Skip to main content

Three projects.
Zero downtime.

Our clients trust us with confidentiality. In return, we detail our projects with complete transparency — figures, timelines, tools, methodology. Sectors anonymised, results verified.

Secteur bancaireMigration

Migration of a SQL Server data warehouse
to Snowflake and DBT

A European banking group needed to modernise its ageing BI infrastructure without interrupting daily regulatory reporting.

623
SSIS packages migrated
14
Project weeks
0
Production interruptions
-58%
Infrastructure cost

Le contexte

The client operated a SQL Server 2016 data warehouse fed by 623 SSIS packages, 89 stored procedures, and 34 SQL Agent jobs. The decision was made to migrate everything to Snowflake + DBT to reduce infrastructure costs and enable the data team to adopt modern practices.

The problem was threefold. Nightly processing windows regularly exceeded 7 hours, encroaching on business hours. The SQL Server hosting the SSIS packages was reaching end of life. And the only developer who understood the package logic had left the company 18 months earlier. Internal teams had no visibility into package dependencies.

Our engagement

The automated audit revealed that 142 of the 623 packages were obsolete (fed by decommissioned sources or never executed in 18+ months). This single finding reduced the migration scope by 23% before a single line of code was written.

The remaining 481 packages were processed by the converter. 338 were automatically converted to DBT models with full test coverage. 143 required manual rework — mainly packages containing dynamic SQL, cursor-based logic, or cross-database references.

Stored procedures containing regulated business logic (ratio calculations, standardised aggregations) were manually rewritten by our senior consultants, with systematic row-by-row result comparison between legacy and target.

⚡The breakthrough: during parallel testing, we discovered 7 discrepancies between the old and new systems — and in every case, the legacy system was wrong. Data quality issues that had been invisible for years became apparent thanks to the migration.

Project timeline

Weeks 1–2

Automated audit. Discovery of 142 obsolete packages. Scoping of forfait.

Weeks 3–6

Automated conversion of 481 packages. Manual rework of 143 complex cases.

Weeks 7–10

Parallel testing on production data. 7 identified gaps corrected.

Weeks 11–12Phased cutover (3 waves). Power BI connection migration.

Weeks 13–14

Hyper-care. DBT knowledge transfer to internal teams.

Insurance / MutualTMA & MCO

Reprise de TMA SAP BusinessObjects
after departure of the historical provider

A regional health insurance company had to replace its maintenance provider urgently after a contract termination, with no loss of service continuity.

1,847
Rapports WebI maintenus
6
Weeks de transition
99.8%
Availability rate
-35%
vs. previous cost

Le contexte

The mutual insurer operated SAP BusinessObjects BI 4.2 SP06 with 47 universes, 1,847 Web Intelligence reports, and 312 Crystal Reports. The entire reporting chain depended on this platform.

The incumbent provider terminated their contract with 3 months’ notice. The problem: no up-to-date documentation, no knowledge transfer, and a BI team of 2 people unable to handle day-to-day operations alone.

Our engagement

We deployed the scanner in week one to automatically map the 47 universes (joins, filters, calculated objects), test all 2,159 reports, and generate a complete dependency graph. In 48 hours, we had a clearer picture of the estate than the previous provider had in 5 years.

By week 4, we had a complete understanding of the environment — documented and reproducible, unlike the oral knowledge of the previous provider. By week 6, the transition was complete and the SLA contractuel actif.

In the process, the audit identified 423 reports unused for over 12 months (no recorded execution). The client decommissioned them immediately, reducing the maintenance scope by 20%.

→On top of maintenance: we planned and executed the conversion of 31 .unv universes to .unx in preparation for BI 4.3, then BI 2025. The conversion is now complete and validated in pre-production.

Maintenance scope

Component
Volume
Universes .unv + .unx
47
Web Intelligence Reports
1,847
Crystal Reports Reports
312
Planifications actives
186
Serveurs CMS
4 (cluster)
Utilisateurs active
380

Results after 12 months

SLA met at 99.8%

Only one SLA breach in the year (OS patch incident, resolved in 3h instead of 2h).

423 archived reports

23% reduction in maintenance footprint, zero business impact.

31 univers converted UNV → UNX

Preparation complete for migration BI 4.3/2025.

-35% on annual cost

Mainly through automation of the initial audit and scope reduction.

Manufacturing industryAudit & modernisation

Estate audit and modernisation
hybride SSIS + SAP BO

An industrial group needed to map a data environment built over 15 years by successive teams, with no documentation, in order to make an informed architecture decision: stay on-premise, migrate to the cloud, or adopt a hybrid approach.

847
SSIS packages audited
234
Univers SAP BO
48h
Audit turnaround
37/100
Technical debt score

Le contexte

The group operated both an SSIS/SQL Server environment (847 packages, 156 stored procedures) for data integration, and an SAP BusinessObjects estate (312 reports, 28 universes) for reporting.

The CIO faced a classic dilemma: they knew the system had technical debt, but had no way to quantify it. Every migration estimate from vendors was based on assumptions, not data.

Our engagement

The full dual-environment audit was delivered in 48 hours. The technical debt score of 37/100 indicated a moderately degraded estate, with critical issues on 3 specific areas: dead SSIS packages (18%), deprecated UNV universes (100%), and hardcoded credentials (41% of packages).

Based on the audit, we recommended a hybrid strategy rather than a full migration: migrate the SSIS data integration layer to Snowflake + DBT (where the ROI was highest), while upgrading SAP BO to BI 4.3 SP05 and converting UNV → UNX (faster, lower risk).

→The hybrid choice made sense: migrating 100% of the estate to a single target would have cost 3x more and taken 18 months instead of 8. The phased approach delivered measurable value from month 3.

Measured impact at 6 months

Overnight processing window : 7h → 45 min

Through Snowflake's native parallelisation and the elimination of obsolete packages.

Infrastructure cost : -42%

Elimination of the physical SQL Server. Snowflake billed on usage only.

Secured SAP BO environmentMigration to BI 4.3 SP05 with maintained SAP support. UNX universes ready for BI 2025.

Complete documentation delivered

Flow mapping, data dictionary, auto-generated DBT documentation. The client no longer depends on a single provider.

Does your project look like one of these cases?

Start with a free audit

Within 48 hours, you’ll know exactly where you stand. No commitment, no access to your business data.

Request a free audit → See our methodology