Hi, Adrian, introduce yourself for a moment.
My name is Adrian, I come from bonprix and I am a project manager in a department that develops software products for the purchasing and supply chain sector. In other words, all in-house products that are primarily related to purchasing. And currently I am mainly involved in forecasting. That is, demand forecasting at product level.
What kind of lead time you have here?
Different. Ideally, it’s over 6 months. But it all depends. We have different types of products, of course. Some products have been around for over 15 years and they have never changed. However, a very large amount of products are only used once for a single collection and never come back.
How many collections do you have per year?
By now we have reached 12 collections per year, one per month.
What was interesting for you? Why did you come to this workshop?
I heard about this workshop from the head of my department. And it actually fits quite well into this forecasting topic, because of course competition details also play a role. Depending on how many products the competition is currently offering in any way, this can have an influence and we don’t take that into account – at least on this technical level.
That means that you are currently forecasting only on the basis of historical stock data?
Correct.
Is there a topic you find interesting for an upcoming workshop?
Actually everything around product attributes. So which you can possibly also extract from images. After all, these are also responsible for the success of forecasting. So I would be interested to know how much and what you can extract from images.
In other words, at the moment you know what sold very well, but you can’t give a title to it, whether it was the short sleeves or a certain pattern?
No, we also use a lot of attributes, but they are all entered manually and that’s something you could think about to automate. Especially when creating patterns for products, you have to create a lot of attributes for a product that may never see the market. And that is of course
some work. This would also be interesting, for example, that you could get automatic attribute suggestions based on photos.