Escaping the Cost-Centre Story Without Overclaiming

Monday, 09:30, at Bluewater Bank. The budget template has arrived from Gwen, the finance business partner, and the tab for the desk is headed "IT (cost centre 410)". Beneath the numbers is one line in red: "Please justify any increase on last year." Rahul runs the three-person desk for about 260 staff across the head office and four branches. He has a ticketing tool full of evidence and no idea which of it Gwen wants. Last year he wrote "the team is extremely busy and demand continues to grow", which was true, and the increase was declined, which was predictable. He has the same sentence ready this year. He also has, for the first time, a year of clean ticket data, and the sense that it is trying to tell him something Gwen would listen to.

The cost-centre story is the version of IT in which it spends money and produces nothing that can be seen on the same sheet. It is half true, which is why it is so durable. IT does cost. What the sheet cannot show is what the cost buys. It buys the hours of other people's work that resumed because a ticket was fixed. It buys the onboarding that happened on day one, the month-end close that ran, the outage that did not occur. Escaping the story does not mean rebranding the desk as a profit centre, which it is not, or claiming outcomes it cannot prove. It means telling a truer story in the organisation's units, with every number traceable to a record. This article is part of our IT's Internal Brand series, and it is about that story and the discipline that keeps it honest.

It covers what the cost-centre story is and why it is half true. Then come the economics in hours, borrowed from the series that owns the arithmetic. Next are the business outcomes the desk can evidence and the ones it cannot. Then it covers the ally in finance and how to recruit one, the roadmap shown once a quarter, and the rule against overclaiming. A short war story and the value narrative outline to copy finish the article.

What the Cost-Centre Story Is, and Why It Is Half True

A cost centre, in the accountant's sense, is any part of the organisation whose spending is tracked and whose income is not, because it has none that can be attributed directly. The desk, the finance team, the building and the legal department are all cost centres, and the label is not an insult. It becomes one when it is the only thing anyone can see. The budget sheet shows the desk's cost every month to the penny. It shows what that cost bought nowhere, because nobody has written it in a form the sheet can hold.

The desk, when asked to justify itself, usually answers in tickets. "We closed 4,800 tickets last year." The number is accurate and lands badly, because to the reader a ticket is a unit of something going wrong, and more of them is not better. The organisation asks in money and the desk answers in incidents. The two sides are not disagreeing; they are not yet speaking the same language, and the translation is the desk's job, since finance cannot do it from the sheet. The earlier article in this series on why IT has a brand whether it wants one found the cost-centre reputation at the root of most kitchen sentences. This article is about the budget meeting where it is written down.

The half that is true should be said first, and plainly, because a story that begins by denying the cost is not believed about anything else. The desk costs what it costs. The question the narrative answers is what the organisation got for it, what it did not have to pay for, and what the next quarter's spend will buy. Those are three questions finance asks of every cost centre, and most of them cannot answer. The desk can, because it has a ticket for almost everything.

The Economics in Hours

The desk's currency is hours, not money, and the whole narrative is easier once that is accepted. The desk spends technician hours. The organisation loses staff hours waiting for fixes and loses more when something is down. Fixes give hours back; prevention and self-service give hours back without a ticket. Finance can convert hours into money with a rate the desk does not need to know and should not guess at. It will do so more credibly than the desk would. The arithmetic for all of this is owned by The Economics of Support. It covers where the money goes, the cost of waiting and the payback of prevention. This article borrows its results rather than redoing them.

A worked example, from Bluewater and stated as one. The desk's three technicians work about 5,000 hours a year between them. The tool says the desk handled about 4,800 tickets last year. The median wait from a ticket being raised to the requester's work resuming was a little under four working hours, with a long tail. Count only the tickets where someone could not work at all. The desk estimates those at somewhere between one in five and one in four. On that basis, the waiting cost to the bank was somewhere between 3,500 and 5,000 staff hours. That is a range, from a fictional desk, and it is stated with its method, which is the only way a number like it should ever leave the desk. The reasoning behind that particular number, and its traps, is in The Cost of Waiting.

