Back from the Grave: The Story of Javelin
We fucked up. We ran out of money. We laid everyone off.
With nothing left, we kept going.
Fighting our shame, burnout, depression, loneliness, doubts, and constant anxiety about paying for next month’s food and rent, we made some progress.
We did it alone, so it took us a few years — but today, we finally launched the first CORE part of Javelin. The platform is live with the first set of incredibly powerful tools to help entrepreneurs validate startup ideas.
I want to take this opportunity to share with you our story, mistakes, and lessons learned. Below is a teardown of all the products we built, and how we got to this point.
Raising $1.5M
Through our participation in Techstars (2013), we raised $1.5M. I knew we didn’t have Product/Market Fit yet, but I believed we would get there with the product direction we had when we raised money.
I was wrong.
Javelin Experiments (Old Javelin)
This is the product we built during TechStars and launched three weeks before our Demo Day. The vision was to be the “Trello for Lean Startup.” I wanted to help people to design experiments for their products.
A lot of enterprises wanted this product and we knew it. We closed six whales for our pilot — and had a pipeline with 21 others. The problem ended up being that, during the pilot, none of our customers really used it.
They didn’t use it despite the fact that I had four customer success reps training them and checking-in on them every day.
The problem was this: the buyer in the enterprise wanted this product but his team members (who actually used it) did not care. They were not early adopters. As a result, going enterprise-first in this market was a dead end.
We decided to pivot. Had the problem been the user experience, we would have stayed the course and iterated. The problem was the employees didn’t need what we were building.
Re-Launching & Scaling Lean Startup Machine
The Pilots Failed
After the pilots failed, we decided to focus on what we thought would be our cash cow: the existing workshop business. This business had already generated $1M in revenue for us, although it was not very profitable considering the time and effort needed to organize workshops.
We thought we could use tech to automate the organization process, improve the marketing outcomes, and enable this business to scale to hundreds of cities profitably.
The business brought in almost $500k in revenue that year.
Though the business brought in revenue, it wasn’t profitable and the tech we built was convoluted. I still believed that if we built the tech properly, the workshop business would have scaled profitably. The event pages were actually quite nice, but the backend used by our organizers (the main part we needed) was a mess.
We were running out of money around June 2014, right before we closed a $600k bridge round. We didn’t have the resources to fix what was wrong with this platform and QuickMVP was looking a lot more promising.
Customer Needs Pivot: QuickMVP
The lesson that we learned in our pilot for Experiments Software, and in so many other places, was two-fold:
- We needed to create a consumer-first product — enterprises were not early adopters.
- People didn’t want/need help designing experiments on a weekly or daily basis.
With QuickMVP we set out to validate that people wanted/needed help RUNNING experiments (not designing them).
QuickMVP helped entrepreneurs set up a landing page and driving traffic to it, quickly and easily, without giving the experiment design much thought.
And I’m not kidding about this, we validated QuickMVP’s landing page builder software with a landing page before writing any code.
Right out of the gate, QuickMVP was massively more successful than Experiments (Old Javelin) ever was.
QuickMVP was more successful both in revenue and usage.
~70 percent of people who signed-up launched a landing page.
~70 percent of those also created a Google Ad.
QuickMVP ended up not being sustainable, but it was clear that this direction was the better option than the direction we were going with Experiments.
The beginning of 2015 was when we knew QuickMVP wouldn’t work.
The main reason was because most people who used it didn’t get value from it. The average customer retention was three months, meaning every month we lost 1/3 of our user base. Most customers would throw up a landing page for their idea, place a Google Ad, and they wouldn’t get any signups.
QuickMVP was great at invalidating people’s ideas, but it didn’t help them pivot. We knew the missing piece was customer interviews. Doing customer interviews is the best way to figure out how to pivot.
The interviews and the pivot allows you to come up with a landing page concept that when tested will have a high chance of working.
Zoom-Out Pivot: New Javelin 1
Learning from listening to hundreds of QuickMVP customers.
Based on what we learned from talking to hundreds of QuickMVP customers and analyzing their cancellation reasons (one of the most important features we put into the app), we needed to do a Zoom-Out Pivot to offer more functionality.
