Product People Advanced a Pricing Platform to Tackle €3M in Annual Errors

The Client: Publicis Media

Publicis Media is the media arm of Publicis Groupe, one of the world's largest communications groups. Its media planning and buying team negotiates volume discounts with TV, audio, and digital media owners on behalf of major advertisers. Those discounts directly affect what advertisers pay and what the agency earns, making accuracy a commercial priority. The team identified an opportunity to modernize a process that combined manual spreadsheets with a 12-year-old desktop application. It estimated that human error in the process cost around €3 million a year, alongside more than 264 hours of manual work each month.

‍

The Mission: Interim Product Manager for a new pricing application

Publicis Media saw an opportunity to turn its deep pricing expertise into a modern platform. After investing about a year in discovery, the team had developed a strong understanding of how discounts should be calculated. Bringing that knowledge into a scalable product meant coordinating dependencies across teams and systems and translating specialist rules into buildable requirements. Product People joined as interim Product Manager during the permanent Product Manager's parental leave, with a clear brief: help the team deliver on that opportunity.

‍

Our Main Quest: Making a complex product deliverable

Launch: Built momentum in the first week

We helped the team move toward delivery quickly by taking over the backlog and team rituals in the first week, setting up weekly working sessions with the business owner and designer, and breaking large requirements into smaller, buildable pieces.

Explore and Conquer

Initiative 1: Turning ambition into a steady delivery flow

• Result: 5 parts of the application delivered, 2 more in development at handover, on roadmap.
  • Because: we planned in small increments, kept several workstreams running in parallel so developers were never blocked, and protected a stable stream of ready work ahead of each sprint.
• Result: everything delivered was reviewed and signed off by the business teams who will use it.
  • Because: we ran hands-on review sessions where users tried the finished features with developers in the room, so feedback landed while it was still cheap to act on.

Initiative 2: Making the most complex part of the product buildable

• Result: the entire pricing calculation area moved from expert know-how to a documented, ready-to-build scope.
  • Because: we mapped its 50+ rules and dependencies end to end with the business experts, wrote the requirement documents and product specs, and kept a decision log so answered questions stayed answered.
• Result: the incoming Product Manager had a foundation for continuing the work.
  • Because: we recorded onboarding walkthroughs explaining the logic and open trade-offs for the Product Manager taking over.

‍

Mini-Missions: Our Side Quests

Alongside the main delivery track, we partnered with the team on opportunities to reduce friction and strengthen the foundations for future work.

Legacy retirement plan: Gave the client a concrete, phased plan for retiring the 12-year-old legacy tool. Together with business and engineering, we reviewed the old application module by module, mapped each function to its counterpart in the new platform or marked it as intentionally dropped, and made dependencies visible in a multi-year retirement sequence.

Reusable shared module: Avoided a duplicate build and diverging sources of inventory data by making a shared product module reusable across teams. When a neighboring team was preparing to build the same capability, we brought both teams together to agree on a shared scope and data model, then documented the interface so the module could be used across teams.

Naming and glossary alignment: Reduced repeated clarification loops between business and engineering and gave new team members a shared onboarding resource. We aligned English naming conventions and built a glossary that connected specialized German trading terminology with the English used by engineering.

Continuous business validation: Kept validation running throughout delivery rather than concentrating it in a single pre-release phase. We established a structured testing and bug-reporting process so business users could test each finished feature in the test environment as it landed, report findings using an agreed template and severity levels, and have those findings triaged into the next sprint.

AI opportunity exploration: Gave the client a prioritized shortlist and feasibility assessment for potential future automation. Together with the data and architecture teams, we identified candidate AI use cases across the offer and calculation workflow and assessed each against available data and existing infrastructure.

‍

Mission Achievements: Delivered Outcomes

📈 A clear path for continued delivery: 5 parts of the application delivered and approved by business teams, with 2 more in development at handover.

📈 The most complex area of the product, 50+ interlocking pricing rules, is fully scoped and documented, so development can continue at pace without depending on individual memory.

📈 Continuous business validation during the mission, with users testing working software as features were delivered instead of waiting until just before release.

📈 A path to addressing the estimated €3 million a year lost to human error, as the platform replaces manual steps piece by piece.

📈 A handover the client could absorb without losing momentum: clean documentation, recorded walkthroughs, and a phased plan for what comes next, ready for the returning Product Manager.