Apply at
Request a demo
Be called back
Even today, 85% of Data projects fail and are not put into production or industrialized.
There are many reasons for this observation, which persists from year to year: poor quality data, large volumes of data, poor acculturation of teams... Another reason is the difficulty of "on-boarding" the business lines in the implementation of the project.
The cause? The notorious gap between Tech / Data and the business, and poor definition and alignment of project objectives and challenges.
However, without this "business" approach, a Data project will remain no more than a technical project.
The best measure of success is when teams adopt Data projects on a daily basis, enabling them to better understand their business and improve their performance.
But how do you get there?
At AdvanThink, we believe that the secret to success is when business teams are the real driving force behind Data projects.
To achieve this, we adopt a business- and end-user-centric approach to the entire data processing cycle. At the end of the project, the deliverable is a Data product, in the form of a visual, easy-to-use Web App that translates the needs of the business, exploiting all their data.
We explain it all!
Why do Data projects struggle to get business on board?
Businesses' adoption of the solutions implemented for them is an essential success factor. Without adoption, no matter how many technical challenges you've overcome, it's as if you've just given up.
There are generally three causes of failure:
- Too technical: the solution deployed is too technical, and business teams are unable to use it, or find themselves dependent on IT or data experts to operate and develop it.
- Too static: the solution is too simple or fixed (dashboard), making its results obsolete or not actionable enough.
- The use of solutions not adapted to the project: this is often the case when solutions or tools already exist in the organization, but do not meet the project's challenges.
This phenomenon is by no means new. Carroll and Rosson highlighted it as early as 2005 in their Technology Appropriation Theory: the appropriation of a technology depends not only on its design, but also on the way in which users make use of it. This is not only true of Data, it's also true of the adoption of any technology: without use, there's no technology!
The conclusion is clear: without business ownership, the project will not succeed!
To maximize the success of a Data project, you need to take this into account from the outset, and ensure that the deliverable will become essential to the business.
The Data Product, a more engaging response
In this context, the implementation of a Data Product is a perfectly appropriate response for the business!
But what is a Data Product? Gartner defines a Data Product as "a curated and self-contained combination of data, metadata, semantics, and templates, including access and implementation logic certified for tackling specific business scenarios and reuse." (Hype Cycle for Data Management 2023 - Gartner). More than just a datavisualization, a Data Product is a strategic tool for the business, enabling it to exploit the data it needs and respond to specific challenges. This tool will become part of their day-to-day business, literally making them more efficient in their decision-making and delivering tangible results.
The Data product design phase is a crucial stage in the implementation of a project. It's the "last mile", the one that can tip the project in one direction or the other.
So it's important to prepare the various stages carefully - and here too, method is important.
Among the unavoidable questions to ask yourself: identify the problem, understand the use case and the key functions that will enable it to be realized, design the ideal solution to the business problem... but we can come back to this in a later article.
To make things more concrete, here are a few examples of the Data Products we produce at AdvanThink on behalf of our customers:
- Real-time bank payment fraud detection application for anti-fraud teams
- Application for scoring customers and prospects according to their habits and behavior, for use by marketing teams
- Application for detecting anomalies in the energy consumption of autonomous buildings
The 3 key functions of a Data Product that works!
At AdvanThink, we're used to creating data products that our customers can't live without.
With more than 1,000 projects industrialized for our customers, we have identified 3 essential best practices in the implementation of business-oriented applications:
- The AdvanThink methodology and our "Bottom Up" approach: the AdvanThink methodology involves the business from the earliest stages of the project. At every stage, end-users and business experts collaborate to create the most efficient Data product for end-users. In particular, our solutions incorporate a "DataOps by Design" approach, enabling the various users to be involved at every stage of the project. Our Amadea and FraudManager software offer a centralized point for supervising all Data projects, making them easier to understand for non Data experts. This collaboration between the various project stakeholders is reinforced by our UX, which provides a real-time visualization of any transformation, enabling a non Data expert to test and explore all his or her data on the fly. This agile, iterative approach has proved its worth, and is now adopted by 100% of our customers.
- Functionalities designed to meet business needs: the Data Product must have been designed with users in mind, for users. Each of its functionalities responds to a business need, which has been involved in its design. It is therefore designed to meet the exact needs of the end-user. It is also designed to be maintainable, scalable and to respond to future uses.
- The importance of real time and early warning: the Data Product must be conceived as the business line's armed wing, enabling it to deal with any unforeseen events and increase efficiency. The ability to implement use cases in real time is essential in certain cases, such as the fight against bank payment fraud. For most use cases, the ability to raise alerts is an essential feature. This will include the ability to automatically warn the business in the event of anomalies or abnormal behavior, or of thresholds being exceeded... An essential feature of the Data-enhanced business!
In short: we offer tools for designing and maintaining simple, high-performance data products... and in French!
Do you have a Data project and want to start out with the best chances of success? Contact our team