Zoom-Out Pivots really suck compared to Zoom-In ones, because it means you need to build a lot more to reach Product/Market Fit.
By 2015, we were almost out of funding (including our $600k bridge round) and only had one engineer left on our payroll (who we ended up laying off two months later). At this point, I knew the only way we could keep going was if I learned to code myself.
Learning to code was an empowering experience.
I was shocked when it only took me two months to get the hang of it and become decent. I actually had a knack for it. I realized how much more control I could have over our future and how much money we could save not having to hire engineers.
Seven months later.
After seven months of developing the New Javelin, it was ready to beta test. I was getting pretty burnt out at this point — and then the worst thing of my entire life happened. My childhood best friend passed away from an accidental overdose.
We nearly crashed at this point. This was too much for me.
My Co-Founder and I took a two-month vacation to Thailand. I needed a break. The past two years caused us both a lot of emotional damage. Shame, guilt, burnout, loneliness, and depression. I cannot tell you how much this vacation helped me clear my mind and recover from my emotional pain.
Some of the happiest moments of my entire life happened next in the beginning of 2016. I caught up with Mark Suster (our lead investor) for the first time in a year and he was great to me.
Thank you for that meeting we had in January 2016, Mark. Can never express to you how it changed my life.
I caught up with other advisors of mine including Farb Nivi and Hiten Shah, who both supported me and made me feel like the path in front of me had endless opportunities. Thank you guys, I can never thank you enough for helping me break through my mental and emotional prison.
The beta test at this point was going so-so.
The product direction was right this time, but the UX and UI didn’t fit together. What we needed was to get hands-on with customers to work out the kinks in the front-end.
I moved to Boulder at this time to explore a pilot with a $B enterprise where the CIO was one of my friends. This almost came together into a great pilot, but the CIO ended up retiring a month after I started working on the deal (bad timing).
Against all odds, we still hung on…
After the opportunity in Boulder didn’t come together, we signed a six-figure deal with Microsoft to run our LSM workshops in their innovation centers around the world. This provided us with a great opportunity to get some money in the bank and some room to breathe again.
Needless to say, our product efforts took a back-seat and the focus shifted to this partnership until the summer of the next year.
Rebuilding Better & Faster: New Javelin 2
In the summer of 2017, Microsoft went through a major restructuring from the top-down, and our champion got laid off. It sucks because he was ready to double our contract size and his proposal and budget almost went through.
Back to coding — and Javelin.
This is when we went back to coding Javelin. At this point, my web development skills had leveled up an insane amount. So I rebuilt the app from scratch very quickly, which allowed us to iterate much faster.
We conducted hundreds of user tests, updating and tweaking the front-end for several months. The front-end was the majority of the work at this point.
In March of this year 2018, (right around my 30th birthday), we finished the core functionality and UI/UX. We then got ready to launch New Javelin 2 in three US cities: Boston, New York, and Washington DC. Jason Saltzman,
Founder of Alley and one of our advisors, hosted our launch parties at his space. Thanks Jason, your support meant a lot.
Now, we’re piloting New Javelin 2 in our own Pre-Accelerator Program in partnership with SOSV, the most active early stage investor in the world. We just kicked-off the program this week and will be working everyday with five high potential startups to get them to Product/Market Fit by using Javelin.
We’re not successful yet, but we can feel how close we are.
It’s been a long journey with many ups and downs (far more downs) but we persisted, survived, and made ourselves stronger. We move forward today as better entrepreneurs, with a renewed energy and better perspective.
Could We Have Done It Differently?
There’s five ways we could have potentially avoided running out of money and been able to get where we are a couple years earlier.
Let’s look at each one of these touchpoints:
1. Raised more money in the beginning?
Possible, given the momentum we had at the end of Techstars. But we weren’t experienced or skilled enough to pull it off, nor did we know we were going to need more than $1.5M. We should have raised at least $3M.
2. Spent less money?
NOT possible. We were scrappy as hell. Neither myself nor my Co-Founder Grace have ever taken salary since starting the company. Our CTO was paid 50 percent less than his previous job.
