Peldi and I talk about moving from California to Italy to lower his cost of living to start a company, juggling different delivery formats for software (and the associated payments challenges), giving amazing support, and bending over backwards to help customers. While he originally didn't want to grow the business beyond himself, Balsamiq is now a team of 23 people based all around the world.
Garrett Dimon: We’re here today with Peldi, the founder of Balsamiq, which is a wonderful wireframing tool. Peldi is based over in Europe, in Bologna, Italy. Still Bologna, right?
Peldi Guilizzoni: Yeah.
Garrett: Go ahead and give us a quick rundown of the history of Balsamiq, how you got started and how you got to where you’re at now, how old the company is, how big the team is, all that kind of stuff.
Peldi: Sure. First of all, thank you Garrett. It’s a pleasure talking to you. I’ve been a big fan of yours since the Buffer days… What was it called?
Peldi: Not Buffer, Sifter.
That’s when I moved back to Italy to start Balsamiq because it was cheaper to start it from Italy than to live in San Francisco. Balsamiq is a bootstrap. We never got any funding other than from customers.
Peldi: It rhymes. The Sifter days. My name is Giacomo Guilizzoni, but I go by Peldi because it’s easier. I live in Bologna, Italy. I worked as a programmer at Macromedia and then Adobe in San Francisco from 2002 to 2008.
That’s when I moved back to Italy to start Balsamiq because it was cheaper to start it from Italy than to live in San Francisco. Balsamiq is a bootstrap. We never got any funding other than from customers. My idea original idea was to start a one-man company and stay a one-man company but the product was too successful for that.
The market told me that I needed to grow. Reluctantly, we’ve been growing and growing. Now we’re eight years old — maybe eight and a half — and we are 25 people, if you can believe, for such a small product.
It’s a wire framing tool so it replicates the experience of sketching user interfaces on a whiteboard or pen and paper but it does it digitally.
Garrett: You originally started out with a desktop application, not the traditional SaaS model and then expanded into SaaS. Can you talk about the decision process that led to that? Is that a long-term shift where you’re moving away from desktop to more online or just constantly planning on juggling both?
Peldi: Sure. I started with a plug-in application. My goal was to create plug-ins. I wanted to create two products a year because each of the products was going to be too small to sustain my company. They’re still around our website, something that says that we are plug-in vendors.
I still haven’t purged all of that. We’re still on the first plugin after eight years.
I do like to build plugins. I like to make, to buy, to build metal ware, make difference systems talk to each other. The idea was to build a wire framing tool for a plastic confluence, which is this extensible wiki that a lot of people use and love.
I built that first. Even during my private data, so many people said, “That’s great but I want it as a desktop app.” I said, “No, this is 2008. Everything is going to the Web. What do you mean, you want a desktop app?” It just kept coming for two weeks.
I was, “No, no.” They said, “I want to work on it offline.” I said, “OK, that’s fair enough.” They said, “I want to save my files on my computer.” My god, that’s so old fashioned. Anyways, it kept coming. At some point, somebody said, “Listen, if you don’t do it, I’ll go somewhere else.”
I said, “All right, I’ll tell you what. I’ll build the offline version for people who buy the plugin so that when they’re offline, they can download their work, continue working, and upload.” Because of the technology I was using, the offline version took an afternoon to build.
The idea was that I was going to build a company just for myself, just by myself, a one-man shop. I wanted to be able to sleep at night. SaaS was completely out of the question.
That was one of the perks of using Flash when it used to be good. I did that and some people said, “Oh, great. I want that but I just want that. Can I buy that?” I said, “I didn’t even know how much to charge.” One of these guys said, “Well, I found this other competitor, this other guy, that’s $79. I think yours is a little bit better.”
I said, “All right, $79, it is.” That’s how I did the pricing for the desktop application. Now, it’s $89. It’s stuck for many years and now we raised it a little bit to adjust for inflation. I wanted to do plugins. Why? Because I didn’t want to deal with desktop, configurations, and especially, I didn’t want to deal with keeping servers running 24/7.
