Skip to content

New back officeAutodistribution2021 → 2026

Rebuild Autossimo's back office, live in production

The back office where the commercial offer shown on Autossimo is configured. A tool in production, with security flaws and error-prone configuration that cost sales. Rebuilt end to end, rolled out across the network, then the old system unplugged.

  • Rebuild
  • B2B back office
  • Security
  • Journey design
  • Nationwide migration
With

Impact

  • 100%of distributors migrated to the new back office (September 2026)
  • 0all-at-once switch-over: migration in waves, old tool kept running
  • 259test cases so the product can be handed over
Before
  • Security flaws, flagged by the CISO
  • Poorly designed configuration, full of errors
  • Errors visible on Autossimo, lost sales
After
  • Role-based permissions, security requirements met
  • A redesigned journey: search, bulk edits, imports
  • The whole network migrated, the old system unplugged

The whole network manages its customers and offer in the new back office. Once the migration was complete, the bridge to the old system was cut: the product moved into day-to-day operation and continuous evolution.

Metrics trackedTime for a routine operationConfiguration errorsDistributor autonomySupport requestsMigration progress by wave

The engagement Autossimo, Autodistribution's B2B e-commerce platform: 23,000 automotive professionals, 3 million catalogue parts, €700M in annual business, several interconnected management systems (ERPs). Freelance Senior Product Owner from January 2021 to October 2026, with business teams, IT, the integrator SQLI and partners.

The brief

Make the tool that configures Autossimo's commercial offer reliable and secure: rebuild it entirely, journeys included, while it runs in production, then migrate every distributor onto it.

The context

The back office is where distributors and head-office teams configure what each garage sees on Autossimo: its account, the brands offered to it, which warehouses are checked for stock and in what order, promotions. Two problems at the start. Security: the tool had flaws and no longer met the CISO's requirements. Reliability: poorly designed configuration, full of errors, which showed up directly on Autossimo and cost sales.

The challenge

The modules at stake were sensitive: the commercial offer, stock across several warehouses, promotions. Revenue depends on them. They had to be understood in depth first, improved along the way, then migrated with no outage while Autossimo still read the old database. And distributors had to adopt new, very technical notions: which brands each garage sees, in which order stock is searched, which warehouse to fall back on when the first one doesn't have the part.

What I did

  1. Understand the sensitive modules

    Module-by-module audit of the existing tool with the CISO, distributors and head-office teams: what each module does, where configuration errors come from, what directly affects the garage's offer. Modules then prioritised by value for the network: commercial offer, multi-warehouse stock, promotions, catalogue error reports, handling of certain customers, logistics alignment.

    • Audit
    • Business workshops
    • Prioritisation
  2. Technical and functional scoping

    Solution chosen with IT and the integrator SQLI. Data model (conceptual and logical) and a new MariaDB database, role-based permissions to close the security gaps, a sync bridge (RabbitMQ queues) to the old Oracle database while Autossimo still read it. Specs and user stories written with SQLI, in 3-week sprints.

    • MariaDB
    • RabbitMQ
    • Role-based access
    • Jira
  3. Rethink the journey with the product designers

    The back office's user journey fully redesigned in Figma, day to day with the product designers: search, bulk edits, imports, exports, duplication, and a co-built design system. The goal: configuration that prevents errors instead of letting them through.

    • Figma
    • Design system
  4. Make it reliable before shipping

    End-to-end testing, from input in the back office to its effect in Autossimo. Test suite rebuilt from scratch (259 cases, 108 suites), every incident turned into a regression test. The destocking module, built and tested, was stopped in 2025: it did not hold up against catalogue constraints.

    • Qase
    • 259 test cases
  5. Roll out and drive the change

    Migration in waves across the whole network, through to full production. Distributors trained on the new back office (an 81-slide deck) and walked through the new concepts, with support at every wave. Field feedback collected in Intercom and fed into the backlog. 100% of distributors migrated by September 2026.

    • Training
    • Intercom
    • 100% of the network
  6. Run it, then unplug the old one

    After the migration: support, access management, evolutions. Then the sync bridge cut: the old system is no longer fed, the new back office is in day-to-day operation.

    • Run
    • Decommissioning

My role on this project

Driven end to end, all the way to production.

  1. UnderstandNeeds gathered with the business, the field and partners.
  2. ScopeScope, business rules, specs and user stories.
  3. PrioritiseBacklog ranked by value, shared roadmap.
  4. Build3-week sprints, coordinating business, IT and vendors.
  5. TestFunctional acceptance testing by journey, in Qase.
  6. ReleaseNationwide migration in waves, distributor training, full go-live.
  7. RunSupport, access, evolutions, then the old system unplugged.

Change management Distributors trained on the new back office (an 81-slide deck), walked through the new concepts (brands visible per garage, stock search order, fallback warehouse) and supported at every wave, until 100% of the network had migrated.

Testing, properly tooled Acceptance test suite restructured in Qase, test cases generated and automated with AI, every fix retested and every incident turned into a regression test.

Migration & decommissioning

Build the target, then switch off the old, breaking nothing.

Autossimo still read the old Oracle database to show garages their settings and commercial offer. The new back office was therefore built on a new MariaDB database, linked to the old one by a sync bridge (a RabbitMQ message queue), kept running until the old database was switched off for good. The whole network migrated in September 2026.

100%
of distributors migrated to the new back office
0
all-at-once switch-over: migration in waves
259
test cases to hand the product over
Old back officeOracleold databaseAutossimogaragessettings, offerNew back officedistributorsMariaDBnew databaseRabbitMQ syncbridge cutswitched offdistributors, in wavesOld back officeOracleold databaseAutossimogaragessettings, offerNew back officedistributorsMariaDBnew databaseRabbitMQ syncbridge cutswitched offdistributors, in waves
  1. 1

    Keep the existing running

    The old back office and its Oracle database stay in production: Autossimo keeps reading settings and the commercial offer there, with no disruption for garages.

    • In production
    • Oracle
  2. 2

    Build alongside

    The new back office is built on a new MariaDB database. A sync bridge (RabbitMQ queues) copies every change back to the old database, so Autossimo always shows the right offer.

    • In construction
    • MariaDB → Oracle
    • RabbitMQ
  3. 3

    Switch over step by step

    Distributors move to the new back office in waves, nationwide. Once the whole network has migrated, the bridge is cut and the old database switched off.

    • No big bang
    • Bridge cut at the end

How I drove the roadmap

Two roadmaps, one trajectory.

  • Own the functional roadmapI build and keep the functional roadmap: modules, journeys, delivery order. The technical roadmap (data model, new database, sync bridge) is owned by IT and SQLI; I align it with mine at every planning session.
  • Prioritise by value and by readinessA module goes first when it shapes the garage's offer and its technical dependencies are delivered. Otherwise it waits, and the backlog is steered towards what can go all the way to production.
  • Stop a module when it has to stopThe destocking module, built and tested, was stopped in 2025: it did not hold up against catalogue constraints. Decision shared with the business, effort redirected to the migration.
  • Explain every trade-offPrioritisation workshops with the business, field feedback qualified in Intercom, 3-week sprints with SQLI: every trade-off is made, explained and reflected in the backlog.
  1. Commercial offer: brands per garageDirectly drives sales
  2. Stock across several warehousesDirectly drives sales
  3. PromotionsExpected by the whole network
  4. Catalogue error reportsMakes the garage's offer reliable
  5. Stopped in 2025
  6. DestockingDoes not hold up against catalogue constraints
Built and tested, destocking was still stopped: it did not hold up against catalogue constraints. The effort went to migrating the network.