silicon Valley myth that people want to hear about this perfect perpetual machine with no salespeople, but it worked for us.

‍

Michael Grinich (06:11):

How did you learn how to do those sales? Can you talk through maybe what that demo looked like in the office? I'm sure a lot of people listening to this are kind of in that transition point where they're trying to figure out how to do initial sales with their first few customers. What did you do?

‍

Edith Harbaugh (06:25):

I'd been an enterprise product manager and I'd gotten on a lot of sales calls. I'd also managed salespeople. So there's no secret sauce. The biggest mistake that technical founders make is think that the demo is all about the product. The demo is actually all about the customer and how you can help them. So the first 10 minutes of any effective demo in my mind we're just asking the customer, "Hey, what do you actually want to see? What do you care about? Why are you investing some time and meeting with us?" And then you tailor your demo for what they want to see. So when John, the co-founder and I were doing our early demos, it was never the same demo every time. It was always like, "Oh, they care a little bit more about rules. Let's show off rules," or, "They care about regions, let's show off regions."

And you always pause in a demo. If you're talking for 20 minutes in a demo, you're doing it wrong. You're not Steve Jobs up on stage doing the keynote. You're in somebody's office and they're giving you some time and you really want to make sure that like, "Hey, this thing that I'm building doesn't meet what you want." And the other advantage you have as an early-stage founder, which I think is underrated, is you could also ask for feedback. And in a lot of our early demos we would say, "Hey, just like this," and people would ask a question, they're like, "Well, does it actually have a user view too?"

And at the time, we did not have a user view. And that's an opportunity where you could pause and say, "Hey, you said you wanted a user view. Let's talk about that. What would you like to see.” And that could spin a disadvantage, which is we did not have a user view. Or an advantage of you're going to help us build this product and your feedback is really important to us. Does that make sense? I'm doing the classic demo thing where I'm pausing now to say did talk for too long and how are you following me?

‍

Michael Grinich (08:09):

It almost sounds like what you're describing is what a lot of product managers do in companies where they go out and do discovery on what to build and ask for feedback from those potential customers.

‍

Edith Harbaugh (08:18):

I always looked at a sales call as a discovery. "Does what we have match what you want? And can I kind of rotate what we have and what you want enough that there's some sort of click?" And a lot of times our early customers loved that. I just had a call with a customer who's been with us since 2016 and I look back and they took such a huge leap of faith and they're just so happy. They're like, "Every time we've really wanted something, you've been a co-partner with us.” And that's just as magical.

‍

Michael Grinich (08:50):

I wanted to ask about pricing because I'm sure during this time you were trying to figure out how to charge for the product and where the value was and just how to navigate that. What was LaunchDarkly early pricing and how did you evolve it?

‍

Edith Harbaugh (09:01):

So I was extremely, extremely shy in the early days. I still remember the legendary story is... We went to see a big potential customer and I was not used to asking for money. I tried to get my nerve together to ask for something. And I asked them for $20 to be part of our beta, not $20,000, $20. And my voice cracked in the middle. I was like, "Can you pay us 20?" And at the time it was my co-founder and I. We didn't have accounting or invoicing systems, we had to go back and make them an invoice. So I think the number one skill as an early-stage founder is just getting comfortable asking for money and you just got to practice that.

Now at LaunchDarkly, we track deals over six figures. We have hundreds of deals over six figures, and it's just practice, practice, practice and having conviction that your product is worth what you think it is. So our early pricing in hindsight was very low. I remember we had customers who were paying us $99, $199 a month. And what happened was, when you start getting more of a product that works, you have more ability to charge more for it and you could also justify it. But it's definitely a process. The salespeople now look back at my early deals, they say, "Oh my gosh, why is that person only paying that much?" I'm like, "Sorry. You do what you do in the early days to get it going."

‍

Michael Grinich (10:34):

I'm curious how the product evolved as you kept doing these sales and kept getting pulled up market to larger customers. How did you need to evolve the service you actually provided customers?

‍

Edith Harbaugh (10:45):

That was interesting. So we originally thought we would be feature flagging. We thought we would be cheerful feature flagging for the masses, take this technology that Facebook, Netflix were using, and democratize it for all developers. If you're a really small company, you don't need a feature flagging solution, you just build it in-house. If you have one or two features, you just hardcoded a database somewhere. What we discovered was that if you're a bigger company, if you're a huge company with multiple teams, that's when you actually don't need a feature flagging, you need feature flagging management. And that was something we did pretty early is, I came up with the word feature management because when we said we were feature flagging, people thought we sold goods to crossing guards.

But when I said we're feature managing, suddenly it became something where it's like, "Oh, feature management." And that was again a click for customers of if you have distributed teams, if you have role-based access, if you have audit logs, if you have SOCs worries about compliance, you actually do want a solution. That's not something that you want somebody working on an hour or two on a weekend if it's in your critical path for deployment. That was a lot of the evolvement of the product was having the core capabilities of really solid features, but also having all the management around it to allow an enterprise to use it at scale.

‍

Michael Grinich (12:08):

You mentioned some of these things around like trust of the platform or compliance. And LaunchDarkly is in the critical path there. Can you talk about how you built that trust with customers? It's a pretty big deal.

‍

Edith Harbaugh (12:19):

Yeah, that's a lot of why, though we have a self-serve path, any sizable customer does want some sort of interaction with sales just to make sure that there's a real company behind it. We take a lot of pride in being stable, being reliable, and just reinforcing that over and over with our customers from basics like we got SOC compliance very early, we have a status page where you can see your uptime and our downtime, and literally, and this sounds very cliche, but just literally telling our customers like you can rely on us.

‍

Michael Grinich (12:58):

I want to go back to the sales team. When you were talking about that early sales process of you going in, founder-led sales, how did you transition that to others? And I'm curious if you can talk about the early sales team you built out.

