DSN: publishers a year ahead of schedule

Open letter from SDDS to DSN stakeholders

“SDDS, the association that brings together software publishers and service providers specializing in payroll and Human Resources management solutions, has taken note of the government’s announcement to rearrange the timetable for the future DSN. Initially scheduled for January 2016, the government will propose, as part of the Social Security Financing Bill for 2016, a timetable for the gradual introduction of DSN to enable widespread use by July 2017.

SDDS, which lent its support to the government from the outset of the project and its expertise by mobilizing its members for concrete implementation from January 2016, regrets this late and hasty reorganization.

The major cause: did we really know where we were going?

As the tests progressed, it was not parameterization or development anomalies that came to light, but cases not foreseen in the specifications, and varied interpretations by the social security organizations (OPS), which a lack of sharing and involvement on their part meant could not be anticipated and corrected within a reasonable timeframe.

As a result, software publishers and strategic project management (MOAS) teams were confronted with an abnormal variety of interpretations of social reality indicators, the result of the sovereignty of the SDAs, a variety that became critical when it came to consolidating everything into a single multi-organization declaration via the DSN. In other words, where we said we were testing the software, we were testing the specifications.

For phase 2, the specifications underwent 5 versions in 12 months, and 6 maintenance logs totaling 68 patches, all spread over a period of 20 months. The software publishers were therefore forced to constantly modify their software to comply with the ever-changing specifications, starting the tests from scratch each time. It therefore seems wrong to blame the delay on the software, simply because it is the last link in the chain, and because it only involved a few thousand companies in these costly tests, rather than hundreds of thousands.

A costly sequential method ill-suited to project constraints

The sequential method imposed led software publishers to implement versions on their customers’ premises that had not been tested by the OPS, who only looked into the matter at the end of the process.

In other words, no advantage was taken of the state of the art in software testing. The companies called in to test, presented as “pilots”, actually carried out unit tests on partial chains, in order to draw in their wake OPS still dubious about the project and compensate for the 7carences of the specifications.

Communication is part of the project!

The DSN project benefited from a new approach to project management: the inclusion of publishers (SDDS) in the steering group. This novelty, combined with a committed and talented MOAS team, enabled us to achieve remarkable results, given the challenges and difficulties of the project. However, this did not compensate for the limited resources mobilized, nor for the project team’s limited prerogatives.

By way of example, communication on the rescheduling of the timetable had a disastrous effect on companies, contrary to what had been expected. In the absence of the rescheduled timetable, what was heard was a postponement sine die, resulting in a demobilization of companies. Here the citizen hears the wise argument “flexibility offered to companies”, but there the company understands “postponed to 2017” and postpones its own involvement.

For such complex and sensitive projects, involving a large number of players and with repercussions for all, communication is also part of project management.