The IBP implementation delivered. The transformation didn't.
The system went live. Eighteen months later, three of six markets were running parallel Excel files. A shorter version of what actually happened.
Panorama Consulting tracked 271 supply chain planning implementations in 2023. Fifty-eight percent exceeded budget. Average cost overrun: 24%. Average timeline overrun: 33%. Only 41% delivered expected business benefits within three years of go-live.
The implementation I'm describing here was in that 59%.
The system worked. The data loaded. The statistical models ran. The go-live happened on time and under budget. The programme director's email to the steering committee said "green" on every milestone. By month eighteen, three of six markets had rebuilt their Excel files alongside the system.
The failure followed a pattern.
Month 1–3: genuine productivity. The consultants are on-site, the steering committee is watching, and the data cleansing done during implementation has improved quality. Forecast accuracy improves 3–5 points.
Month 4–6: the first data crisis. In this case, promotional history that had been loaded without promotional flags — the statistical model was forecasting from a history that included uplift periods without knowing they were uplift periods. The demand planner who knew this had been correcting for it manually in her Excel file for nine years. She raised it. The fix took eleven weeks. She never fully stopped using the file.
Month 7–12: consultant departure. The implementation contract covers go-live and stabilisation. By month eight, stabilisation is declared. What leaves with the consultants: institutional knowledge of design decisions, ability to escalate vendor issues quickly, and the energy keeping the steering committee engaged. What remains: a team that can operate the tool but cannot configure it, and a vendor support contract that handles bugs but not adoption.
Month 13–18: parallel files everywhere. Each file is a reasonable response to a real gap — NPI process too slow for commercial launch speed, customer-specific forecast formats that don't fit planning buckets, promotional correction logic that was never fully built. Each file is also a fracture in the "one version of truth" the implementation was sold on.
Three things actually failed.
The process was designed for the organisation that was aspired to, not the one that existed. Future-state assumed commercial would provide structured monthly volume input. Commercial provided deal-by-deal information two weeks before execution. The gap was never closed.
The incentives didn't change. Planners were measured on forecast accuracy. When the system's baseline was wrong, the fastest path to the right number was the Excel file. The system recorded the outcome. The file contained the reasoning.
The change management programme ended at go-live. It measured whether users could operate the system. It did not measure whether they trusted it enough to stop using alternatives. After go-live, the budget was exhausted. The resistance — suppressed during implementation by programme energy — re-emerged.
Total cost: $4.2M implementation fee, $1.1M in fixes, 2,400 person-hours of internal time managing the system-versus-file gap. Forecast accuracy improvement: 4 percentage points. Inventory reduction: 8% in two markets, increase in one.
The lesson is not that IBP doesn't work. The lesson is that the implementation contract prices the technology and the go-live. The transformation is priced separately — in Year 2 resource, incentive redesign, and sustained executive attention — and that price is rarely quoted upfront.
Full post-mortem with sources: IBP rollout, month 18