The number is useful not because it is large but because it points at a lever. Suppose the desk can show that the median wait fell from four hours to two across the year. It has given back something between 1,700 and 2,500 staff hours. That is a sentence finance can put on a sheet beside the cost. Equally, if the wait rose because the desk lost a person, the same arithmetic shows what the vacancy cost. That is the honest foundation of a request to fill it. Hours are the desk's evidence. Money is finance's translation. Keep the two jobs separate and both numbers are believed.

The Business Outcomes the Desk Can Evidence

The desk enables outcomes it can prove and outcomes it cannot, and the narrative contains only the first kind. A business outcome, for this purpose, is a thing the organisation wanted to happen, that happened, and that the desk's record shows it had a hand in. The new branch opened on the Monday with every till working; there is a ticket for the branch setup with the dates in it. Thirty-one starters had a working account and laptop on their first morning; the onboarding tickets say so. The month-end close ran without the reporting server failing; the change record shows the disk added the week before. Each of those is an outcome, each is evidenced, and each is a sentence a finance director understands without translation.

The outcomes the desk cannot evidence are the ones it is most tempted to claim. "IT enabled the bank's growth." "Without the desk, nothing would work." "Our self-service portal increased productivity across the organisation." These may all be true in some sense and none of them can be traced to a record. A reader who doubts one of them is entitled to doubt the rest. The table draws the line, with the tempting claim, what the record can actually support, and the sentence that goes in the narrative. The technique of turning tickets into sentences a non-technical reader will act on is in Translating Tickets Into Business Language. The table is that technique applied to the budget meeting.

Notice that the sentences in the third column are smaller than the claims in the first, and more persuasive. A finance director who reads "29 of 31 starters had a working laptop before 09:00" believes it, can check it, and will remember it. The same director who reads "onboarding is now seamless" has read a slogan and will treat the rest of the page accordingly. Smaller, evidenced claims compound into a reputation for telling the truth. The large ones compound into a reputation for slides.

The Ally in Finance

Every budget round is easier with one person in finance who understands what the desk does. That person is recruited by giving them what they want, which is predictability. Finance business partners are judged on whether their numbers surprise anyone. The easiest cost centre on their sheet sends the same one-page narrative every quarter. It forecasts its replacement cycle a year out and says what it will ask for before it asks. It never presents a number it cannot source. Gwen, at Bluewater, had never been offered any of that by IT. She had been offered "extremely busy", once a year, in the week she was busiest.

Recruit the ally by asking, rather than presenting. Book twenty minutes with the finance partner outside the budget round and ask three things. How would they like the desk's value shown? What in the current numbers can they not explain to their own director? What do they wish they had known earlier in the last two years? The answers are the narrative's outline. Then give them the hours arithmetic and let them do the money, in their rate and their format. A finance partner who built the money number is its defender in the room. One who was handed it is its auditor.

Give the ally a forecast, because it is the thing they can least get from anyone else. Make a one-page list of the laptop replacement cycle, the licence renewals by month, and the two systems that will need replacing within three years. Give roughly what each will cost in technician time. Update the list each quarter and date it. Finance partners keep this kind of list for every other department from invoices and guesswork. A desk that hands them one with its own numbers in it has moved, in their eyes, from a cost that happens to a cost that is managed. That is most of the escape from the story, and it was done with a list.

The Roadmap Shown Once a Quarter

The quarterly roadmap is one page that says what the desk will do in the next three months. It gives what each item will cost in technician hours and in money, and what each one removes or enables. It is shown to the same people every quarter. It sits beside the quarterly review of trends, which has its own shape in The Quarterly Review: Trends, Not Months. It looks forward where the review looks back. A desk that only reports what happened is reacting. A desk that also says what it will do next, and then does it, is planning, and planning is what the organisation pays other departments for.

