In conversation with
The founding engineer of Troops.ai on starting a company, learning from past mistakes, and being a mentor
A hobby becomes a profession
Hi, I’m Greg Ratner, the founding engineer at Troops.ai. My road into engineering was initially nonexistent because I had no intention of going into computer science or technology. I come from a family of artists, so I grew up drawing and painting and expected that to be my path forward in life.
However, in the late 90s and early 2000s, I got into animation and was fascinated by the era of state-of-the-art flash games that were popular at the time. When I started doing that, I realized that I needed some engineering and technical knowledge to catch up with the people making the cooler, more advanced games, so I dabbled in computer science and took programming classes in high school.
I was always really excited about the possibility of typing things into my text editor and creating things that people could use. I noticed that doing this in tech still allowed me to express my creativity, and I ended up majoring in computer science at Worcester Polytechnic. College is where I started my first company, and it was my first foray into entrepreneurship. I loved building things from scratch and seeing how people would use what I built.
Unfortunately when I graduated in 2006, the dot-com bubble had recently burst and we were about to go into a recession. This was a little bit of a reality check as I started interviewing for jobs, and let me tell you, it was not easy.
The trials and tribulations of interviewing
When I graduated college, the job market was shaky at best. I think a lot of people graduating this year will understand this feeling of graduating into the unknown at a time where finding jobs is hard.
At this time, managers were looking for people with more experience and were wary of hiring junior engineers. I remember interviewing for months and months and exclusively getting rejections four to five times a week. I got stuck in a rut of waking up, going to an interview, going to class, checking my email for the rejection, and repeating the whole process the next day. It was extremely frustrating and demoralizing, to say the least.
By the last month, I was just going through the motions. In a weird twist, the combination of lots of interview practice as well as the expectation that nothing was going to happen made it so I had a lot of confidence and became good at interviewing. To my complete shock, in my last week of five interviews, I ended up with five offers!
So, for those looking for jobs right now, I just want to say to not give up and that there is hope. I know it’s even harder with the pandemic, but stick to it and learn from each interview opportunity to push yourself forward.
Vistaprint and beyond: the road to Troops
The role I accepted from those last interviews was as a junior engineer at Vistaprint, which at the time was a startup that was about to go public. While not a very sexy premise, printing business cards and such, the company still operated from a very entrepreneurial mindset.
From the tech side, they recruited smart, interesting people and focused on the culture before it was the norm. All of the company’s decisions were driven by data. Alongside company culture, one of our main “pillars” so to speak was experimentation. When it came to establishing industry standards, they got it right, and I stayed there for three years.
After Vistaprint, I moved to New York City and immediately took a job where I was in way over my head. It was a small e-commerce company selling perfume, and I joined as the sole engineer of an $8 million a year e-commerce operation. I was feeling good after coming out of Vistaprint, but I didn’t realize the extent to which I didn’t know how to do most things, and I had no one to ask for help. This was a pivotal sink-or-swim moment for me because I was either gonna level up and make things happen or it was going to be a total disaster.
This job was a place along my path that I learned that if I worked hard enough, I was resourceful, and I built my network, I was able to solve so many problems and reach far beyond what I was capable of. I only stayed at that company for about a year before I bounced around to a few different places, working at a digital agency with clients, building an election online platform from scratch for a nonprofit, even starting my own company.
Introspective learning from starting a company
The company was called Salon Pulse, and it was basically an OpenTable for hair salons. I paired up with some friends so we could build this platform, but the idea around this was mostly based on hubris. I didn’t know the industry that well, and I just wanted to apply a non-pattern to an industry with a large market that didn’t have one yet. In hindsight, it was a naive approach, and I think a lot of entrepreneurs will understand the feeling of being so excited about starting a business that they dive into something just because it seems like a good idea.
As I did market research, I had a rude awakening where I truly understood how much needs to go into building a company from scratch and dedicating all your time, energy, resources, and mental space to it.
However, I learned so much about what motivates me as an engineer, like what kind of work I want to do, the type of people I want to work with, and more. After Salon Pulse, I continued in leadership positions by managing a tech team at another digital agency. It was a great experience, but marketing was not where my heart was at and I wanted to get back to hardcore software and SaaS tools. It was at this point that Scott, one of the cofounders of Salon Pulse, and I reconnected, and he had an idea that would end up becoming Troops.
