The Quarterly Review: Trends, Not Months

Thursday, 10:15, at Bluewater Bank. Marcus, who runs the five-person desk, has sent a one-page monthly report for twelve months without missing one. He is proud of that, reasonably. In the quarterly leadership meeting Yusuf, the chief operating officer, has a folder open with all twelve pages in it. "These are good," he says. "I read them. I have a question I can't answer from them, though. Are we better or worse than we were a year ago?" Marcus has twelve answers to that question, one per page, and no answer. Each page said what changed that month. None of them said what the twelve months added up to, because none of them was asked to. Twelve photographs, taken a month apart, and nobody had stood them in a row.

The quarterly review is one page, once a quarter, that answers three questions over a longer window than a month. It asks what changed over the quarter, as a trend rather than an event. It asks what the desk did about it, with the number each action moved. It asks what the desk needs next, with a horizon longer than the next report. It is the monthly page's older sibling. The monthly page asks "is there anything to decide this month?" The quarterly page asks "are we getting better, and what would make us better faster?" The second question cannot be answered by any single month. That is because a month is weather and a quarter is the beginning of climate. This article is part of our Reporting to Leadership series. It is about the page that stands the photographs in a row.

It covers why a quarter is the right window and the three trends worth a page. It covers reading a trend honestly; what the desk did, with results; and what the desk needs next. It then covers the review meeting itself; the fourth quarter as the year; a short war story; and the quarterly review page to copy, in that order.

Why a Quarter

A month is too short to see a direction and a year is too long to act on one. A quarter is the window in between where both are possible. A help desk's numbers move with the organisation's calendar: a quiet August, a month-end spike, the week a new system arrives, the start of a term or a financial year. Any two months can differ for reasons that have nothing to do with the desk. A monthly page that reads too much into one movement will be contradicted by the next. Three months smooth the worst of that. Twelve months drawn, with the most recent three discussed, shows the direction and the season at once. The method for separating them is the subject of Trends and Seasonality.

A quarter is also long enough for an action to have a result. An article published in the first week of a quarter has two or three months of after-period by the review. A clear-out of old tickets has had time to show whether the backlog stayed down or rebuilt. A change to how tickets are routed has had time to settle. The monthly page can say "we did this"; only the quarterly page can say "we did this, and here is what happened". That sentence, action and result in one breath, is the most persuasive thing a desk can put in front of leadership. It needs a quarter to exist.

The quarter matches leadership's own rhythm, which is the practical reason. Most organisations review something quarterly: budgets, performance, risk. A desk whose page arrives in the same week as those reviews is read as part of running the organisation rather than as a report from a department. Find out when the quarterly meeting is and send the page the week before. A page that arrives on its own, in a month when nobody is thinking in quarters, is a monthly page with more history and gets read as one.

Twelve monthly pages do not add up to a quarterly review, which is why Marcus could not answer Yusuf. Each monthly page compares one month with the one before. A desk can drift a long way in twelve small steps, each of which was "a bit up". The quarterly page is where the steps are drawn as a line. A line is a thing a reader can see the direction of without arithmetic. The quarterly page is also where the monthly spreadsheet earns its keep, because the twelve rows are already there, waiting to be drawn. (If there is no spreadsheet, the first quarterly review is the afternoon you start one.)

The Three Trends Worth a Page

Three trends fit on a page, and they are the three directions leadership cares about. Is the desk busier, is it slower, and are people happier with it? The first is demand: tickets created and closed per month over twelve months, with the change in the organisation that explains any change in the line. Bluewater is the fictional desk this article uses for its figures. Created ran from about 350 a month at the start of the year to about 420 at the end. The explanation was a branch merger that added about seventy staff in the second quarter. A demand trend without its cause is a worry. With its cause it is a fact about the organisation, which is what it was all along.

The second is delivery: how long people wait, as the median time to resolve each month. This is paired with the count of tickets older than thirty days at each quarter end. The two are read together because they fail in opposite directions. A desk can keep its median down by working the easy tickets and letting the hard ones rot. The thirty-day count catches it. A desk can clear its old tickets by stopping everything else, and the median catches that. Bluewater's median ran 5.2 hours in the first quarter, 6.8 in the second when the merger arrived, 6.1 in the third and 5.4 in the fourth. The thirty-day count ran nine, twenty-seven, six and eight. The two lines together say the merger hurt, the desk recovered, and the old-ticket spike was cleared and stayed cleared.

