Smart Waste’s chief product officer talks to Kurrant Talent about the challenges he faced and the importance of data when it came to designing the smart waste collection software for the company’s Waste Vision solution.
As Chief Product Officer you have designed the smart waste collection software for Waste Vision. Can you explain what are the main challenges you have faced in this journey?
The crucial point in the software development was the waste collection planning component. Many existing solutions at that time used a plan board or manual planning that was still relying on the team manager collection planning experience. So the challenge I had with the team was to create a solution that could show the expected collection need for the coming days based on the target filling level at collection that we had set. For that the model we have set up was based on the previous five weeks filling data history provided by the IoT sensors deployed on the containers. We collect and store more data (up to five years) but the model only looks at the most recent five weeks to estimate the container filling speed. I was very lucky to have in my team some very good developers who actually understood the operational requirements.
Before Waste Vision you had been for 15 years with Transvision a route optimization solution provider. Beyond software development what is the relation between route optimization and waste collection?
Waste collection has undergone a drastic change in the past 10 years, moving from a very static planning schedule to dynamic collection route planning. In the past, we used to plan on collection route per week-day and per area, only putting the collection points in an optimal sequence in order to reduce the collection route length. This was resulting in collecting at an average filling level of 50%. With the addition of forecasting features that creates dynamic collection areas in a few minutes or even seconds, we can now collect 70% to 80% full bins, while still having no overflowing bins. This dynamic planning is based on route optimization algorithm similar to what we used at Transvision.
Smart Waste collection solutions enable to maximize filling level at collection while eliminating overflows. With access control it also enables to increase recycling rate. These are operational and cost benefits for the governments, cities and waste collectors. How can citizens also benefit from these smart solutions?
Citizens can really look forward to reducing their recycling taxes as smart waste solutions will help rewarding their sorting efforts. When an access control system is in place, citizens that separate their recyclable waste from the non-recyclable one end up paying only for the left-over waste, a very small part that cannot be recycled. In fact, in the Netherlands, the “pay-per-deposit” scheme already exists and some areas are experimenting the credit system. Some citizens already get credits for delivering recyclable waste such as plastic. On another level, smart solutions improve citizens comfort as it enables them to locate the closest and non-full waste disposal sites by using their smartphone.
Software represents the biggest part of the price of a smart waste collection solution (vs the sensor). Is it related to the cost of each component?
No, especially if we consider the leasing model for smart waste solutions which is foreseen as the most promising one in the future. In that model, the sensor together with the software will be leased for a fixed monthly or yearly price per container, inclusive of the warranty and maintenance costs. This, of course, enables the garbage collector to predict and save on operations costs and CO2 emissions, while making sure the sensors & software are always up to date. Therefore, the cost structure is not made transparent and it is difficult to make the difference between the sensors and the software costs. In fact, I believe that the electronic part has a higher cost.
You live in the Netherlands who has, from far, the largest installed base of bins equipped with smart systems including filling level sensors and access control systems: why is this country so advanced when it comes to smart waste solutions?
One reason is the fact that the ground is mostly sand in the Netherlands – versus rocks for instance in France or Switzerland, which makes it very easy to install underground collection containers. But while you install these large shared containers, you cannot really control when and how much waste is deposited by the citizen. This is why the filling level sensors and access control systems have been largely deployed on the underground container systems. Another reason is the fact that some collectors had their offices located nearby universities and have, as a result, adopted dynamic waste collection models that were presented by some Ph.D. thesis.
Talking about the Netherland, you are French and have studied in France. What brought you to the Netherland? And how would you describe the working style there?
I have studied Mathematics and Engineering in France and my favorite subject was already Optimization. I ended up living and working 12 years in Denmark where Transvision offered me the opportunity to lead their growth in the Netherlands. Few years later, I joined Waste Vision as their Chief Product Officer.
The Dutch people are very pragmatic in their way of learning and managing. They accept technology much faster than in France. They are okay to try new things, without proven market research, and just experiment whether it works or not. I have also enjoyed the fact that there is very little hierarchy here with a pretty flat organization. You are really empowered and you get to do what you want within your scope of work, as long as you can show some results.
You have decided to move to your next professional challenge. What motivated this career change?
I am moving from being a Product Owner as Chief Product Officer (CPO) to a Chief Technical Officer (CTO) role which is a step up. The company called Plotwise is a pioneer in using data science like artificial intelligence and machine learning for route optimization. It is, in a way, very complementary to the IoT waste management area but much more technical. At a more personal level, I will be managing and coaching a technical team of 16 developers and data scientists from all around the world, rather than directly handling daily technical issues. That is a really nice challenge.
What are your key objectives for the first month in your new job? What would you like to do differently from your previous work experiences?
It has only been a few days but let’s say that my objective at this stage is to understand and learn my team’s way of doing it, before seeing where I can best contribute and eventually help them avoiding some doing common mistakes. Like in my previous jobs, I like to be open minded and take time to understand and validate first. Going too fast and jumping into conclusions is, in my opinion, a big mistake.
You might also like: