Giving Every Operating Cost Line a Named Owner

The owner of a cost line is whoever can stop it. Not whoever spends it, not whoever raised the purchase, and not whoever has the word budget in their job title.

That definition does most of the work in this piece, and it disqualifies more people than you would expect. The office administrator who books the courier is not the owner of courier costs. The delivery lead who signs off subcontractor timesheets is not the owner of subcontractor costs if the volume is set by what sales sold. The person who can cancel the contract, decline the renewal, change the specification or stop the activity is the owner. Everybody else is a spender, and spenders do not defend lines.

The other half of the rule is the one that makes finance uncomfortable, so I will say it plainly: finance owning every cost line is the single most common reason variance meetings are useless. It happens for a good reason. Finance is the only function that can see all the lines, so finance builds the budget, and building it looks like owning it. Then the month closes, a line moves, and the controller writes a sentence explaining a department’s spending to that department’s manager. Everyone nods. Nothing changes, because the person who wrote the explanation cannot act on it and the person who can act on it did not write it.

The test, and it has a clock in it

One question decides ownership, and it is deliberately narrow.

Within ninety days, and without asking anyone above them, can this person materially change this line?

Ninety days because a quarter is the shortest interval over which a management decision shows up in the numbers. Without asking anyone above them, because a person who has to escalate is a proposer rather than an owner, and a matrix full of proposers describes an organisation where the owner of everything is the CEO.

Three answers are possible and all three are legitimate results.

Yes. They own it. Their name goes in the matrix, they get the pre-filled submission at budget time, they approve the spend on the route below, and they write the commentary when it moves.

Yes, but only with a decision from above. They own the operation of the line and somebody else owns the level of it. Name both, in two columns. This is the case for most headcount-driven lines, and pretending it is a single ownership is how a manager ends up being asked to explain a variance caused by a hiring decision they argued against.

No. Nobody in the company can change this line within ninety days. Do not manufacture an owner. Mark the line as having no in-year discretion, assign it to the finance lead for accuracy rather than for control, and read variances on it as errors rather than as behaviour. That last point is the practical payoff and it is covered further down.

The matrix

Keyed to account groups rather than to individual accounts, because a chart of accounts at this size has a couple of hundred lines and a matrix with a couple of hundred rows is a matrix nobody maintains. Group the accounts, own the group, and let the detail sit underneath.

The right-hand column is the one people skip and it is the point of the exercise. A line whose commentary duty is unassigned produces commentary written by finance, which is the failure this piece exists to prevent.

Account group	Owner test result	Owner	Second owner, where the level is set elsewhere	Spend route	Who writes the commentary
Salaries and wages, by department	Yes, but	Department head, for hours, overtime and mix	Owner or CEO, for headcount and pay levels	Headcount driven	Department head for the operating part, owner for anything caused by an approved hire or a deferral
Subcontractor and outsourced delivery	Yes	Delivery or operations lead	Sales lead where volume is contracted	Discretionary purchase, per job	Delivery lead, naming the jobs
Commission and variable pay	No in-year discretion	Finance lead, for accuracy	Owner, for the plan design	Formula driven	Finance lead, and the variance should reconcile to revenue or the plan is being applied wrongly
Recruiting, agency fees and onboarding	Yes	Whoever is hiring	Owner, for the approval to hire	Discretionary purchase	The hiring manager
Training and certification	Yes	Department head		Discretionary purchase	Department head
Software and subscriptions	Yes	One named person per application, never a committee		Recurring contract	The application owner, on renewal and on seat count
Telecoms and connectivity	Yes	Operations or office lead		Recurring contract	Office lead
Marketing programmes and events	Yes	Marketing or sales lead		Discretionary purchase	Marketing lead, by programme rather than by month
Travel and client entertainment	Yes	The department head whose people travel		Discretionary purchase, per claim	Department head, naming the trips
Vehicle and fleet running costs	Yes	Operations lead		Mixed, recurring plus usage	Operations lead
Repairs, maintenance and consumables	Yes	Operations or facilities lead		Discretionary purchase	Operations lead
Freight and courier, outbound	Yes, but	Operations lead, for carrier and method	Sales lead, for what was promised to customers	Recurring contract plus usage	Operations lead, with the sales cause named where it is one
Insurance	No in-year discretion	Finance lead	Owner, at renewal only	Recurring contract, annual	Finance lead, and only at renewal
Rent, occupancy and property costs	No in-year discretion	Finance lead	Owner, at lease event only	Recurring contract, long term	Finance lead
Professional fees, recurring	No in-year discretion	Finance lead		Recurring contract	Finance lead
Professional fees, project or dispute driven	Yes, but	Owner		Discretionary, event driven	Owner
Bank charges, merchant fees and interest	No in-year discretion	Finance lead		Formula driven	Finance lead, and a variance here is usually a volume or a rate change, both of which are explainable in one line
Depreciation and amortisation	No in-year discretion	Finance lead		Formula driven	Finance lead, and any variance is a capital timing question rather than a cost one
Bad debt and write-offs	Yes, but	Finance lead, for the provision	Sales lead, for the customers	Formula plus judgement	Both, and the sales half is the one that changes behaviour

Reading the spend route column

Amount thresholds are not in this matrix on purpose. Who may approve what at what size is a separate control and it is already set out in the AP approval and payment run matrix, which is the document that tells your accounts payable process what to do. Repeating a threshold grid here would create a second copy of it, and two copies of an authority table diverge within a year.

What this column adds is the route the money travels, because the route determines when the control is available.

Recurring contract. The decision was made once, possibly years ago, and it renews itself. There is no monthly approval moment. The only control is the renewal, and the only useful discipline is a renewal calendar with the owner’s name and a notice-period date against every contract. If the notice period passes unnoticed, the owner did not own it for another term. Most companies at this size discover on their first pass that they cannot produce a complete list of their recurring commitments.