The third is experience: satisfaction as good ratings over replies, each month, with the reopen count as its check. Satisfaction alone can rise because the furious stopped replying. Reopens alone can fall because the desk stopped confirming fixes. Together they say whether people got what they came for and whether it stayed fixed. Bluewater's satisfaction ran 48 good of 55 replies in the first quarter's average month, then 50 of 58, 56 of 61 and 58 of 60. Reopens were falling from about fifteen a month to about nine. That is a trend worth a sentence: "people are happier with the desk than a year ago and fixes are holding better, and both improvements followed the confirmation step we added in the spring".

Three, not six, because the page has room for three charts and a reader has attention for three directions. The three charts are the three described in Charts That Help and Charts That Hide: created against closed, the backlog by age, and the trend line. They are drawn here over twelve months with the dates on the axis and a sentence under each. The quarterly review is the one page in the year where charts earn their space, because the question is direction and a direction is a shape. Everything else the desk tracks stays on the scorecard. If one of those numbers is the quarter's story, it gets a sentence, not a fourth chart.

Reading a Trend Honestly

A trend is twelve points, not three. The first honest thing a quarterly page does is refuse to call anything a trend before it has the points. In the desk's first year the quarterly review draws the bars it has and says in plain words that the line will appear when there is a year to draw it. The second honest thing is to say what season the quarter sits in. A fall in demand in a quarter that contains the summer is a season until it has been seen in a quarter that does not. The same three months a year earlier are the comparison that settles it. The template has a column for them.

Tell a step from a drift, because they have different causes. A step is a line that was flat, jumped in one month, and stayed at the new level. Examples are a merger, a new system, or a change in how tickets are counted. A drift is a line that rose a little every month with no single jump: a desk slowly losing capacity to a growing organisation, or a process slowly silting up. A step has a date, and the page names what happened on it. A drift has no date. That is what makes it dangerous and what makes the quarterly page the only place it will ever be seen. That is because no monthly comparison is large enough to notice it.

Never claim credit for a movement the desk did not cause, and say so when the cause is not known. Demand fell because a department left. The median fell because the supplier finally answered forty tickets at once. Satisfaction rose because the survey was moved onto the resolving reply and more people answered it. Each of those is a true explanation that gives the desk no credit. A quarterly page that offers it is believed when it does claim credit the following quarter. The cautious form, "the median fell by about an hour; the supplier's catch-up accounts for most of it and the new routing for some", is the form that survives a reader who was in the room when the supplier caught up.

Where a number needs smoothing, say how. A three-month rolling average, the average of each month with the two before it, takes the jaggedness out of a demand line without hiding the direction. The page says "rolling three-month average" in the subtitle. A rolling average on a chart that does not say so is a chart that has been quietly improved. Smoothing is honest when it is labelled and a small deceit when it is not, and the difference is four words.

What the Desk Did, With Results

The middle of the page is the desk's actions over the quarter, three to five of them. Each comes with the number it was meant to move and what the number did. This is the section that turns the quarterly review from a weather report into a record of work. "Published the remote-access article in week two; remote-access tickets fell from about fifty a month to about twenty-five over the quarter." "Cleared the twenty-seven tickets over thirty days with five days of Lena's time in month one; the count at quarter end is six, and the supplier's queue is now reported monthly." "Added a confirmation step before closing; reopens fell from about fifteen a month to about nine." Action, number before, number after. Nothing else.

Report the action that did not work with the same shape and the same calm. At Bluewater, Marcus had announced self-service password reset in the first month of the quarter and expected password tickets to fall. They did not, because the directory team's rollout slipped to the following quarter. The announcement reached people before the feature did. The line on the page read: "Announced self-service password reset; password tickets unchanged at about sixty a month; the reset feature itself arrives next quarter, and the announcement will be repeated when it does." A page with one honest failure on it is believed about its three successes. A page with four successes and no failures is a brochure, and leadership has read brochures before.

