The Cut Is Decided. The Hard Part Lands on You
By Robert Bradshaw, Founder and President of WiserOwl
The number usually arrives locked in and from somewhere above the contact center: a staffing reduction target calculated to lower operating costs and free up capital for elsewhere in the business. It lands on your desk as a plan already in motion.
Reduce staffing. You're not asked if it's a good idea. You're tasked to execute it.
And as it settles in, you realize you've been handed not one, but two decisions. The first is who to let go. The second — the one that keeps you awake at night — is how deep staffing can be cut before service breaks and revenue follows it out the door.
Whatever you decide, you'll have to prove it to the people who handed you the number.
You Don't Size a Cut by Guessing
You have a workforce model to help you.
Since 1917, Erlang C and its variations have converted workload into the staff needed; factor in shrinkage and you have your agent requirements. Riding shotgun is average cost per contact, which is used to measure the financial targets.
But contacts have become exceedingly complicated. A single contact today can cross an IVR, touch a bot, and land with agents whose salaries, tenure, and pay loadings look nothing alike, with layers of technology cost woven throughout.
This complexity is what's causing problems with staffing models, because the decisions you have to make rely on answers that an average was never designed to tell you.
It's also throwing your cost-per-contact shotgun rider off the wagon because there's more happening than one rider was ever meant to cover. And while the workforce number is important, the cost savings is the number the business is waiting on.
Who Goes — and Who Decides
According to an April 2026 Gartner survey of 321 service and support leaders, 63% are reducing frontline headcount gradually through attrition. This reactive approach can seem preferable to a proactive one, since a gradual reduction leaves room for course correction.
But an attrition cut is still a cut — it just lets chance decide who goes.
And regardless of the approach, there is a blind spot — cost savings.
In day-to-day operations you measure work from the ground up: activity by activity, contact by contact, agent by agent. You cost that same work from the top down — taking a top-line dollar and allocating it.
But a contact can require several technology solutions on top of more than one live agent, and that top-down average was never designed to provide the exact cost of each resource in that contact. Its job is to fold the paid hour and the productive hour into the same number.
This means the cheap agent may be the expensive one, and the agent who left may have been the one carrying the queue. You won't know which until after they're gone — and even then, you still won't know the actual financial impact.
How Deep is Too Deep
Your workforce model will show service collapsing at each depth of staff reduction, and that side of the model keeps improving — newer Erlang variants, simulation, machine-learning, each one sharpening the same answer: how many people are necessary.
What none of them was built to sharpen is the financial half I mentioned earlier: how much of what the operation pays for actually reaches a customer.
Put these two halves side by side and the depth question splits cleanly.
Your workforce model decides where service breaks. The cost number decides whether the reduction handed to you is one the operation can afford to absorb — because if part of every paid hour never reaches a customer, then there is capacity in the operation that no headcount report shows.
What this means is staffing reduction isn't just about headcount. It's a function of how much of what you're paying for you're actually receiving — and no average, and no ranking, was designed to answer that question.
What Reduction Looked Like When Someone Measured What They Were Getting
A few years ago I worked with a health insurance provider — a B2B contact center. The mandate came down: one-third of the staff had to be reduced.
The contact center leader had all the activity and quality metrics. What they didn't have was a clear view of what the work actually cost — not as an average but built from the ground up, activity by activity, contact by contact, using data the organization already had in the ACD, HR, and payroll. How deep that data goes depends on what’s captured.
The answer that came back made retention decisions fair and transparent to everyone, including senior leadership. With actual cost attached to actual activity, the conversation about who stays started including how much pay was actually reaching customers.
And it answered how deep to cut, because it showed where paid capacity wasn't reaching customers — capacity the operation was funding and not receiving. That, more than anything, is what made a one-third reduction survivable: the contact center knew in advance that the room to absorb it was already inside their operation.
It Just Couldn't Be Seen Until it was Measured
The operation transitioned to one-third fewer agents. Customer satisfaction held. Revenue held. And in the six months that followed, voluntary turnover was close to zero — which, if you've run a contact center through a reduction, you know better than to expect.
A big part was the people who stayed understood why they stayed, and how they were being valued on a level playing field.
I want to be precise about what happened, because it's easy to overclaim. We identified where the costs sat and where the capacity gap was. The leader made the decisions, and the organization realized the outcomes.
In short, the measurement let the person responsible make the call with eyes open, in a way everyone could see and understand.
Staffing is Just One Decision in a Bigger Set
Staffing reduction and retention are big challenges, but they sit on your desk beside other decisions asking the same thing:
- What did the work cost before AI?
- What did the automation actually save?
- What did the deflection initiative actually deflect, and at what cost?
The answer you need in every case is what the specific work actually cost before — and most operations don't have that number. It's nobody's fault. Few organizations had to ask for this level of detail, until now.
But once this question surfaces, the answer has to be built: more than 700 operations, in a non-negotiable order, before any cost is totaled. This is construction, not a spreadsheet.
So, if you think a staffing number may one day land on your desk, the question worth getting ahead of isn't how you'll hit it, or whether natural attrition will spare you the hardest choices. It's whether you know what you're paying for, what you're actually receiving, and what you can safely remove before service, customer experience, and revenue begin to suffer.
You can return to the August 2026 Newsletter here too.