How data science and machine learning can predict the future for the supply chain industry
To understand the story of Sixfold, I must first tell the story of Palleter, which was the failed logistics startup precursor to Sixfold. This story is what transformed our company into what it is today. One of our founders, Märt Kelder, was one of the first investors in Bolt back in the day (Europe’s equivalent of Uber, a unicorn ride-sharing company founded in Estonia). He figured that the same network exercise that Bolt was using could also be applied elsewhere in transportation, not just in ride-sharing like Uber, but maybe it could be applied to public transport or logistics. And from there, Kelder pitched that we try to emulate this network model in logistics, an “Uber for trucks” concept. Soon thereafter, he brought me onboard and another co-founder (we had previous experience in company building), and then we began the process of building out Kelder’s “Uber for trucks” concept. What we tried to do in essence was this: instead of having the same shipper try to obtain a usual contracted carrier to move a load, we created a very efficient “spot market”, whereby the most logical truck (the truck with the closest proximity, loading space, and fit other parameters), would pick up the shipper’s load on the back haul. For example, if there was a truck going from Paris to Berlin, and then back from Berlin to Paris, it would receive offers. Somewhere on the back trip, the truck would stop somewhere on the route (so it wouldn’t have to steer off its course), it would fill up the few remaining places it had left in its trailer, and then the carrier company could earn extra money. At the same time, the shipper would also receive a very well priced, quick and convenient service. This way, everybody would be happy.
Not only that, but the service would be good for the environment as well. Instead of sending a truck to pick up a load that was 200 kilometers away, the shipper could get a truck that would debark just 2 kilometers away. The initial idea was simple, easy and foolproof. But it failed. This is because the transportation industry works very differently from the ride-sharing Industry of Bolt or Uber. The biggest reason why logistics transportation is different from ride-sharing is because trust in the network is necessary. In other words, when an Uber comes to pick a driver up, the driver and the person being driven don’t care about the relationship between each other. But it turns out that in the transport industry, people are actually willing to pay extra money and spend more time on deliveries with the carriers that they already have trusted relationships with. They would rather do this than pay far less for carriers with whom they have zero experience or relationship with. Part of the reason is because if a carrier is late on delivering, the penalties can be very expensive. There is no guarantee that a random carrier will deliver on time, and it adds stress to the mess.
In addition, if the carriers were to use Palleter, they would have to confide 100% trust in us as a full service solution. We would have to fulfill 100% of their shipments. Imagine if we told shippers the following: hey, here are 10 companies who might have trucks passing near you, just hit them up, call them and deal with them yourself. This would simply not fly, as there is no trust being established if we are just handing over random companies to carriers that don't trust the shippers. And we did not want to be a full service solution. We didn’t want to be in the business of digital freight forwarding or brokering. Our goal was to become a big data company that figures out the network efficiency, but doesn’t actually fulfil shipments.
Because there were so many complications, we decided to shut down Palleter, and not take the 1M EUR in funding that was offered to us from different investors. However, all was not lost. We wrote a blog post about our failed experience, and it went viral over the internet. A variety of different companies reached out to us, and in the end we picked TPG as the buyer of Palleter’s assets and founded a new company using Palleter’s existing technology, Sixfold. TPG is one of the world’s largest PE companies, so we were surprised that they teamed up with 5 guys and a dead startup. However, the reason why they wanted to work with us is because a year earlier they had bought a majority stake in Transporeon, a leading logistics transportation platform in Europe. Technology wise, Transporeon tried to do something very similar to us, but because the company lacked the proper talent, it was difficult for them to innovate the solution. TPG figured that Palleter had a shot at breaking big with Transporean.
So, Transporean teamed up with us, and founded a new company Sixfold. Palleter was a marketplace. Our technology connected thousands of trucks geographically into a real-time network, and tried to sell a capacity on them without knowing if they were interested in it, if they even had a capacity, etc. In Sixfold, we made a crucial change in the business case. We used the same tech of connecting trucks into one large system, but instead of creating a marketplace, we provided visibility between the carrier and shipper who were already trusted partners and doing business together. In other words, the shipments were already happening between the trusted partners, but they lacked eyes on the shipments.
We figured that with Transporeon’s customer base, we have a go-to-market, we have the perfect tech, we have an excellent team, and we had a shot at making this a success, and we got the investment from TPG. There was still of course a huge unknown because there were not yet major successful companies in Europe that had tried to do what we did.
Can you elaborate on some key AI use cases that your company provides?
Maybe the best use case is our Live Border Map. When the Covid lockdowns were announced in Estonia and Austria (where our offices are), it ruined the team's weekend plans and forced everyone to stay at home. So we decided to have an internal hackathon. We noticed that our customers were getting worried with all of the uncertainty caused by various lockdowns and restrictions. It was difficult to get an operational understanding of what was going on. In particular, one of our biggest customers didn’t know through which border crossings they should send their shipments. The crossings between Italy and Austria had already been closed down, and the customer had to figure out how to reroute their shipments As time passed, it looked like more borders would close down and it would become very difficult to get perishable goods through. We figured that probably we already have enough valuable data and insights on this. Hundreds of thousands of vehicles were connected to our platform and from their location data we could aggregate an understanding of how much time does it take to cross each border and how long are the queues.. Gradually, we integrated data science and ML tools to map things out and already by Monday morning, we published our border map as an open-access and free product that anybody could use to check delays on most were in European border crossings.Coincidentally, our free border map went live the same morning when the whole Europe was locked down.
We were able to use ML to analyse thousands of vehicles driving through hundreds of border crossings in real-time. After a week or two of releasing the border map, hundreds of thousands of people began using it, and that’s when the entire logistics industry learned our name. Everyone from Amazon to European Commission, central banks and governments began to utilize our data. We basically got 10 years worth of marketing attention in just two weeks and became Europe’s market leader in real-time transportation visibility according to Gartner. We instantly became a brand name and it helped us to secure many customers over the next year.
You have launched an impressive visibility data sharing initiative. Would you care to elaborate on this program for our readers?
We provide our carriers with the option to share their visibility data through Sixfold with any of their customers. In our space, some visibility providers still try to institute barriers because they want to protect their network, not allow their carriers to easily share data, and push shippers to join them. We believe that visibility roaming should be open to everyone. Any carrier must be in charge of their data and they should be able to share it with any of their customers, no matter what visibility platform those customers are using. For example, it doesn’t matter if you are a Sprint or AT&T user, you should still be able to call each other without barriers. The same concept applies here.
What are your future ambitions for Sixfold?
Our customers are doing business also outside of Europe. Therefore further geographical expansion in the US and other markets gets more of our attention from now on. We are also integrating more modes to our platform. We first began with FTL (full truck loads) and provided visibility for freight on the road. Now, we’ve also kicked off Ocean, Ferry, and Rail tracking shipments that are often carried through multiple means. In addition, because we’ve teamed up with Transporean (which has a massive amount of additional shipment related data), we are working to add a suite of transport execution tooling unlike anything that exists in the market right now. Lastly, our focus has always been more on the shipper’s side, but in the beginning of this year we began providing open visibility data (OVD) to our carrier customers, and we are now working towards adding more tools for them.