Troops in the beginning
Scott was a great sales guy and was someone that I knew I wanted to link up with because I had programming skills, but no experience in sales. Scott reached out to me with a new venture he was working on to get my thoughts. The vision was rooted in the fact that SaaS tools have some of the worst interfaces that people are forced to use for their functionality. The idea was to connect all of these systems with a messaging interface to communicate across systems so people didn’t have to learn anything new. You could ask the assistant about, say, a sales pipeline and the stats for the week, and it would respond to you with the right information. I loved this idea, and Troops evolved to be a central nervous system for backend tools. It created a workflow so if anything happened in any system, it would notify the right person so they could take the next step directly in Slack or Microsoft Teams. The core of being able to pull signals from the system and bring them to a messaging interface is still at the root of the company today.
Bringing past learnings to a new venture
There was one main lesson from my past that I brought to Troops. It’s something that was ingrained in me from my Vistaprint days, and it was to never invest resources in something that we haven’t validated with customers. With any project, build the simplest version you can and put it in front of customers to get initial thoughts and reactions. Does it resonate in a direction that we can go in that will result in product-market fit?
Collecting feedback and using it with a combination of intuition and market research is the best way to go. Ultimately, tech is just a tool, and it's easy to use a specific tool as a crutch. But if it's not the correct tool for the job, you’re always going to run into problems. Focus on what the customer needs and work backward from there.
Dealing with tech challenges
Our problems started when we started to scale. Early on we decided that we were going to hire senior-level people who had made mistakes in the past and knew how to avoid them when building software. Our DevOps pipeline and deployment processes did not exist. We used a couple of different tools to help with specific tasks, but nothing was automated. This worked well when we were a team of three or four, but the moment we had six or seven, velocity ground to a halt. Our deploys kept getting more and more backed up, and we would be deploying something with a whole new week of work stacked behind it. All the pressure would end up on one person doing the deploys, and this also led to an ownership crisis on the team. That’s when and why we brought in automation and feature flagging. The main tool we ended up using was LaunchDarkly, which was a game-changer for our velocity. We were an early customer of theirs, and it helped us build out our DevOps pipeline, develop our own checks and balances, and it helped us unlock our growth and scalability as a team.
Reliability and investing in new tools
We are fairly conservative and try to stay away from shiny objects, but we like to depend less on humans and adopt reliability tools where we can. When it comes to managing infrastructure and managing reliability, we love it because we’ve seen the dividends.
We’ve found that the less you have humans involved in managing your uptime and performance, the better the outcome. One of the choices that helped us along the way from a software perspective was using typesafe languages for deployments with Scala on the backend and TypeScript on the frontend. They allowed us to catch several issues before they got to production and rely on a compiler to catch other things. We’ve taken advantage of ecosystems, especially Amazon, to evolve and move towards managed solutions. We moved to RDS very quickly on Amazon, and as that product evolved it had multiple availability zones with automatic failover. Over time, Amazon made things more and more serverless and we were one of the first people on AWS to jump on that.
Troops in the future
Looking to the future for Troops, I’m excited for the new platform we launched in October because of the new tools we can connect to (Zendesk, Hubspot, and more) and the fact we can add more messaging interfaces. We are also moving towards a place where we can bring information to different teams from systems that they may not have access to. This allows for more cross-department, cross-system signals so people can take action. All these things allow us to connect more dots so teams can empathize more with each other and work together on solving problems.
Mentorship and paying it forward
I was inspired to become a mentor because having mentors in my life was a transformational experience, both in terms of how I handled my day-to-day and my mental health. I was a relatively early member of the New York startup scene, and I felt a gap of not having the right resources and feeling very alone. Then when I started Troops, everything was moving so fast and I knew I needed advice from those who had gone through this process before, look at things through a different lens, and give me advice based on data instead of my emotions at a given time. I connected to people through the First Round Fast Track mentoring program, and my mentor was amazing. He had been the CTO of many different companies and was able to give me an even-handed perspective on the challenges I was facing. Over time, I had people start reaching out to me with something brand new who I would have informal conversations with, a cup of coffee, all that stuff to pay it forward like those who helped me.
One of my friends got me into the Oculus Quest and it blew my mind. I approached it with a decent amount of skepticism and it’s been freakishly good. It’s like being in Ready Player One, and things are only going to get more advanced. For example, I’m not into golf or a golfer by any stretch of the imagination, but I’ve been playing minigolf on the Quest and have loved playing with friends and catching up. It’s so much nicer than Zoom because there’s no fatigue, and you get to chat and do something fun at the same time!