How to build a responsible full stack

Whatever their raison d'être, but with specific business and regulatory needs including Corporate Social Responsibility (CSR) obligations, many companies choose to invest in the development of customized solutions rather than using market solutions. This complementary activity of design and development of management applications, often requiring the involvement of a specialized company, mainly falls into two categories of the Responsible Digital Challenge:

• Either the creation/adaption of tailor-made applications to support the company's CSR strategy (IT for Green)
• Or the reduction of the environmental and social impact of applications linked to the company's operations (Green for IT)

Created to manage products or services, this software ensures that the promise is implemented for a defined end-target and controlled by the issuing company, returning to the original meaning of the word responsibility.

3 examples of responsible digital best practices

Restrict the number of features of the application

Many software applications in a company can already partially produce the same result or an approximation of a "new" expected result expressed in a project (e.g.: managing schedules is possible on various tools such as Notion, Teams, Assana, iLuca, Jira, etc). Identifying redundancies and priorities then allows for arbitration. By focusing on the business value via the expected user experience, certain functionalities can be questioned: choose not to develop, to integrate third-party services rather than develop from scratch or to develop the bare minimum to create bridges and allow continuity between tools. Let's look at how responsible digital thinking can limit risks:

• Economical: Rethinking the scope with a digitally sober approach typically lowers the direct development bill, which can easily exceed tens of thousands of euros, but it also may reduce the time required to train future users on the new solution, a cost that is frequently overlooked in internal project costings.
• Environmental: Data centers are constructed using scarce, non-renewable, and non-recyclable metals (such cobalt), which are harvested using polluting methods under unfavorable working circumstances in underdeveloped areas. Rationalizing the need for server resources helps to avoid the multiplication of orders and operations of these costly machines in terms of GHG (Greenhouse Gas) emissions.
• Legal/social: Limiting the number of features means the end user (employee) spends less time learning how to do their job.

Ask for or gather information that is specifically required for the service to operate properly.

Data minimization benefits both service providers and clients, whether internal or B2B, B2B2C, or B2C. The general public, employees, and businesses are becoming more and more conscious that their personal and professional data, no matter how modest, has value and that there is a risk of fraud or improper appropriation. Some information systems are reluctant to be used or even rejected because of areas of uncertainty and reliability concerns. Thinking about the need to obtain this data, its direct use for the proper functioning of the service and a clear specification of user/administrator rights is a responsible approach that will limit the risks of rejection. The benefits will be:

• Economical: The cost of security to provide for anonymization procedures, reliable authentication services, and other means of preservation increases with the sensitivity of the data's classification. Less data means less cost and a budget focused on essential data.
• Environmental: A data center consumes the water of 6.5 Olympic swimming pools per day, can consume the energy of 80,000 homes, and its end-of-life terminals generate several million tons of waste each year. By not storing uncontrolled and unused big data, you can help reduce the demand for machines in your data center.
• Legal/social: User transparency and trust facilitate the adoption of solutions, limits potential data leaks/cyber-attacks that could jeopardize or paralyze one's business, prevents prosecution and sanctions for non-compliance with the GDPR.​​​​​

Keep technologies up to date and regularly iterate the functionality of solutions 

The maintainability of software, from a functional and technical point of view, is a critical issue, especially for companies that do not publish software but have developed custom solutions. To remove this risk, it is important to establish a long-term methodological plan, integrating good user-centered design practices from the MVP (Minimum Viable Product) onwards and in all future iterations, so as not to blindly pile up developments from version to version.  The reflection around the roadmap aims to prevent known pitfalls in the software publishing world:

• Economical:  
  • (Technical) A project whose code cannot be taken over, due to a lack of resources trained in the technology, runs the risk of being redone from scratch. In 2022, some technologies were widespread, whilst older ones were only used in a few niche industries.
  • (Functional) A tool that does not adequately or no longer serves a true need is abandoned in favor of workarounds, which may incur additional costs on top of not using the already-purchased tool. A project that is not reviewed is likely to lose its meaning over time, as the business and the initial tasks it was designed to address evolve. This goes beyond the realm of technology. The primary reason for appropriation failure is the target users' lack of involvement in the project, which is followed by a poor phase of collection/expression of the demand.
• Environmental: Some technologies use more energy than others: Perl uses 79.58 times as much as C, Python uses 75.88 times as much as C, Ruby uses 69.91 times as much as Rust, which uses only 1.03 times as much energy, C++ uses 1.34 times as much energy as Ada, and Java uses 1.98 times as much energy as C.
• Legal/social: The Standish Group published a study in 2018 stating that 31.1% of projects are stopped before completion and are never deployed, 52.7% exceed their budget forecasts by 189%, impacting the company's activities to a greater or lesser extent, and possibly leading to social plans and job losses.