We spoke with Senior Advisor Mika Laukkanen about a project that did just that. In this interview, he shares the story of how an early machine learning model navigated market volatility and initial skepticism to become an indispensable tool on the sales floor.

‍

Who are you and what do you do?

‍I'm Mika Laukkanen, and I work as a Senior Advisor at Data Design. In practice, I participate in various data and AI projects.

‍

You have extensive experience in utilizing machine learning and AI. Is there a particular project that has stuck with you?

‍That's a tough question; there have been quite a few projects over the years. Perhaps pricing cars using machine learning is one of the most memorable.

‍

Why?

‍It was from a time when these kinds of solutions were just emerging, and very few of the projects that were started actually made it into production. This was one of those that did.

‍

What was the project about?

‍The goal was to utilize machine learning for pricing used cars. The idea was that it would save salespeople's time, standardize pricing, reduce pricing errors, and improve sales efficiency. An additional option was the ability to estimate the value of the inventory and its development based on the predicted prices.

‍

What kind of solution did you end up with?

‍We implemented a predictive model based on historical prices and available car data. The end result was an application where a salesperson enters a license plate number, and the application returns an estimated selling price and, for example, a predicted time to sell.

‍

‍

What was the reception like for the solution?

You could say that, initially, the reception among the salespeople was mixed. For instance, they quickly noticed if the model returned clearly distorted predictions, which was good feedback for developing the model further.

This also made it very clear that an objectively "correct" price doesn't even exist; the same car can be priced in different ways. And ultimately, this can also depend on the salesperson's own experience and perspective.

In any case, the use of the application has since become an everyday tool.

‍

What kinds of challenges arose in the project?

‍Perhaps the most memorable challenge was that car prices change quickly. However, the training data spanned a long period, meaning the old prices were no longer comparable with the present moment. We had to find and test solutions for this.

‍

Any other interesting observations?

‍There was clear variance in prediction accuracy between car brands. Let's just say that for so-called premium brands, the predictions worked with a smaller margin of error. Additionally, newer cars get more accurate predictions than older ones, which is a fairly predictable characteristic.

‍

Could the same solution concept be applied elsewhere?

‍Yes, certainly. I believe a similar concept could be applied to pricing things like apartments and many other products. The main thing is that we have enough high-quality data.

‍

What was the best part?

‍The fact that the results went into production and further development. That's when you know you've done something right.