Insight

I’m Julien Rollon, I’m in charge of Vekia’s deployment in Mr. Bricolage Group. I have a professional background that is mainly indistrual. I’ve done master’s degree in the domain industrial management with a specialization in ERPs and information systems. 

I have spent twelve years in the industry in an international environment. I joined Mr. Bricolage in 2015, in order to deploy Vekia solution through machine learning. 

Today Vekia solution is part of nearly 200 stores throughout France of size more or less large, from the type of store called “city” that is 800 square meters to the one called “specialist” that is bigger than 4500 square meters. Meaning that in terms of references, we can have from 16 000 references to 90 000 managed in stores. 

It is fully radical, I have settled into Oracle Business Suite, into SAP, into SEDGE for the calculation of needs, of the production, we were still on an archaic system where there were sales and forecast analysis that used to be done manually or through Excel files that we had to pilot.  

Machine Learning, AI compose this powerful data analysis made in a short amount of time, in particular during night time when everyone “rest”, it is an impressive gain in terms of data and of results and performance. For my part, coming from the industry and knowing the old systems, I was like most of the people, a little bit dubious concerning this way to work. But in reality, for me to be credible towards stores and to see that the result is real, I have to visually expose these results at the level of stores and realize that the reliability rate is beyond 90%. 

For instance, if we take a key product at Mr. Bricolage such as the white spirit, Vekia’s software tell us that in 3 weeks we will sell 100 of it, at the end of these 3 weeks we sold 90 of the 100. Nowadays, having a tool this effective in terms of reliability and forecast is a big added value and it is comforting for stores. 

Even when the economic activity is not meeting the expectation, we have defined standard management rules with a couple of adherents. Nowadays, by deploying the solution in a standard way I win around 10 and 20 days of stock lifetime per year. 

Therefore, there are some stores that win a little bit less stock lifetime, maybe because they are less constant and maybe fearful concerning the configuration and the confidence in the tool. But some other can win tens of coverage days per year. 

If we talk about the adherents, the problematic of the information system of infrastructure management is that they don’t care about it, the added value that I bring them is the fact that I pilot by myself, on Vekia solution. When I say piloting, I mean in terms of data but also in terms of computer processing. And the fact that I have my expertise and my rigor of industrial manager allow me to support them to optimize their stocks and resupplies.  

It is true that concerning the tools and IS, there are two philosophies: 

There are many companies that will say that the user has to adapt itself to the system, and on the other hand other companies will say that the system has to adapt itself to the user. 

For my part, in order to help Mr. Bricolage’s stores, I have made the system powerful, intuitive and the most productive. Personally, I can perform individual configuration in companies’ office packs where Vekia is interfaced.  

We have created different processing in stores. Some processing by article that we call “mass change” where the adherents can, depending on different management policies adjust their minimum stock whether by nomenclature, by family, by supplier or even by product. They also can define a default Supply Chain with Vekia’s solution if they feel like they are a bit lost in their choice. We have to tell them what can the solution do, and also give them the ways and process for them to indirectly be able to have the control on the tool functioning. 

Of course, I do have some piloting and monitoring indicators that notify or alert me if a management policy is not judicious regarding the store functioning.  

Also, the fact that stores, on every order proposal line and for every product have their current and projected stock coverage, it is the data with more added value of the tool. 

People in stores usually think that there are some products that represent good sales and some others less sales and wonder why there are not more products ordered to answer the customer demand. Also, in stores we are a lot focused on stockout.  

With this Vekia’s tool I realized that there are “good stockout” and there are especially products that are not in stockout but that are in potential disruption of supplies because we don’t have enough stock to hedge future sales that we are going to realize. There is 90% of chance that we realize the sales that the tool predicted. 

The tool is able to tell us that there is no interest in buying a certain product (that might be expensive to store) now, it will tell us that we should buy it in 5 months in a half instead. 

When we tell this to a retailer, he usually says: ”Julien, I need this product in stock”. But when we project, look forward and we talk to him about his treasury, and that we realize that this product volumetry can represent in certain stores 30% of their stock value.  

If they have 1 million € of stock and they apply this 30% of stock value for this product, it represents 300?000 € of their stock. And then we look at the cash-flow problem that is created.  

In the end, it is better to optimize your cashflow. 

What is important is to consider the cashflow, and explain the importance of the cashflow for the store, and all the optimization of the supply. 

I also track the ROI indicator, meaning the return on investments of the stocks. In fact, the minimum of information we need to analyze and make sure we are confronted to a well-managed store is to check that over a 12 months period, it generates a gross margin that is superior to his stock value at every end of the month. Vekia’s tool provided a lot concerning this aspect. 

And I realize that to this day, that I have 30% of the stores that I manage, (I manage a total of 200 stores) that generate a margin over a 12 months period that is lower than their stock value.  

Let’s visualize an income statement, the first line we read is the turnover, then it is the line of stock value. If we realize that even at this level the store is not profitable and we didn’t even add fixed expenses, wages, taxes and other taxes, we notice that these stores, have a low financing capacity that is negative. So here the profitability is not present at all. 

The goal is to have the most powerful ROI, some stores I managed can get an ROI 2 or 3 times higher. For instance, over a 12-month period, a store that has 1 million € of stock will generate 3 million of gross margin. We then know that we have this gap that will allow the store to deduct wages, taxes, fixed expenses. And logically, these stores will have at the bottom of their page a line with a positive self-financing capacity. It means that their activity in a fiscal year is profitable. 

I tend to say that “we don’t see the end of it”. As we can say, data is great but we have to be able to manage it, and use it the right way. Some data are not useful, it is necessary to choose the right ones, and to work with it in an efficient way. 

Managing a volumetry of data represents the need to know that there is a certain number of data that we will not make use of. The data that we will not use represent a cost, because this data is stored on servers. That is why it is important to identify which data is useful for our company in order to keep and save the useful ones.