Why the savings number deserves suspicion
Outsourcing savings are calculated against a baseline, and the baseline is negotiable in ways the audience rarely sees. Include the fully loaded cost of every person touching the process, count the systems generously, assume no internal improvement over the contract term, and a thirty per cent saving appears; run the same exercise against an honestly improvable internal operation and much of it evaporates. Providers know this arithmetic better than buyers do, because they run it for a living. The first pressure-test of any outsourcing case is therefore not the provider's price but the credibility of the internal number it is being compared against, and who constructed it.
What the deal team tends not to model
- Transition is staffed by the provider's best people, and the service you experience in month two is not the service you will receive in year two, after the A-team has rotated to the next transition.
- The process knowledge held informally by your longest-serving people leaves precisely when they do, and the deal makes many of them redundant just before the provider needs what they know.
- The retained organisation is designed as a leftover rather than a function: too thin to govern the contract, too senior to do the work, and staffed by whoever was not transferred.
- Everything not written into the service description becomes a change request, and the provider's account team is incentivised to find those gaps: that is where deal margin is recovered.
- SLA dashboards can be green while the business experience deteriorates, because the metrics measure what was easy to specify rather than what the operation actually depends on.
- Insourcing back is priced by nobody, so by year three the provider is negotiating with a customer that has no credible alternative.
The questions that separate a deal from a dependency
Before signature, ask the provider which named individuals from the transition team remain accountable in year two, and what your contractual remedy is when the answer changes. Ask what proportion of their revenue on accounts like yours comes from change requests beyond the base fee, and get the last three accounts' figures if they claim it is low. Ask your own team who will manage this contract day to day, what that function costs, and why that cost is absent from the business case. And ask the exit question in specific terms: what does it cost, in money and months, to bring this operation back in-house or move it to a competitor in year four, because whatever that number is, it is the ceiling on your negotiating power for the life of the contract.
Test the operating model, not just the commercial one
A well-priced deal with a badly designed operating model still fails. Pressure-test the interface: which decisions your people still make, which the provider makes, and what happens when the two disagree at 4pm on a month-end Friday. Pressure-test the improvement clause: providers commit to productivity gains in the contract, but the mechanism that shares those gains with you needs to be enforceable, not aspirational. And pressure-test the people plan on your own side: retained roles need career paths, or the capable people leave and contract governance devolves to whoever could not. Organisations that treat the retained function as a proper design problem, staffed before transition rather than after, have a categorically different experience of the same deal.
What veterans of these deals notice first
Operators who have signed outsourcing contracts, managed them through renewal and unwound one or two read these deals with a distinctive checklist. They look at the ratio of deal-team effort spent on price versus service description, because the description is where the money moves later. They look for the year-three provisions (benchmarking rights, exit assistance, key-personnel clauses) that the deal team traded away for a better headline rate. And they ask the question internal teams avoid: is this outsourcing an operation, or exporting a problem the organisation has not diagnosed? A process that is broken in-house arrives at the provider still broken, but now with a contract wrapped around it and a margin on every fix.