Where an action killed a recurring cause of tickets rather than treating the tickets, say which cause and what it had been costing. That is the kind of result leadership remembers. The quarterly hunt for those causes, the top-five list and the one source killed each quarter, is the subject of The Quarterly Hotspot Review. Its record of what was killed and what it saved is the same section of the same page written from the other end. One line per kill: "retired the eleven oldest laptops; laptop faults from that group fell from about twenty a month to none."

Keep the actions to those with a number attached. The desk did a hundred things in the quarter, and ninety-five of them were the job. The page lists the five that changed a number, because those are the ones that answer "what did the desk do with the quarter?" in a way a reader can check. A list of everything the desk did is a timesheet, and the quarterly page is not where the desk proves it was busy. The demand chart already did that.

What the Desk Needs Next

The quarterly ask has a longer horizon than the monthly one and connects to a trend rather than a month. A monthly ask is "five days of Lena's time next month". A quarterly ask might be "a protected week each quarter for the two technicians to kill the quarter's top hotspot". Or it might be "a conversation between you and the housing supplier's account director, because the desk's contact has had a quarter and the thirty-day count is theirs". Or it might be "agreement in principle to a hire, with the case to follow". The ask names the trend it would move and by roughly how much, as a range. It is one ask, because a page with three asks has a reader choosing between them and usually choosing none.

When the ask is a hire, the quarterly page carries the headline and the case goes on its own page. The case is built from the capacity arithmetic in The Headcount Case: hours available against hours needed, backlog growth, response decay, and what one more person changes. The quarterly review is the right place to first say the word, because the demand trend is the evidence and the page is where the trend is drawn. It is the wrong place to make the argument, because the argument is a page of its own and the review has three other jobs. "Demand is up about a fifth on a year ago with the same five people; I will bring the capacity case to the next review" is the quarterly sentence. The next review has the case attached.

The page also says what the desk is carrying: the risks and waits older than the quarter that have no decision attached yet. The backup fault open since the spring. The nine tickets waiting on the supplier. The server that cannot be patched until the application vendor answers. These are not asks. They are things leadership should know the desk is holding. Each carries a date it has been held since and the name of whoever outside the desk owns the next step. A carried risk that appears on three quarterly pages in a row becomes an ask on the fourth. The three pages are the evidence that the desk gave the owner every chance.

The last block is next quarter's plan: the three things the desk will change, and the number each is meant to move. That way, the next review can report against it. "Publish articles for the top three request types; expect those types to fall by a quarter to a half." "Move the survey onto the resolving reply; expect replies to rise." Writing the expected movement down in advance is the thing that makes the next quarter's results section honest. A result compared with a prediction is a finding and a result compared with nothing is a story.

The Review Meeting

The page is sent a week before the meeting and the meeting is fifteen minutes. The two facts depend on each other. A reader who has had the page for a week arrives with questions, and fifteen minutes is enough for questions. A reader who sees the page for the first time in the meeting needs the fifteen minutes to read it and asks nothing. Open by reading the three trend sentences aloud, one each, in under a minute. Then stop and let the questions come. The questions are the meeting. The page was the preparation for them.

Expect three questions and have the answers in your pocket. "Why did that go up?" wants the cause behind the trend, in the organisation's words. That is the translation work in Translating Tickets Into Business Language. "What would it take to fix it?" wants the ask, which is already on the page, and the answer is to point at it. "How does that compare to other places?" is the question with no honest answer. The honest reply is that the desk's own year-ago figures are the only comparison it can vouch for. That is because other organisations count differently and the desk cannot see how. Say that plainly. It is more credible than a borrowed number, and it will be remembered as such when a borrowed number from somewhere else falls apart in the same meeting.

Leave with the ask answered or dated. "Yes", "no", or "bring me the case by the next review" are all acceptable outcomes. "We'll think about it" is not. The polite way to close it is "shall I put it on the next quarter's page as still open?" That makes the carried item visible and usually produces a decision in the corridor afterwards. Write the outcome into the ticket system or the desk's own notes that afternoon. The next quarterly page reports what was decided and what happened since. A decision nobody wrote down is a decision that will be remembered differently by everyone in the room.