Keep it to five or six items, each with a month, an owner, a cost in hours and a line saying what it buys. For example: "retire the old file server by end of quarter; about 40 technician hours; removes the system behind a third of our slow-document tickets". Items that touch people's working day get a note that a change notice will come. Items that are requests for money or people get a pointer to the case being made separately. The case itself, with its evidence and its pushback, is the subject of The Headcount Case. The roadmap only says that it exists and when it will be made.

Then, next quarter, show the same page with each item marked done, slipped or dropped, and the reason, before showing the new one. This is the step most desks skip and the one that does the work. A roadmap that is never reconciled is a wish list. A roadmap reconciled four times a year is a track record. The finance partner can see from it whether the desk's estimates are to be trusted. (They will be wrong at first. Say so, and show them getting better.)

Never Overclaiming

The rule is that every number in the narrative traces to a ticket, a change record, a timesheet or an invoice. It is stated as a range where the record is uncertain. It is one the desk would say straight to the finance director's face without a slide to hide behind. The reason is practical rather than moral. The first number that is challenged and cannot be sourced discredits every number on the page, including the true ones. The desk gets one such moment per finance partner. The arithmetic discipline is the one used for self-service in Deflection Math Without Fooling Yourself: assumptions stated, ranges not points, the pessimistic end shown beside the hopeful one.

The overclaims below are the ones that appear most often in IT value presentations, and each is cut from the narrative on sight. They are listed as a checklist because the temptation returns every budget round, usually at about eleven at night.

• A total saving in money that the desk calculated itself, rather than hours that finance converted.
• A return-on-investment percentage of any kind, from any source.
• "Revenue enabled", "growth supported" or any outcome the desk's record does not show it touched.
• "Countless", "seamless", "significant", "dramatic", or any adjective standing where a number should be.
• A prevented-outage value with no record of what the same failure cost before.
• A point estimate where the record supports only a range.
• Any figure borrowed from a vendor, an industry report or another organisation's desk.
• A comparison with last year where the counting rules changed in between and the change is not stated.

Two tests catch most of what the list misses. First, could a sceptical reader reproduce the number from the record with the method stated on the page? If the method is not on the page, add it or cut the number. Second, would the desk lead say the sentence aloud to the finance director across a desk, with no slide, and defend it? If the honest answer is "I'd soften it", soften it on the page. A narrative that passes both tests is shorter and duller than the one the desk wanted to write. It is also the only one that works twice.

What Escaping the Story Looks Like

Escaping the story means the desk's line on the budget sheet has a second column that finance wrote. It also means the finance partner can explain the desk to their own director without IT in the room. It is not escaped by a presentation, however good, and it is not escaped in one budget round. It takes four quarterly narratives, four reconciled roadmaps and one forecast that turned out to be about right. That is a year, and most of the year is spent being boringly accurate. Boringly accurate is the brand. It is a better one than "extremely busy".

The narrative also feeds back into the rest of this series. The hours returned by self-service and prevention are what the monthly note reports. The outcomes evidenced from tickets are what the kitchen eventually repeats. The forecast is what makes the next request for a person land as planning rather than pleading. If you run QuerySolv, the standard reports on volume, backlog by age and response and resolution times give the narrative its ticket figures directly. The event history on each ticket is where the dates for an outcome such as the branch opening are found. Any tool with the same reports and history does the same. If yours has neither, a monthly export to a spreadsheet is enough to start. The narrative does not need a better tool. It needs the same honest page, four times a year, until the sheet has two columns.

