This the second post in a retrospective series documenting CarmaLink’s history from 2008-2021. Content written and provided by Kevin Holmes and Chris Suozzo.
2010 - 2013
Moving into 2010 we now had a number of products, concepts and potential clients with various markets we could move into. There were people approaching us regarding custom ECU tuning projects, monitoring for new/learning drivers, commercial IoT opportunities at local and world-scale, and even an IoT enabled snowboard.
A major focus for the executive team at this time was trying to suss out the best business outcomes from all the opportunities being presented to us. We wanted our investors to see their returns asap so we could continue to build trust and grow our team to take on new challenges. It became clear after a few projects that the most promising were those with a business-to-business approach. We wanted to focus squarely on innovating our IoT hardware products and the systems ingesting their sensor data.
After making a number of consumer-focused products like Doctor Auto and the Sword IoT prototype (Bluetooth) we were very energized about the “green” aspects of this product and how it could also be utilized to help the general public and commercial fleets save money on fuel costs. It was not long ago back in 2008 when gasoline had gone over four dollars a gallon in the USA setting records and starting a shift towards cost reduction and sustainability for those who used significant amounts of fuel for business purposes. We knew that was a billion dollar market but needed to figure out how to tap into it. The idea of reducing emissions and slowing climate change also was an important core value for the founding team and investors.
2010
Towards the end of 2009 and going into 2010 we now had a number of products, concepts and potential clients with various markets we could move into.
We had been working on reaching out to everyone in our networks, trying to find an application for the technology the four of us had developed in a literal garage bay we had rented out in a small town outside of Albany, NY. A friend suggested that we look into a data collection and measurement approach, as he had done private security work and knew various pharmaceutical companies were having significant issues with the theft and hijacking of trucks in other countries.
This was a bit of a light-bulb moment for us. What if we could take parts of what we had built and turn it into a telemetry gathering product for businesses with high-value payloads?
We returned to the drawing board and realized the OBD2 connection to a vehicle’s computing system could be huge when combined with GPS and other sensor telemetry data. We stripped out the off-the-shelf ELM 327 chip that we were licensing and decided we would build our own, and it would be compatible with Bluetooth, GSM/cellular wireless tech along with having a number of onboard sensors to measure important events.
We began retooling our SEO and sales goals. Reaching out to our networks, looking for anyone interested in such a service or device.
Almost immediately after making this pivot, we started seeing quite a bit more interest in our business and the in-vehicle OBD2 interface technology we had built. This was when a company named Logica approached us with a course-altering project that helped define the next decade of our lives and the business.
They wanted us to help them build a vehicle telemetry product. The goal was to save millions of gallons of diesel and gasoline, which meant billions in savings for clients who needed solutions like this to connect their fleets into a centralized platform that managed their operational complexity. Logica had the frontend system taken care of. Still, they needed a partner to provide the hardware and telemetry ingestion systems “at scale” with a requirement of one million vehicles operating and pushing telemetry at any given time. The task was monumental, and there wasn’t much time to deliver within their provided project timelines.
Logica’s requirements had actually lined up quite well with Carma’s at this time. We had been building a “big data” ingestion stack before that was even really a term, along with taking an API-first approach to our platform as we saw the success AWS was having with their platform’s similar approach at the time with the launch of EC2 and further cloud platform products.
The only hangup was that we were not even remotely ready to launch this new concept. The hardware was mostly ready but we had only months to get the firmware working and deliver it by hand to the clients for an initial pilot program that would take place in a number of cities across France in the summer of 2010.
We scaled up hiring, and brought in Embed Inc. to help round out the staffing to get our SaaS and hardware stacks up and running for the pilot that was only months away. At this point we became a round-the-clock operation in the old house we had rented in Delmar, NY. Our project manager (Bernie Walters) went out and bought cots for us so we could take turns sleeping in our offices between meals and long days of coding and testing. To speed up development timelines we were building OBD2 emulators as “bench testing” rigs so we didn’t need to keep flashing changes onto our IoT device and running out to the parking lot to plug the device in and validate our changes were working.
We spent long nights in meetings with our Logica compatriots who were mostly located in India and Europe, trying to hammer out any final requirements, running demos to show our progress, and preparing travel plans for some of us to spend half of our July in France to ensure everything went off without a hitch.