Everyone who worked on Lean Startup Machine’s operations was underpaid significantly. All of our developers were hired remotely, meaning for their experience and skills we got a 20–30 percent discount off of what they would have been paid if they lived in New York or San Francisco.
3. Spent more effectively?
Absolutely — and yet we were being careful. However, if I was a technical CEO (as I am now) it would have been a game changer. Instead of hiring a CTO, two extremely experienced developers, and a VP-level Director of UX, I could have executed on my vision.
We would have needed fewer, less experienced people and actually been more successful. Since becoming technical, my abilities are equal to five or more developers because I’m full stack. I know exactly what to build, and I’m a work horse.
4. Earned more profit?
Absolutely — we would have been profitable early on in the game. Again, if I was a technical CEO, the Lean Startup Machine platform would have scaled and actually been profitable.
The product was too convoluted, and not really understandable by the average company, partly because I was looking at it from the perspective of a marketer. I didn’t have the experience then — and was not the full-stack developer/marketer hybrid I am now.
5. Raised again before running out of money?
MAYBE — We raised a $600k bridge 12 months after the seed round. To raise a larger bridge or raise again after the bridge we should have to put a lot more enough effort into thinking about and planning to raise, starting the very day our seed closed.
Honestly, QuickMVP wasn’t really investable as a brand, and that was where our traction was. We should have kept the product under the Javelin brand. That would have made us look more organized from the outside as well.
Overall, I didn’t communicate well enough with our investors, including our lead. Communication is a broad, useful key — but I was too focused on trying to get things done. I was too inexperienced to know how to continually get our lead to buy into our vision, especially when we had to pivot.
Biggest Mistakes
To summarize the points from above, there are three core concepts that I attribute to our initial failure.
1. Being a Non-Technical CEO.
If I was technical when we first raised, we would have been able to spend our funding a lot more effectively and our previous products would have been more successful.
Technical ability would have helped me know how much money to raise and how to prioritize our efforts. I’m a technical CEO today, and this change has given me an immense amount of confidence and control over our operations and execution.
I have the educated vision now, where I can see how these skills will help us save millions in the future. I have the “know how” to able to quickly iterate on our products and deftly move them in the direction they need to go.
2. Doing Too Much At Once.
Building the Lean Startup Machine platform and QuickMVP at the same time was a mistake. I thought we could handle it. I thought going slower was a greater risk because we needed to show growth.
Hiring less people would have just ended up in getting less done. The problem is, I was doing so much I couldn’t communicate effectively with our investors. With multiple revenue streams and cost structures, managing our resources was ridiculously complex and difficult for our small team.
Now, I’m obsessed with being focused and keeping things simple. I know how to leverage the few activities needed to have the biggest impact. Everything needs to be organized and methodical ahead of time or we won’t do it.
3. Lacking Experience and Maturity.
If I was more experienced, I would have done everything differently and better (I know, we all say that — but it’s true).
If I had more maturity, I would have been able to overcome the difficult emotions and fears involved in building a high growth business under conditions of extreme uncertainty.
I would have been able to do the right things faster (but this is part of the learning process — each time).
I’m grateful for the hard lessons and pain, as long as I don’t repeat the same mistakes again.
My Ask.
I’ve spilled my guts. Now we need your help.
We just launched Javelin on ProductHunt.
Please click here to check out the product that took us so long and so many sacrifices to build.
Javelin is a platform that guides you through turning your ideas into successful new startups. This software is based on our experience teaching entrepreneurs for the past eight years.
*UPDATE: We just won 🏆#1 Product of the Week! We need just 200 more votes to beat Apple for #1 of the month. Thank you for support!
It’s been a difficult journey to figure out how to simplify the startup process into tools that anyone can use, but I think we’ve finally got it.
ProductHunt (founded by a 2x Lean Startup Machine alumnus) is a website where users vote on the best new products. By launching Javelin there, and getting votes, we can reach thousands of entrepreneurs and makers. We can have a big impact on the global startup community.
