Marco at FashionTechDays 2017
Roubaix, France, October 25, 2017
Hi everyone, I am Marco of Evo Pricing.
In order to find added value, to improve profitability, you have to find where there is some waste and in retail there is a lot of waste.
Waste in the retail sector is mainly generated if, when customers go to a store, they cannot find the items they are going to buy, they do not find their size. In this situation there is a lot of waste because customers will not buy, and you leave money on the table.
If you make price reductions because you are afraid of not selling your excess production and you cannot even sell it by greatly reducing the price, you lose money and there will be a lot of waste that you can turn into profitability.
What do we do?
We use artificial intelligence to make an estimate of the probability of selling each item and we do it with data we take from the company. We normally use data from past sales but that is not enough and to improve forecasting we also need to use external data. We will use data that can better predict the actual demand, not the data on past demand and we’ll check what the competitors’ prices are, we will use the time, we will use geography, the location of the stores to understand it.
But, to be even more precise, we also use data collected by store managers. Every Monday morning, we make a proposal to each store on the items it should have on sale in the coming week. Everything is calculated with our algorithms, but the sales staff can better understand whether an item will sell or not, and they may refuse it or ask for another article.
We collect this data from 1,000 stores, for example, from our customer, and all of these data form a collective intelligence using the algorithm, along with other data.
In this way we can be extremely precise with our forecast and when we have a very precise forecast. . . When we have an accurate forecast, we can make decisions very easily and very quickly on item prices. We can decide when to do promotions, and when not do promotions to avoid losing money. You can make decisions about what items should be in stores every day or every week and, above all, you will have in the store what will actually be sold, and you can make forecasts for production or purchases. So you can prepare better for the season to come. And you can measure all those things because, normally when doing promotions, you do not really measure what’s going on. If you are measuring this, you can make very quick changes and you can improve forecasts. Algorithms will always say something that is the right thing for that moment.
We are a startup that was born 4 years ago. We have 2 offices: one in London, where I am based, and another in Italy, in Turin. We work a lot with universities, we do a lot of research – we worked with this institution to do research. Giuseppe, my colleague, spent 3 months here doing work for his doctorate. We work with the London Business School, ESCP, Harvard, the University of Turin. We are about twenty people, we are growing a lot and we deal in retail, and data science for retail.