I can still remember driving four hours to Boston that night, putting a half dozen Pelican cases filled with padding, anti-static-bag wrapped circuit boards, and 3d printed cases onto the conveyor belt at Logan airport and thinking “it’s actually happening!” Fast-forward 14 hours and we landed in Paris during an ongoing transit worker strike. Guess how hard it is to rent a car in France when public transit isn’t working? We spent hours going to nearly every car rental location in Charles De Gaulle Airport before we could finally find what we needed, a Renault Kangoo van. The exact same van our clients were going to be using for our postal service pilot program with Logica.
I have a fond memory of Chris and I blearily wandering around the Paris business district our first night, trying to find food, and everything had closed except a small bakery. We made do with some sparkling water and each of us walked back to the hotel with a baguette smothered with some smelly soft cheese. The baker wouldn’t take my 50 Euro note that I had gotten from the currency exchange. I don’t remember how we resolved payment, I think I gave him a ten-dollar bill, which was probably 10x the cost of that food with the exchange rate at that time, but I was so tired I didn’t care at that point.
The following day, we were up and at it bright and early. Chris and I knew we were close to getting our prototype IoT firmware fully functional, and we wanted to ensure everything would go smoothly for the pilot starting in less than two weeks in Lyon, a few hours south-east of Paris. Well, we plugged that sucker into our first French car - it turns out Renault didn’t follow the OBD2 standard entirely. Our product didn’t work. I can still remember sitting with Chris in the parking garage, in the passenger seat of our Kangoo at 8:30 am, both of us thinking “oh sh** we’re in big trouble.” I don’t think either of us slept more than a few hours over the next ten days.
We were completely blind-sided by this compatibility issue to say the least. We immediately contacted our team back in New York and broke the news. We needed resources and we needed them fast. They overnighted test hardware to help us debug the issue and we spent that next day driving around Paris trying to find the right FedEx location that received our hard-core debugging tools and extra circuit boards. We were going to make this work, or die trying.
Chris and I sat in our comfortable but small hotel room in the business district (conveniently located next to the parking garage where our Kangoo was sleeping peacefully.) We pair-programmed our way through nearly every single issue until the early morning hours. Walk through code, validate logic, burn the firmware, hike up the stairs to the Kangoo, plug in in our wireless telemetry unit, cross-fingers, and go for a test drive up and down the parking garage. We’d call our VP of Sales (Tom Sinopoli) who was back in the hotel room and ask him to validate the right data was coming back into our (homemade) IoT observability platform. Tom (who passed away in 2018, RIP) was not the most technical guy, so it was certainly not the ideal scenario I think about when trying to iterate quickly with crazy deadlines.
The day to launch our pilot program with the French Post came. We packed up the Kangoo with a few dozen (hand-built!) IoT prototypes and drove across the beautiful French countryside. It also reminded me of upstate NY dairy country, and felt a bit like home which was nice, because I really needed to feel something at that time beyond the doubts that were flying around in my head about if any of this would work.
We pulled into a rural post office. A Logica employee had joined us who told us to say we also worked for Logica if anyone asked. This was the first time a client would ask us to pretend we worked for them and were not our own firm, and it certainly wouldn’t be the last either! Everyone wanted us to be “their” team.
The French postal employees were not happy. Why? Ah, right. They pointed at the black-box and said the only English words that they’d utter during this entire transaction: “GPS?” I being the naive tech entrepreneur that couldn’t see why anyone wouldn’t want this cool new tech!? Was … a bit taken aback. We had spent years getting to this point, months of nonstop development, sleepless nights and blurry days. They didn’t want it? No matter, it wasn’t their decision. We installed twenty of these in their fleet of postal vans and drove back to Paris. Our pal from Logica took us out for the best Indian food I’ve ever eaten in my entire life (still.) He promised us this “wasn’t even that good” and said if we ever made it to India he’d show us what real food tasted like (I missed that trip but Tom went and he confirmed the food was incredible!)
The celebration started, we were going to install more of these in another fleet outside of Paris the following day. Life was good and we were going to make it.
The next morning we are up early and excited to keep rolling out our new product. We saw from our remote debugging platform that the initial telemetry data was rolling in and thinks looked good. As we are packing up our bags and burning our latest firmware revision onto each board and manually assembling - we got a call. It was from another Logica employee. A postal employee said that their car had lost control, and it was our device that had caused the incident. The pilot was halted immediately. Over the course of the next few days, all of the pilot devices we had installed were actually smashed to pieces before being returned to us. We were absolutely devastated, but Logica was surprisingly keen on moving forward. They said if the French didn’t want the hardware, they had dozens of others who did and could start using the technology immediately. From their point of view the pilot was a success. We had delivered from their point of view, and it was only a matter of refining the platform. They were convinced our product didn’t cause the loss of control in the pilot vehicle. All was not lost.