The idea was that I was going to build a company just for myself, just by myself, a one-man shop. I wanted to be able to sleep at night. SaaS was completely out of the question. Plugins gave me all the advantages of being… SaaS was someone else’s server.
That is the sweet spot because you got all the benefits of a single version, recurring revenue, but you don’t have to host anything. I started with plugin and then desktop. Desktop exploded. It was 90 percent of revenue at the beginning.
Gradually, years later our customers started asking more and more for SaaS. Then by that time the company had grown, we were three, four people. I had hired this guy who’s on our server side. I hired him originally to build a plugin for another system, but he’s a total DevOps guy.
He was like, “I’ll do the SaaS, don’t worry. I’ll help you keep the servers running.” I was like, “OK, let’s do it.” Our customers really ask for it. This was 2010, maybe. A while ago, a while after I launched. We did that and that took forever, but we did it, and we launched it. It’s being going well.
Right now, we have these three business models. We have the Desktop version, which is a one-time deal. Recently, about a year and-a-half we made it so that only a minor version of dates are free. If you buy for three, you get 3.X, but when it’s four you have to buy again.
We still haven’t released the 4, and the 3 is being around for a year and-a-half. It’s going to be every couple of years we’ll do a major update. Then we’ll try to give existing customers upgrade pricing. Basically, the revenue there is new customers and volume licenses, but no repeat customers. There’s no support fee there for Desktop.
The plugins we have both an annual licensing model, where you buy the plugin and it comes with one year of maintenance, and if you want to continue to get updates and support then you pay again. Next year, you pay half the amount every year after that. This is the same pricing that Atlassian does for Confluence.
We modeled ours after theirs so that it’s easy for them to buy. Atlassian now has a SaaS version of their own wiki, and so we have a subscription based version of the plugin as well. That’s monthly or yearly recurring revenue. Then we have our own SaaS and that’s monthly or yearly recurring.
The plugins for the last few years have been about 20 percent of our revenue, pretty stable growing a little bit but not too much. Then you’ll see the chart of desktop going down a little bit and SaaS just keeps going up and up.
That’s what I expected, what I wanted at the beginning. The world is getting more used to working in the Cloud they start to believe that the data is actually safer in someone else’s Cloud than on my laptop. That takes a while especially for us because most of our customers are businesses. It’s slower than the general public.
That’s what’s going on. That’s the break up. I really like having all three models, because some months maybe Desktop won’t do well, but the plugins do really well for some reason. Plugins is more of an enterprise kind of customer… even though it’s not an enterprise sale it’s still a self-service online.
It’s nice to have this it’s like a diversified portfolio with one product.
That’s a mistake that I made by going with all these different products, all these different variations of these products. We have one tool, but we have 40, 50 SKUs in our database.
Garrett: Does the support and managing all the billing and payments, does that get significantly morecomplicated or is that just set up to be a pretty subtle little difference or it doesn’t matter too much?
Peldi: That’s for sure more complicated. That’s a mistake that I made by going with all these different products, all these different variations of these products. We have one tool, but we have 40, 50 SKUs in our database. There’s all the volume licenses for Desktop, that’s half of them, but there is also a bunch. There’s all the other ones too.
That’s complexity. That’s really complex. What it means is that, for instance, we killed two plugins. We used to have a plugin for FogBugz and one for XWiki, and we killed them. They were bringing in some revenue, and not negligible, but not enough to justify the amount of complexity for sales support especially. We killed those just to simplify.
We built our own billing licensing server over time and we have a person dedicated to that full-time. It’s expensive but it’s also a competitive advantage. I still always see it as this is all the infrastructure that we were building for the first product. The second product, if we ever have to build another one, we can just reuse all this stuff that we’ve built.
I still always see it as this is all the infrastructure that we were building for the first product. The second product, if we ever have to build another one, we can just reuse all this stuff that we’ve built.
Garrett: Given the way the word grew organically and theproducts evolved and changed, and you started creating new ones, is there anything you could have or would have done differently to make that easier or have you pretty much just done the best you could under the circumstances?
