Hello Hassen, can you introduce yourself briefly?When we started the project with Keepeek , I was in charge of Business Intelligence of Groupement des Mousquetaires.
My missions have changed and I am now in charge of New Technologies (approximately 100 people) . These new responsibilities give me a broad vision on the urbanization of the information system.
Genesis and project objectives
How did the idea of this project emerge?In the beginning, at the end of 2013, I became interested in the new INCO law that increases the level of “product information” brought to the attention of consumers. It also specifies the contractual value of the pictures shown online. This new legislative package is particularly careful in the food sector. It had a direct impact on us. So I decided to present my direction a project to standardize our visual management processes and information products. I also approached trades teams who were highly motivated to work on this new project.
Was there an emergency to launch the project ?Yes, we had to act quickly. The INCO law was going to be enforced at the end of 2014. With existing devices, it was impossible to answer correctly. Besides preparing my file, I mapped our processes and I perceived prospects for reducing substantial costs. These are the two reasons that accelerated decision making.
Was this cost reduction objective realistic?It is important to understand that we are a group with many stores. Historically, each branch had its own image repository in very diverse forms. It could be simple raw hard drives or small databases. Management costs of these historic bases were important. The establishment of a common tool allowed then to suppress many costs associated with these numerous repositories. Moreover, these subsidiaries were sometimes selling the same products. It was very common to buy several times the same visuals with the same studios. In this case, cost reduction is particularly objective. We divided by at least two our production budget with better quality guaranteed.
Finally, for a large part of our base, we entrusted a studio the production and conservation of our visuals. A contractual clause did not allow us not to own the visuals. We had recurrent renting costs for visuals usages. It seemed appropriate to change the model and become master of our images. The last point concerns again the INCO law. The financial penalties for failure to comply are particularly incentives.
All these reasons acted as a catalyst to build this central media repository, the BIG project – Base Images Groupement.
What are the primary goals that you set?The main objective was very simple: we had to be able to meet the INCO obligations before the end of 2014. This required intermediate goals:
1- Collect all content different historical repositories and standardize these repositories.
2- Standardize the acquisition process and renewal of the visual products with our studios or directly with manufacturers.
3- Connect the BIG base with the other applications including e-Commerce websites to automate the dissemination of the visuals.
Why integrate your suppliers in this internal project group?This is a major challenge of the project. Given the number of suppliers and products (over a hundred thousand references), we had to integrate all stakeholders in the standardization process. It is also in their interests to benefit from updated visuals and information on our sites. A special device was deployed. Now, whenever you want to reference a new product, a request is sent to the provider to request a product sample. This sample is physically sent to our photo shoot provider, he receives it and shoots it according to a precise quality charter. Finally, he deposits the photos in BIG automatically. Sometimes it is too expensive to send samples (e.g. household appliances). We have introduced a variant that allows suppliers to submit their photos directly in BIG and enter the information required for indexing.
How do you guarantee the quality of the base?Images base that are pollution sources are numerous. The products often change, we estimate a 30 % turnover in our database. We must be vigilant on the renewal of visuals associated. Our quality approach is based on three axes.
Axe 1 – Automation First, we automate what we can. When adding new photos, BIG detects only potential quality problems on the file. BIG base is permanently connected with our internal product repository. We collect regularly the delta between the two databases to answer two questions : do the photos correspond to a product ? Does each product have many photos?
Axe 2 – Suppliers involvement We also involve our suppliers in quality supervision. We have deployed a specific web application (BIG suppliers) enabling them at any time to check the quality of the visuals or information we have for their products. They may invalidate products that seem not conform to them. It seemed logical to share the responsibility with our suppliers who have a portion of the information.
Axe 3- Quality team Automatic controls on required fields are produced by Keepeek application. Subsequently another partner, Kalista Solutions, controls manually, to check what the machine could not detect. Kalista particularly checks the quality of the image to be sure, for example, that the box of a product was not dented when it was photographed . Once quality control is done, the picture is completed by a number of information that comes from the repository (including the bar code , the wording of the product, …).
Who are BIG users?BIG is for many users of the group but also for external users. Internally:
We are currently working on the commercial management system outlets. In this way, the outlets will also have access to BIG content via another tool.