A Short War Story: The Four-Hundred-Percent Slide at Oakfield Transit

Marcus ran the desk at Oakfield Transit, a bus operator with about 300 staff. In his second budget round, he presented a slide saying that the self-service portal had delivered a return of four hundred percent. He had built the figure the night before. He multiplied an estimate of deflected tickets by an estimate of minutes per ticket by a rate he had guessed at. Then he divided by the licence cost. Jo, the finance partner, asked in the meeting how the number had been calculated. Marcus had the spreadsheet but not the sources. He could not say which of the four inputs came from a record, because none of them did. The slide was withdrawn, politely. Six months later, the desk's case for a second technician, which was well-evidenced from the ticket history, was deferred for a year. A note from the finance director asked that "the IT numbers be reviewed by finance before submission". That took the case out of Marcus's hands and added a quarter to it. The year without the second technician cost the desk a backlog that peaked at about 190 open tickets. It cost Marcus, he said afterwards, most of his credibility with the one person who could have been his ally. One invented number does not cost you one argument. It costs you every number you present afterwards, for as long as the person who caught it is in the room.

The Value Narrative Outline

The outline below is the one-page narrative, in the order a finance reader expects it. The cost is stated first, then what it bought, then what did not happen, then what was given back. Next comes what next quarter buys, then the ask if there is one. At the bottom is the source of every number on the page. Fill it from the ticket history, the change records and the hours arithmetic, in ranges, once a quarter. The last section is not optional; it is what turns the page from a presentation into evidence.

THE VALUE NARRATIVE — ONE PAGE, ONCE A QUARTER
Desk: ______________   Quarter: ____________   Written by: ______________   Finance partner: ______________

1. WHAT THE DESK COST (finance's number, stated first and plainly)
   Budget line this quarter: per the finance sheet.   Technician hours worked: about ______

2. WHAT THE ORGANISATION GOT (tickets, in the organisation's units)
   Requests handled: about ______ .  Median time from request to work resumed: about ______ hours.
   Outcomes with a record (one line each; ticket or change reference beside it):
   - ______________________________________________________  ref: ______
   - ______________________________________________________  ref: ______
   - ______________________________________________________  ref: ______

3. WHAT DID NOT HAPPEN (prevented work, with what the same failure cost last time)
   - ______________________________ done on ______ . Last time this failed: ______________ (ref: ______)
   - ______________________________ done on ______ . Last time this failed: ______________ (ref: ______)
   - ______________________________ done on ______ . No previous occurrence on record.

4. WHAT WAS GIVEN BACK (hours, in ranges; finance converts to money)
   Staff hours returned by faster fixes: between ______ and ______ (method: ______________________)
   Technician hours returned by self-service / prevention: between ______ and ______ (method: ______________)
   Where these hours went instead: ______________________________________

5. WHAT NEXT QUARTER BUYS (the roadmap, reconciled against last quarter's first)
   Last quarter: ____ items done, ____ slipped, ____ dropped. Reasons: ______________________
   Next quarter:
   - ______________________ by ______ ; about ______ technician hours; removes / enables: ______________
   - ______________________ by ______ ; about ______ technician hours; removes / enables: ______________
   - ______________________ by ______ ; about ______ technician hours; removes / enables: ______________

6. THE ASK, IF ANY (one line; the full case is a separate document)
   ______________________________________________________  Case attached: [ ]  Evidence from: ______

7. WHERE EVERY NUMBER CAME FROM
   Ticket history export dated ______ ; change records ______ ; hours from ______ ; invoices ______ .
   Counting rules unchanged since: ______ (or: changed on ______ ; comparisons adjusted: [ ])
   Numbers finance has reviewed: [ ]   Numbers stated as ranges where uncertain: [ ]

Rahul sent Gwen the first narrative in place of the sentence about being busy. It had the waiting-cost range in it, the branch opening and the onboarding count. A line at the bottom said which export every number came from. She replied with two corrections to his hours arithmetic, both of which made the number smaller and more defensible. She also asked a question about the replacement cycle. It was the first time finance had asked IT about the future. The increase was approved in part, which was more than the previous two years combined. The desk was still cost centre 410. It had simply acquired a second column.

Frequently Asked Questions