Peldi: The timing was interesting. When I started, the cloud for knowledge workers, local drive, this online wikis that were just not for nerds. That was really the hot new thing. I thought that just like Confluence, there were going to be a bunch of other platforms for me to plug in to.
That was going to be, I thought, the primary way that people would want to use my tool. Instead, what happened is that people are totally comfortable with having 30 different SaaS products on their monthly bill using one tool for one thing.
Sure they integrate with each other like Slack and all the integrations, but it’s a loose integration of online services. I was gambling that people would buy the platforms, and there would be like three or four major winning platforms and then that they would want everything on that same bill.
That was a mistake on my side, plus it was my weakness because I really love building that code that makes integration plugins. For me, it was fun we went down that route fully. If I was starting now, I would just do a web version with Desktop clients that require a web subscription.
In fact, we’re moving that way. The iPad app that we’re working on is probably going to be free. The app is free, but you have to have a subscription to save the data in the Cloud. That’s what Microsoft is doing for Office for iOS. There’s big software vendors paving the way for us little guys.
I think that If you’re doing any authoring tool the future is complicated meaning that you have to have native clients for all desktops and mobile platforms, and you have to have a Web client for Web people who are not on their computer…
Garrett: I think that pendulum swings both ways a little bit even internally at Wildbit I see us adopting tools, and it kind of expands, expands, expands, and like, “Oh gosh, this sucks. How can we contract this and focus on a fewer tools. Not necessarily one, because that’s impossible, but really condense that.”
Peldi: That’s what Basecamp did last year, they sold everything else. At first they were adding stuff and then they focused. It’s a part of the evolution of a software company. It’s cool. I like that.
Garrett: Yeah, absolutely.
Peldi: In general, the answer to should you do SaaS or desktop? I think that If you’re doing any authoring tool the future is complicated meaning that you have to have native clients for all desktops and mobile platforms, and you have to have a Web client for Web people who are not on their computer but they still want to make some edits. The data has to be able to be saved locally and in the Cloud. That is expensive!
Garrett: That’s hard tobuild. You can’t build for every platform right when you launch. You have to kind of choose your battles and expand.
Peldi: As a consumer that’s what I want. I want the power, and the speed, and the keyboard shortcuts of desktop, but I want my data backed up in the Cloud automatically saved. Then if I’m away, or on my phone, I want to be able to make minor edits that way too. That’s what’s going to be required which sucks for the little guy.
Garrett: Yeah. It could still happen though. The tools are getting better for making those.
Peldi: That’s true.
Garrett: That’s pricey. What are the other things that… I love, especially reading the blog and Twitter is talking about sending out t-shirts and stuff and the horror of, “oh my gosh, we don’t measure the ROI, we just do it.” Is that attitude towards marketing, and taking care of customers, and showing them they’re appreciated?
Just doing those little things, is that something you ever second guess or you just always felt good about it, and what advice would you give to other people getting started, because I know I wanted to do a lot of that when I was independent, but it takes time.
You have to have somebody dedicated to be able to keep track of that and ship it all out. How does that scale, and how did you get started, and where do you see that going and how is it continuing?
Peldi: It’s actually not that much work. At the beginning, the swag step it was just me and I was sending a little bottle of balsamic vinegar to a few people, the handwritten letters. I did that for the first few months, and then it became impossible, because I wanted to do more, we had too many customers to thank and to appreciate.
In the end what we do now is pretty automated. What we do is we have an account on Spreadshirt which is the t-shirt maker, and we upload the designs — they have a little editor where you can design little products.
Then you have a store. We have two, one in Europe and one in the US so that our customers pay less shipping. Then what we do is we pre-buy these prepaid cards with Spreadshirt and we buy 40 at the time. We get codes, and then these codes, basically, they should be able to cover a t-shirt and shipping.
Whenever we want to send somebody a gift, we tell Val, “Hey, Val, can you send a shirt to this guy?” She has a completed email, she plucks in the code, marks it as used on the Wiki, and the guy, the receiver can just use it. That way, they choose their own model, their own size, their shipping address. We’re not involved after that.
