Somewhere in your last filing there is a sentence announcing a cost savings initiative, usually with a number and a period attached to it. It was reviewed by finance, agreed with the chief executive, and read by every analyst who covers the company. Everyone treated it as a commitment about spending.
It is also a commitment about work. A cost that is going to be lower next year is a piece of work that has to be done differently, by someone, starting on a date. The filing names the amount. It does not name the mechanism, and in most companies of this size nobody in the reporting chain is responsible for producing one.
Over the trailing twelve months, SEC full-text search returns 239 filings describing a cost savings initiative. Those are filings rather than companies, so the count of distinct issuers is lower. What the population has in common is more useful than its size: a public number, a public period, and an internal organization that was already fully occupied before the number was announced.
A Target Is Not A Mechanism
The target arrives at the operating level already divided. Finance allocates the announced figure across functions, and each function receives a share it now owns. Three kinds of answer come back.
- Headcount. Positions are removed or left unfilled. This is countable on the day it happens, which is why it is always the first answer.
- Procurement. Contracts are renegotiated, vendors consolidated, discounts taken. Also countable, and largely exhausted after the first cycle.
- Efficiency. The same work, done with fewer hours. This is where the remaining balance is assigned, and it is the only one of the three with no standard method behind it.
The third category is where the initiative usually stalls. It is written into the plan as a percentage rather than as a change to a specific process, because the person who owns the number does not own a way to change how the work is performed. They own a budget line and a team that is already busy.
A saving assigned to "efficiency" with no named process change is not a plan. It is the amount by which the other two answers fell short.
Why The Cost Comes Back
There is a well-established asymmetry in how costs move, and it is worth knowing before you commit to a period. Selling, general and administrative costs do not fall as fast as they rise (Anderson, Banker & Janakiraman, Journal of Accounting Research, 2003): across 7,629 firms over twenty years, selling, general and administrative costs rose an average of 0.55% for each 1% increase in sales, and fell only 0.35% for each 1% decrease. The wider review of asymmetric cost behavior (Banker & Byzalov, Journal of Management Accounting Research, 2014) attributes the asymmetry to deliberate resource-commitment decisions rather than to accounting artifacts. Capacity is expensive to remove and expensive to rebuild, so managers hold it.
The pressure of a published target changes that behavior in a specific direction. Earnings targets and managerial incentives affect the stickiness of costs (Kama & Weiss, Journal of Accounting Research, 2012): when managers face incentives to avoid a loss, avoid an earnings decrease, or meet an analyst forecast, they bring forward the reduction of slack resources rather than letting it follow activity. The reported cost falls in the period being reported. Whether the work falls with it is a separate question, and the filing does not ask it.
What happens next is familiar to anyone who has run operations through one of these cycles. The removed hours reappear as overtime, as contractors, as a backlog that becomes a quality problem two quarters later, or as a senior person doing junior work because the junior position was the one that was cut. The evidence on whether workforce reductions improve performance is, in the authors' own word, equivocal: an examination of downsizing across industries (Guthrie & Datta, Organization Science, 2008) finds the effect conditional on the industry the firm operates in rather than reliably positive. A reduction is a change in who does the work. It is not, by itself, a change in what the work is.
Where The Savings Actually Are
In a company between sixty and two hundred million dollars in revenue, the recoverable hours are rarely in the work itself. They are in the coordination around the work: finding the current version of something, asking a colleague for a status, re-entering the same values into a second system, checking that an approval that was promised in an email actually happened, and rebuilding a report because the source moved.
None of that appears in a budget. It appears as fully occupied people who cannot take on anything more, which is exactly the condition the initiative is supposed to relieve.
| What the savings line says | What has to change for it to be true |
|---|---|
| Reduce contractor spend in operations | The process stops needing a person to chase it |
| Improve productivity in the shared service center | The same request stops being entered twice |
| Consolidate reporting effort | The numbers come from one place that is always current |
| Reduce overtime at period close | The steps that must happen in order actually happen in order |
Each right-hand cell is a piece of software, and each one is small. That is the part which is consistently misjudged. The instinct of a finance or operations leader who has lived through an enterprise system implementation is that any software answer means a year and a seven-figure budget, so the answer is refused before it is priced. The work that sits around the enterprise system is a different size of problem entirely, and it is the size that most savings targets are actually made of.
Make The Saving Measurable Before You Claim It
A saving you cannot count is a saving you cannot report, and most process-level savings are uncountable in the state the process is in today. If the work happens in a spreadsheet, an email thread and a weekly meeting, there is no record of how many items were handled, how long each one took, how many were reworked, or how many are waiting right now. The initiative therefore gets reported using the only measure available, which is headcount, which returns you to the first answer.
The order that works is the reverse of the usual one. Instrument the process first, so the current cost is a measured number rather than an estimate. Then change the step. Then report the difference from the same instrument, in a form somebody outside the department can check.
This has a second benefit that finance leaders tend to value more than the saving. A process that records its own volume, timing and exceptions is a process you can say the status of without asking a person, which means the savings claim in next year's filing rests on a system rather than on a manager's assurance.
First Steps
- Take the savings figure and write the mechanism next to each part of it. Headcount, procurement, or a named process that changes. Any amount that cannot be assigned to one of the three is unallocated, whatever the plan calls it.
- For the largest process-change amount, ask the person who runs that process how they would count today's cost. If the answer involves an estimate or a sample rather than a record, the saving cannot be evidenced later either.
- Write down the single step in that process where work waits for a person. Waiting is where the recoverable hours concentrate, and it is the step a small system removes most reliably.
Price The Mechanism Before The Period Ends
A published savings number has a period attached to it, and the period is the part that cannot be renegotiated quietly. The gap between a target and a mechanism is normally one conversation with somebody who builds this kind of system, followed by a few weeks of work.
A fifteen-minute call — nothing paid, nothing signed — takes one process from your savings plan apart: where the hours currently go, which of them a system can remove, what would have to be built, and what it costs. If the honest answer is that the hours are already thin and the saving is not there, that is what we will say, and you will have avoided spending against it.
Where something does have to be built, that work runs as a monthly engineering partnership: one prioritized delivery lane, with the measures each release has to meet written down before it starts. The work runs in a repository we own, and paid-for deliverables transfer to you monthly, with the documentation and run guide. After that, $15,000 a month puts one accountable person on keeping it running, with a written monthly report of what ran and what changed. That report is also the evidence your savings claim needs, produced continuously instead of assembled at the end.
References
- Anderson, M. C., Banker, R. D., & Janakiraman, S. N. Are Selling, General, and Administrative Costs "Sticky"? Journal of Accounting Research, 2003.
- Banker, R. D., & Byzalov, D. Asymmetric Cost Behavior. Journal of Management Accounting Research, 2014.
- Kama, I., & Weiss, D. Do Earnings Targets and Managerial Incentives Affect Sticky Costs? Journal of Accounting Research, 2012.
- Guthrie, J. P., & Datta, D. K. Dumb and Dumber: The Impact of Downsizing on Firm Performance as Moderated by Industry Conditions. Organization Science, 2008.
- U.S. Securities and Exchange Commission. Form 8-K. Item 2.05 covers costs associated with exit or disposal activities.
- U.S. Securities and Exchange Commission. EDGAR Full-Text Search. Filing counts cited are from full-text search over the trailing twelve months.
