The cost target was clear. The path to it wasn’t.

At a glance

Organization

PE-backed, multi-site services company, $950M revenue.

Situation

Sponsor mandate of $25M annual run rate from a $210M controllable cost base — equal to 11.9% of the identified base.

Complication

Dozens of savings ideas mixing structural cost elimination with productivity, deferred hiring, vendor assumptions, technology dependencies and discretionary reductions — several claiming the same benefit.

Entry point

Zero-Based Transformation™ across the full lifecycle, GET through PROVE.

Result

$21.8M modeled annualized structural run-rate reduction — 10.4% of the cost base — with $18.1M modeled as realized in year one against $9.6M of one-time cost to achieve.

A clear mandate and a long list

The sponsor had done its part precisely. The target was specific, the cost base was identified, and the timeframe was stated. There was no ambiguity about what had to happen or who owned it.

Management responded the way most management teams respond: it collected ideas. Function leads submitted what they believed could be removed, the finance team aggregated the submissions, and within a few weeks the total exceeded the target. On paper the mandate was already met.

The list was not padding. Most items were genuine opportunities identified by people who understood their own cost lines. The problem was in the aggregate rather than the parts.

The list mixed things that are not alike

Read line by line, the savings pipeline contained at least seven different kinds of claim, each with a different probability of reaching the P&L and a different chance of staying there:

• Actual cost elimination — work that would stop, roles that would not exist, spend that would end.
• Productivity — the same work performed with less effort, which only becomes a saving if the released capacity is removed or redeployed.
• Delayed hiring — real cash in the year it happens, and usually reversed in the next.
• Vendor savings — some contractually secured, some anticipated from negotiations not yet held.
• Technology assumptions — savings contingent on a system change that was itself unfunded.
• Organizational restructuring — structurally sound, but with severance, transition and retained-cost effects not netted.
• Discretionary-spend reductions — the easiest to report and the fastest to return.

Several items also overlapped. A vendor reduction and a workflow redesign could both claim the same underlying cost, and both were counted. Aggregated without that distinction, the pipeline could total whatever number was asked of it.

What operating model can run the business at the required cost — and which savings can actually reach the P&L?

That is a different question from where can we cut, and it is the question a sponsor is really asking. A cost program that hits its number for four quarters and then reverses has not changed the cost structure. It has borrowed against it.

Why a savings pipeline totals to any number

The conventional approach is to manage the pipeline: track each idea to a stage, apply a confidence weighting, and report the weighted total against target. It is disciplined project management applied to a set that was never validated.

Weighting does not help when the categories are incommensurable. A 70% confidence on a contractually secured vendor reduction and a 70% confidence on an unfunded technology dependency are not the same 70%. Nor does pipeline management detect double counting, because each item is tracked in isolation by the function that proposed it.

The deeper limitation is that a pipeline says nothing about whether the business can still operate at the reduced cost. Savings are removed from cost lines. Work continues to arrive.

Rebuild the cost structure around what must still be done

Zero-based logic refuses the incremental baseline — here, the assumption that next year’s cost structure is this year’s minus a percentage. The sequence starts from the economic requirement and rebuilds toward it.

First, the economic requirement

The $25M target was restated as what the business would have to look like to operate at that cost: the volume it must carry, the service levels it must hold, and the obligations it cannot shed.

Second, the work and capabilities that must remain

Read against that requirement, the question is not which costs can be cut but which work the business must still be able to do. That distinction is what prevents a cost program from removing capacity the business will have to rebuild at a premium within a year.

Third, the rebuild

The cost structure was then rebuilt around the retained requirement, across four levers: organization design, external spend, workflow and productivity, and the technology estate. Each was sized against the rebuilt structure rather than against an idea list.

Fourth, structural separated from temporary

Every claim was classified as structural or temporary and reported as such. Deferred hiring and discretionary reductions were not disqualified — they were simply not allowed to count toward the run-rate figure, because they do not persist.

Fifth, ownership and evidence

Each major claim was assigned to a named owner and driven toward financial evidence rather than a reported intention. A saving is not a saving until Finance can see it in a cost line.

Where the structural reduction came from

The overlap test

The four levers were tested against one another for double counting, because a savings pipeline of this shape almost always contains it. The 62 roles in the organization lever and the 41 released by workflow redesign are distinct positions in distinct functions — the organization lever removes management layers and spans, the workflow lever removes effort from transactional processing. Vendor reductions were tested against the technology lever so that a rationalized application and its retired support contract were not both claimed, and the workflow lever excludes any productivity gain enabled by a system change already counted under technology.

Why $18.1M rather than $21.8M in year one

The $3.7M difference is phasing, not shortfall, and it reconciles by lever: the organization actions carried a consultation period and landed with roughly nine months of in-year effect ($1.8M outstanding); the vendor reset followed contract renewal dates falling in the third and fourth quarters ($1.1M outstanding); and the technology rationalization completed after a parallel-run period ($0.8M outstanding). Each of those actions was complete and in the run rate at year end, which is what makes the $21.8M an annualized figure rather than a forecast.

Modeled inputs. Role counts, fully loaded compensation, addressable spend and removable percentages are modeled for this case and stated so that each lever is reproducible from its denominator. The residual against the sponsor target is not a miss to be explained away — it is the difference between a number that was set and a number that survived execution and evidence.

What was decided

The sponsor was told, early, that $25M of structural run rate was not available at acceptable risk to the operating capability — and that $21.8M was. That conversation is the reason the number held.

The $9.6M cost to achieve was approved alongside the target rather than absorbed quietly during delivery, which is the other reason. A restructuring funded without its severance, migration and duplication costs will either miss the run rate or find the money by not doing the structural part.

Temporary actions continued where they were sensible, and were reported separately as timing benefit rather than run rate. Nothing was hidden; the two categories were simply not allowed to blend into one total.

Each lever carried a named owner with an evidence path to Finance. Where a saving depended on an unfunded technology change, either the dependency was funded or the saving was removed from the bridge.

The results

Modeled annualized structural run-rate reduction is $21.8M, or 87.2% of the sponsor target and 10.4% of the identified cost base. Year-one realized value is modeled at $18.1M, or 72.4% of target — the $3.7M gap being phasing, reconciled lever by lever above rather than asserted.

Against that, one-time cost to achieve was $9.6M, giving a net year-one cash effect of $8.5M and a payback of roughly five months on the annualized run rate. A cost program reported without its cost to achieve is not a cash number, and a sponsor will ask for it.

A cost program that had reported $25M against the same base would have looked better for three quarters. The difference is that this one was still delivering in the fourth.

Why the result held

Structural savings were separated from deferrals at the outset, so the run rate did not quietly reverse in the following year. Nothing counted toward the structural figure that depended on a decision being deferred rather than made.

The business could also still operate. Because the rebuild started from the work that had to remain, capacity was not removed and then repurchased at contractor rates — which is the most common way a cost program returns its own savings.

What this means if you carry a cost mandate

If the target has been set and the path to it is a list of ideas with confidence weightings, the honest number is not yet known — and the sponsor is being asked to fund an assumption.

Evidence basis

The documented pattern is that value leaks throughout delivery and beyond it, and that the programs which hold their value embed the discipline into normal operations.

• McKinsey, Losing from day one (December 2021 global survey, n=1,034). Reports that 55% of transformation value loss occurs during and after implementation, while successful programs capture value faster and embed transformation disciplines into normal operating processes. Cited as context.