As for measuring the ROI, I have a condition that’s extremely rare. I’m allergic to metrics. It’s a medical thing. The medical community hasn’t really put their finger on it, but to me, it’s a real thing.
It’s pretty easy for shirts and stuff like that. Also, on Spreadshirt, you can set how much profit you want to make off of each sale. We set zero so that it costs the cheapest possible for people to buy it. We don’t expect to make money off of the swag, we never made it for that.
I think it’s rare that people just buy for real from the store. It’s mostly set up for us to send gifts to people. As for measuring the ROI, I have a condition that’s extremely rare. I’m allergic to metrics. It’s a medical thing. The medical community hasn’t really put their finger on it but to me, it’s a real thing.
We track nothing. We have analytics hooked up on the website. I think it’s been three years since I looked at it for five seconds. We don’t track downloads. We don’t track anything. The only thing we track is revenue and profits.
Garrett: That, to me, I think I wish I was in that camp but it makes me nervous. It always made me nervous, not looking at things and not checking in on them but in hindsight, especially now, having sold Sifter, I look back and that’s one of the things I get way too worried about, that it wasn’t really doing anything meaningful.
Peldi: Listen. I think we’re blessed that we found product market fit right away, right?
Peldi: The product has been successful from day one. In fact, the problem has been, how do we support all these people? We have too many customers? Which is a fantastic problem for that. Instead, if more commonly you have the problem of, “This is not making enough money for me. I got to optimize,” then of course, look at the funnel. We don’t have a funnel.
Of course, try to squeeze every percentage at every layer. There’s also another way. Talk to your customers. See if you can make the product better so that they love it more and they tell their friends.
Garrett: Talking to customers is going to beat out analytics any day, right?
Garrett: Because analytics, all they can do is give you the facts, whereas talking to customers, you can understand the why. That’s the difference. It’s easy to open up analytics and look at numbers. It takes more time and effort to find customers, reach out to them, schedule time with them, thank them and do all those things.
Peldi: It only does if you’re introverted programmer, like we all are, of course. I don’t want to talk to people, I can get the same info from numbers. No, that’s not true.
Garrett: You absolutely can’t.
Peldi: In analytics, you just spend hours and see, whatever you want to see, you’re going to be able to see it. It’s completely a waste of time.
Garrett: I would almost go so far…
Peldi: A low-budget usability study where you go to someone and say, “Share your screen, try to log in in my ap. Try to sign up for my app.” In 20 minutes, watching someone, you’re going to have two pages of notes of things that you’re a hundred percent sure you must do today because your app sucks. Analytics is not going to tell you that.
Garrett: Analytics isn’t going to do that. I would almost go so far as to say that you’re better off not even setting up analytics from day one and instead, until you are growing so fast where you’re looking to optimize some of that other stuff…Even then, if you’re growing fast enough, you’re going to find other places…
Peldi: No, if you stop growing. I plan on optimizing when we stop growing, but it hasn’t happened yet. I’m sure we’re leaving a ton of money on the table, but it’s fine…
Garrett: Even then, if you stop growing, I think there’s plenty of opportunities to just go sit down with people. You’re never going to not find an opportunity to improve it just by watching somebody use your software…
Peldi: I know, for sure. The software does mature after a while. Right now, took us eight years, and the product is good. It’s pretty good. We have two people in tech support with 600,000 customers. It’s good. It’s solved.
We say, from the beginning, I chose to compete on usability and customer service. It’s funny because these two things are really the same thing. Customer service is part of usability. Usability, or user experience, means that every interaction that anyone has with our company has to be excellent.
Garrett: That’s a lot. I have a feeling we may have already touched on this but what would you pin as Balsamiq’s unique reasons for success? What’s made Balsamiq successful that other companies haven’t been able to replicate or couldn’t replicate?
Peldi: We say, from the beginning, I chose to compete on usability and customer service. It’s funny because these two things are really the same thing. Customer service is part of usability. Usability, or user experience, means that every interaction that anyone has with our company has to be excellent.