Discretionary purchase. A decision is made each time, so the approval route applies and the owner is genuinely in control. This is the only route where a monthly variance reflects a monthly choice.

Headcount driven. The line moves because of a hiring decision, a departure, or hours. The department head controls hours and mix. The level was set in the budget. Splitting the commentary along that seam is what stops a variance conversation turning into an argument about a decision that was already taken.

Formula driven. The number is computed from something else: revenue, volumes, a rate, an asset base. There is no discretion at all. The correct response to a variance is to check the formula and the driver, and if both are right, the variance belongs to whatever drove it and should be explained there rather than here.

The collapsing rule

Most companies at this size have between four and eight people who can own anything. That is not a defect and the matrix should not be built as though there were fifteen.

Collapse like this. Two account groups belong to the same row when the same person owns them, the spend arrives by the same route, and the commentary would be triggered by the same kind of event. When all three match, merge them. When any one differs, keep them apart even if the owner is identical, because the split is what makes the commentary specific.

And never split a line to create an owner. If facilities and office costs are genuinely one person’s, do not divide them into six accounts so that six people can each have something. An owner with one line they control is worth more than six people with a sixth of a line each, and the second arrangement makes the budget meeting longer without making it better.

What you are left with in a company of this size is usually something like: the owner or CEO holds the headcount level, the deal-driven professional fees and anything at a lease event; two or three department heads hold their own discretionary and headcount-driven groups; one operations or office lead holds facilities, fleet and consumables; and the finance lead holds everything marked as having no in-year discretion. That is five or six named people covering the whole operating cost base, which is a matrix somebody will actually maintain.

The lines finance genuinely owns, and why labelling them matters

The temptation is to hand the fixed lines to a department head so that every line has an operating owner. Resist it. It produces a manager who is accountable for rent, which teaches everyone that accountability in this company is decorative.

Marking a line as having no in-year discretion buys you something concrete: it changes what a variance on it means. If insurance moves in a month when there was no renewal, that is not spending behaviour. It is a posting error, a prepaid schedule that was not amortised, an invoice hitting a wrong period, or a genuine event nobody told finance about. All four are worth finding, and all four are found faster when the reviewer’s first instinct is “this line should not have moved” rather than “somebody spent more”.

That is also the argument for keeping these lines in the matrix at all rather than leaving them off it. A line with no owner listed reads as an oversight. A line explicitly marked as having no in-year discretion reads as a decision, and the difference shows up the first time somebody new joins the finance function and has to work out which lines to chase.

The commentary duty, which is the part that fails

Ownership that stops at budget time is not ownership. The test of the matrix is what happens at the close.

The rule: the owner writes the sentence, finance edits it for accuracy, and finance does not write it. A controller who writes twelve departmental explanations every month is running a reporting exercise that has been detached from the business. The commentary sounds better and means less, because the person with the knowledge of what actually happened is not the one composing it.

To make that workable, the trigger has to be defined in advance or you get either everything explained or nothing. Set a rule at the start of the year: a line requires commentary when the movement exceeds the threshold the company set for its own management reporting, or when the year to date position has crossed the plan regardless of the month. Both conditions, not one. A line that is slightly over every month is the more dangerous of the two cases and the monthly test alone never catches it.

Two more conditions that keep the practice honest.

The commentary explains a cause, not an amount. This is the same discipline as everywhere else in the reporting pack and it is worth repeating in the budget-owner context because owners find it harder than finance does. “Freight was over because we air-freighted two orders after the plant shutdown” is a cause. “Freight was over by a lot” is a caption.

And the owner names what happens next, or explicitly says nothing happens. Half the value of a variance explanation is the sentence after it. If the cause is going to recur, the owner says so, and the reforecast picks it up. If it was a one-off, the owner says that, and next month’s reviewer knows to check.

Setting it up without a reorganisation

You do not need a new org chart to do this, and if the matrix implies one, the matrix is wrong rather than the company. Work with the people who exist.

1. Print the trial balance at account level for the last completed year, sorted by amount, largest first.
2. Draw the line at the point where the accounts below it add up to something you would not spend a meeting on. Everything below the line is grouped into a single row with the finance lead against it, permanently. Do not apologise for this. A matrix that spends equal effort on the largest and the smallest lines is a matrix that gets abandoned.
3. For everything above the line, apply the ninety day test out loud with the person you think owns it. They will disagree with some of your assignments and they will be right about roughly half of those.
4. Where the answer is “yes, but”, write both names down. Do not resolve it by picking one.
5. Where the answer is no, mark it and move on. Resist the urge to find someone.
6. Circulate the matrix before the budget cycle starts, not with the submissions. An owner who first learns they own a line by receiving a form to fill in has been assigned rather than asked.
7. Re-read it once a year, at the same point in the cycle, and change it when people change roles. A matrix naming somebody who left is worse than no matrix, because it tells the next reader that nobody has looked at this in a while.

What we do not do

I do not build an owner matrix that requires a role your company does not have. If the honest answer is that four people own the whole cost base, the matrix has four owners in it and it is a better document for saying so.

I do not allocate a shared cost across departments in order to give it several owners. Allocation is a reporting choice, and using it to create accountability produces managers arguing about the allocation method instead of about the spending. One owner, the one who can cancel it, and the others are consulted.

And I do not write the departmental variance commentary. Drafting it is quick and it is exactly the thing that makes the whole apparatus pointless, so I will chase an owner for a late sentence for as long as it takes rather than supply one. The controls work is worth having only if the people it names do the parts assigned to them.