By Matthijs Roobeek, Growth Analyst
As Picnic is relatively small, and because we’re growing so fast – you’re given responsibility straight away. Two weeks after I started, I was doing the order forecasts. That’s a lot of responsibility. Many of our processes rely on the forecasting; the products we buy, the shoppers and runners we plan for, the development of the fleet. In order to write the forecasts, I used the existing models based on historical data. By incorporating the number of orders already placed, we saw real improvement in the model. The process is ongoing, and each day we use new tools to develop more involved and accurate models.
Next to this, I’m also coordinating the Sunday delivery roll-out. This is interesting, but we have to be careful. We have to be sure that capacity and amount are aligned. The main challenge is doing it fast enough without blowing up the system. In the beginning, it was hard to predict the popularity of Sunday as a delivery day. Now we know the numbers, we can gradually roll it out.
When customers install the app, they are put on a waiting list. I determine how many people we let in each day. Right now, this is done manually, but if we have an accurate model of how invites turn into orders we can automatically calculate the invitations. That's what I'm currently working on. It’s not a big project yet, but when I have a working version we can dedicate more time to it.
When you see an opportunity that can benefit the company, you’re given the freedom and trust to work on it. Your influence is very direct – and that’s incredibly motivating. As a kid, I was always playing with LEGO. Do you remember the excitement when you built a vehicle and got it to work? When I see my models and algorithms performing, I get that same feeling.