Our website has to be excellent, our support channels have to be very approachable and easy to get to. You get great support when you reach out to us. Our product, first and foremost, has to be really solid and really well-done and fun to use, which is very hard to do.
We basically look at all of this as a single focus. It’s all about trying to provide our customers the best experience possible and really caring for their success, even more than our own sometimes. In some cases, we have customers who lose data.
They work on a prototype for a while and they get printed out to PDF — you can do that, interactive PDF — but then they lose the source file, for some reason. It could be our fault. It could be their fault. They write to support.
What we do is, we say, “Send me the PDF, I’ll recreate the source for you and send it over,” which is crazy. We love it.
Garrett: It’s not crazy. That’s the thing. It’s not crazy.
Peldi: It’s not crazy at all.
Garrett: It’s just something that at first look, you’re like, “No, there’s no way to…that’s not sustainable.”
Peldi: You know what, it’s a chance for our tech support people to use the app for a couple of days intensely and they find a bunch of bugs. The customers are going to love us forever. It doesn’t matter. Everything is a good…
Garrett: That’s so true.
Peldi: Doing the right thing for our customers ends up being the right thing for us as well. For instance, we have these support contracts where on the EULA, we say, “The desktop app is not really supported.” This is because at the beginning, it was just me and I wanted to be a one-man shop.
I said, “No, I cannot sell a cheap desktop app because I’m going to have too many customers to support. I want to build a plug-in so that I have few customers that pay me a bunch of money so I can support them.”
The EULA said, “If you’re a plug-in customer, you get support and you pay for it once a year, etc. If you’re a desktop customer, you’ll only get forums to support so that you guys can support each other.” We’ve never done that. We answer every tweet, every Facebook message, every forum post, every phone call.
Why? Because if you have a problem, it means we have a problem with the app. It’s probably a bug that we got to fix. Why wouldn’t we want to try to help you?
Peldi: I don’t know. I think it’s really tough. The reason we are successful is maybe because we are pretty good of all the different steps. We try to do our best. We try to be excellent at every single thing that we do — we call it the golden puzzle.
Which is that whenever we get a mention of something that we did that’s not our core and someone tweets out like, “Balsamiq has the best terms of services of every SaaS I’ve ever seen,” that’s a golden puzzle piece for us. We really try to collect those for everything.
“Oh, the shopping cart experience was great.” Yes! “I spoke with this very knowledgeable and kind customer person.” Yes! In every single thing we do, we try to make it so that it’s super valuable to our customers.
You lose that ability to stay enthusiastic and invest the time to make those wonderful experiences. Instead, it’s like, “What can I do to get this done so I can get back to my core product?” But the thing is, every touch point with your company is the core product.
Everything is a core product. That’s right. The software is just a small piece of it. The documentation, so important. The tutorials, so important.
Peldi: Everything is a core product. That’s right. The software is just a small piece of it. The documentation, so important. The tutorials, so important. There’s a concept of the whole product. I think it was in Crossing the Chasm where they say, “You think you’re selling software but really, there are so many other things that are just as important. Without those, you wouldn’t buy.” If you think that something sucks, that happens, but I have a hack for that that I use on myself, which is, I treat every task as a learning opportunity.
Just framing it that way makes me go, “All right, let’s do this. I got to write a EULA? I’m going to Google and read a bunch. It’s a research project. How fun is it that you get to do research and you get paid for it?” It’s like being an academic but not really.
Garrett: It’s not a task to be completed. It’s an opportunity to have fun and do something speical.
Peldi: To learn something new, right?
Peldi: That’s actually why I wanted to be a solo founder, because I wanted to learn all of it. Everything, I wanted to know what are all the steps.
Garrett: Obviously, great customer support. Marketing but not so much analytics. What’s the ratio in a company of people doing support/marketing type things versus people actually creating and shipping products?
Peldi: I had to look it up because I never track anything. I went back and I did a little spreadsheet. If you do marketing and support — and then there’s people that do both. At the beginning, it was me doing that, so I account for both. It’s been about 50/50, and it’s been stable that way for years actually. Never intentionally.
I also did the ratio of developers to total employees, just the developers just because it made me curious. That’s been between 50 and 70 percent since 2010. A little bit developer heavy but…
Garrett: Starting to even out now?
Peldi: Yeah. We have two people in marketing, which we never had for a long time.
Garrett: We’re coming up in about 30 minutes, but there’s a couple of key questions that kind of dovetail nicely that I really want to make sure we get to. One is, what would you say, looking back, are the key, one or two, inflection points in the company and decisions that had to be made around those?
As a parallel to that, what was the most difficult thing that was either the hardest decision, or the hardest thing to get through emotionally or whatever it is, to get to where you are now?
Peldi: One we’ve talked about already that I went from being a plug-in vendor to a desktop vendor before I even shipped. They say, “No business plan ever survives the first impact with customers,” and I was no exception. That was big.
Then the initial explosive growth of the first few years was just completely unexpected. I had to hire people, which was not my plan. When I hired five, I said, “This is it. We’re not going to hire any more. I don’t know how to manage more than five people. This is my first time.”
We tried to stay stuck at five. I even wrote a blog post saying, “For the next two years, the plan is for nothing to change.”
Which is fantastic to think about now. We tried for a few months and failed because customers kept on coming, and we were struggling. That was a tough moment. Right at that time, we got a serious acquisition offer, very serious, very big. We were only two, three years old.
We had five, six people, and this was a big acquisition offer, and so we came really, really close to selling. I told a story about it in the Business of Software talk last year, but you had to be there to hear it. We edited it out. Emotionally, that was really tough.
Saying no to all that money was really hard, but glad we did. We weren’t ready. We’re still having too much fun, we still don’t know what we’re doing, we’re still learning too much, it’s still our baby. I’m glad that we didn’t sell. Those are the key things.
Lately, it’s been more stable. It’s different problems because now we’ve got to run a company of 25 people. Very different than a company of 10, or a company of 5, or when I was by myself. I’m learning how to do that.
Different challenges, and still a ton of learning, so I’m still having a good time. Revenue-wise, we’re growing but at a nice comfortable pace.
Garrett: Right on. I think that’s good for people to hear, when people face that decision to sell, and it’s incredibly tempting, and ultimately decide against it, and are happy with the decision, because I think that’s one that you can second guess for eternity.
Especially if you don’t have explosive growth and then you’re like, “Maybe I should have just gotten out at that point and taken a job.” I think so many people, too, are wired to struggle with working for somebody else at this point because they might be like, “We need to stop this shipping t-shirts to customers. That’s a waste of time and money,” and you’re like, “No, no.”
Peldi: For sure.
Garrett: That becomes a really difficult thing to accept when a company wants to run things differently for any reason.
Peldi: The thing about us, though, is it’s never been about growth. We’ve been blessed with more money than we need since the beginning, pretty much. We’re all paid well. We give a ton to charity. We’re fine. We don’t need to grow that much. It sounds very stressful to grow that much with all these customers to support.
We’re totally fine with our little niche. Sure, someone could come and try and kill us, someone who’s willing to put their work to support more people, but so far we’ve competed against venture-backed companies. After a couple of years, they go away because our niche is too small.
Garrett: They run out of money.
Peldi: Yeah. I get a VC call once a week, pretty much, and at the end of the call, we talk about the business, and the market, and I’m always saying, “There’s no way that we can grow 10 times in five years.”
I ask them, “Our niche is too small. We’re already dominating the niche. How do you imagine us doing 10X in five years?” and they said, “Well, we’ll pair you with these other startups, and build a suite, and then go after Microsoft and Google.” That’s when I say, “Good luck. Try somewhere else.”
Garrett: There’s no way.
Peldi: There’s no way. That’s not something I’m interested in. Go try and do it with someone else.
Garrett: Have fun.
Garrett: This has been great. I really appreciate it. I think there’s a lot of pearls of wisdom in here, so thanks for taking the time.
Peldi: Thank you, Garrett. My pleasure. Any time.