‍

Edith Harbaugh (13:10):

For a long time I had this half hope that we would get to this point where it would be self-serve, that me going out and selling was just that people needed to build trust in the company. The forcing function for getting our first salesperson was that I became the bottleneck. So at the time, we're an eight-person company and I became the bottleneck that people would ask for quotes, people would ask for demos. I also worked with an external lawyer, did the red lines. There's actually a lot of work in sales. And I think there's this cliche that salespeople just play golf all day. No, absolutely not. There's actually a ton of work in it. And so I realized that I was the one blocking sales just because I was also at the time the CEO of an eight-person company and anybody who's been the CEO of an eight-person company knows how much other work you're doing.

I was raising money. I was writing blog posts, I was giving talks. So I hired our first salesperson and that was actually incredibly difficult. The issue is that everybody thinks their startup is special and wonderful, and of course it is, but to take the job as a first salesperson is an incredible leap of faith. If you're an Oracle salesperson with a $1.6 million quota, why on earth would you go to a small startup? So we found this great guy who was out of New Relic who thought it would be exciting to work at a small company and that was fun. He took, as I said, this huge leap of faith on us. And that was the first knowledge transfer that I did of, "Here's how you should approach this field." And what I realized, in hindsight, was that our salespeople were the experts in our field. They were the trusted advisor.

They thought that they were far less technical than their buyers, which of course is true. They were not hands-on keyboard coding salespeople. They were salespeople, but they knew a lot more about feature management than our customers sometimes. And if they would help educate the customers, the customers were very grateful. I still remember I went on a sales trip to New Zealand with one of our salespeople and the customer had gifts for the salesperson, which in my mind was just unheard of to show up on site and the customer's like, "Thank you so much. I'm so glad that we are using LaunchDarkly. Here I got you a local New Zealand gift to repay you for all you've done."

‍

Michael Grinich (15:42):

I don't know if I've ever heard about that before. That's pretty incredible. That's pretty unique. How did that scale? Tell me about when you went from one salesperson to many, what had to change? What did you evolve as a team?

‍

Edith Harbaugh (15:52):

So we went from me selling, and I was in hindsight, the best Business Development Representative (“BDR”) and Account Executive (“AE”) because I cared so much. But that just did not scale at all. We hired our first salesperson, great guy, and then after we got our series A, we hired our first VP of sales. And enough time has passed, it'll just tell this story. He was this great guy who never really sold to engineers, but he said, "I want to try selling to engineers." I'm like, "Wonderful." And he lasted about three months because after three months he came to me and he said, "Edith, I hate selling to engineers. They're persnickety. They don't really like PowerPoints. And they ask hard questions." I'm like, "But yeah, I'm an engineer. This is who we are." And so he quit. But the two salespeople that he had hired stayed. So at this point, actually he'd hired three salespeople.

And the funny thing is, so I was acting VP again for another six months kind of unwillingly because I had thought, this is great, I hire this VP, I get to not worry about sales anymore. So I was acting VP and the weird thing is it was super fun and those three salespeople are still with the company, but that was again, this education in that there's actually a ton of work. So then we hired our second VP Sales who was wonderful. And the thing about doing a job for a while is that once you hire somebody to do it, you, (a) Realize how hard it is and (b) let them run with it. So when we hired our second VP, he's like, “Great, I'm going to start to hire managers. I'm going to scale. I'm going to do this.” I'm like, “Wonderful, tell me how I can help.” But there was no question in my mind about why we needed a sales ops person or why we needed managers because I had seen firsthand the nitty-gritty of running it myself.

‍

Michael Grinich (17:44):

What advice would you give early-stage entrepreneurs, kind of the next generation of founders building companies right now, having seen LaunchDarkly's growth and them wanting to have similar type of growth and success for their businesses?

‍

Edith Harbaugh (17:57):

Our success was not overnight. So we started in 2014 and we had this period in the early days where we got one customer a month. And we got one customer a month because I sent out an investor update every month with our dashboard, and I was determined to not let a month go by without adding one customer. So literally, I would cajole, beg, convince people to try, and if somebody came in through a lead form, I would write them back and say, "Do you want to get on a call? Do you want to hear what we're doing? Do you want to visit from us?" And that period was very, in hindsight, dark because we weren't paying ourselves much at all. We were getting one customer a month. It was hard. But the thing that kept us going was that the customer we got was usually pretty happy.

So what I told the team back then is every customer we get right now will be ten two years from now. Because happy customers get more happy customers, developers switch jobs. If they like something, they'll take it with them. And that did end up working out for us. Like now we have customers, as I said, because we've started so long ago where they've used us at multiple jobs and that feels really good, but the period of growth takes a long time. You asked before about do you have this sleek system where you automatically know if somebody who tries your SDK from a hidden Gmail account is going to turn into lead? It's like, "No, you don't." It would be really nice if it worked like that, but you don't. You're just taking this assumption that if you put a good product out into market that people will use it.

‍

Michael Grinich (19:47):

I think that's really sage advice. Edith, that's all the time we have for today. Thanks so much for joining us and sharing more about LaunchDarkly.

‍

Edith Harbaugh (19:54):

Thanks.

‍

Michael Grinich (20:00):

You just listened to Crossing the Enterprise Chasm, a podcast about software startups and their journey moving upmarket to serving enterprise customers. Want to learn more about becoming enterprise ready? The WorkOS blog is full of tons of articles and guides outlining best practices for adding features like single sign-on, SCIM provisioning, and more to your app. Also, make sure to subscribe to this podcast so you're first to hear about new episodes with more founders and product leads of fast-growing startups. I'm Michael Grinich, founder of WorkOS. Thanks so much for listening and see you next time.

‍