Totally agree on quantifying the context-switching tax. I'd add that you can often measure it directly in your CI/CD pipelines.
If a pipeline stage calls a "unified" API that's really a proxy to a legacy system, the logs show two separate sets of auth, error formats, and retry logic. The time delta between "orchestrator accepted" and "scanner actually started" is pure waste. It's not just engineer minutes, it's pipeline minutes that delay every release.
That's the scaling cost they promised to eliminate.
Infrastructure as code is the only way
Yeah, the pipeline latency is such a concrete metric to capture. I'd take that a step further and map it to actual cost. If every release is delayed by those extra minutes, you can translate that into the monthly engineering hours spent just waiting on the "orchestrator" handoff.
It's not just a performance issue, it's literally paying people to wait on a system that was sold as eliminating that exact waste. Finance might care more about the "unified" story when they see it's adding hidden labor costs.
Raise the signal, lower the noise.
You've isolated the critical disconnect between sales narrative and operational reality. The invoice is actually your most powerful piece of evidence, but you need to read it through a FinOps lens.
Pull the last three months of invoices and map each line item to a discrete service component: the central dashboard SKU, the legacy SAST module charge, the orchestrator fee, the separate data egress for the reporting database. If it's a single platform, you'd see one charge for "Unified DevSecOps Platform." Instead, you're seeing an aggregated bill for a portfolio. This proves you're paying for integration labor, not an integrated product.
Present the invoice breakdown alongside your screenshot of the three UIs. The financial fragmentation mirrors the architectural fragmentation. Management understands paying for three things but only getting the value of one.
Always check the data transfer costs.
The invoice breakdown is perfect, but take it one step further with the finance team. Ask them to pull the "discount schedule" from the original deal. I've seen those discounts applied per-module, not to the "platform" total.
That proves even the vendor's own finance system treats them as separate products. When legal sees the discount matrix mirrors the three SKUs you're pointing to, they can't argue it's just a billing quirk.
Keep it simple.
That RBAC point hits close to home. If you set up a role in the "unified" admin panel but it doesn't propagate to the legacy SCA module, you've just created a critical shadow admin by accident. I've seen teams open tickets for weeks because the vendor says "just replicate the permission in the other system."
For the compliance angle, check if your audit logs have different source identifiers for actions taken in each UI. If they do, you've got multiple systems of record, which is a nightmare for any regulatory audit. That's a solid, non-engineering risk you can document.
security by default