The Fourth Quarter Is the Year

The fourth quarterly review is the annual review, and it is the same page, not a different document. The three trends now have twelve points that are all this year's. The comparison column is the same quarter last year for every number. The actions section lists the year's kills and the year's failures. The ask is the year's ask. The temptation to produce something grander for the year, a deck, a long document, a presentation with a theme, is strong and should be resisted. A reader who has learned to read the quarterly page in two minutes will not learn a new format for one meeting. The page's whole value is that it is the same page.

Add one paragraph to the fourth page and nothing else: the year in five sentences. They are written the way the three monthly sentences are written: what changed, why, what the desk did, what it cost the organisation in hours when things went wrong, and what the desk carries into next year. At Bluewater that paragraph said that demand rose by about a fifth with the merger. It said that the desk absorbed it with the same five people by removing about a third of the password and remote-access tickets at source. It said that the median wait was back where it started by the fourth quarter. It said that the one bad quarter cost the organisation somewhere between a hundred and two hundred staff hours in waiting. It said that the desk would need a sixth person by the second half of next year if the growth continued. Five sentences. Yusuf read them aloud at the leadership meeting, which is the only evidence a report ever needs that it worked.

Keep every quarterly page, in order, in the same folder as the monthly spreadsheet. Four of them are a year and twelve of them are the history of the desk that nobody else will ever write. If you run QuerySolv, the standard reports on volume, backlog by age, resolution times, reopen rates and satisfaction are the five columns of that spreadsheet. Whatever tool you run, the spreadsheet is the record, because it holds what was reported at the time rather than what today's definitions make of old tickets. The desk lead who inherits the folder in three years will read the quarterly pages first. Make them worth it.

A Short War Story: The Twelve Small Steps at Acme

Acme was a manufacturer growing fast enough to open a second site. A desk lead called Jo sent an accurate one-page report every month for two years and never drew a line through any of them. Each month was "a bit up" on the one before. Created rose from about 280 a month to about 420 over the two years as the headcount grew. The median time to resolve drifted from about five business hours to about eleven. It moved in steps of twenty or thirty minutes that no single monthly page could fairly call a problem. Nobody above the desk noticed, because nobody was shown twelve points in a row. The request for a fourth technician arrived not as a trend on a page but as a complaint from the production director. His planners had been waiting two days for things that used to take an afternoon. The hire was approved as a crisis measure, and took a quarter to make and a quarter to train. The organisation ran at roughly double its old wait for the six months in between. The production director counted that in hours his planners spent waiting: something between four and seven hundred. Jo's monthly pages had every number in them. A drift that is invisible from one month to the next is obvious in twelve points, and nobody had drawn them.

The Quarterly Review Page

The template below is the page, as plain text so it pastes into an e-mail with the three charts attached or pasted beneath it. The three trends block has three columns, this quarter, last quarter and the same quarter last year. The third column is left blank until the desk has a year of history rather than filled with a guess. The actions block lists three to five with a verdict each; "too early" is a verdict. The carried block and the plan block are the two most often skipped and the two that make the next review possible. Fill in the sources line last and never remove it.

HELP DESK QUARTERLY REVIEW — ONE PAGE
Quarter: ____________   Covers: ________ to ________   Prepared by: ______________

THE THREE TRENDS (twelve months drawn; figures are the quarter's monthly average)
                                    This quarter   Last quarter   Same quarter last year
  1. DEMAND
     Tickets created / closed ..... ____ / ____    ____ / ____    ____ / ____
     What drove the change: ___________________________________________________
  2. DELIVERY
     Median time to resolve (h) ... ____           ____           ____
     Over 30 days at quarter end .. ____           ____           ____
     What drove the change: ___________________________________________________
  3. EXPERIENCE
     Satisfaction: good / replies . ____ / ____    ____ / ____    ____ / ____
     Reopens per month ............ ____           ____           ____
     What drove the change: ___________________________________________________
  Charts one, two and three beneath, dated axes, one sentence under each.

WHAT THE DESK DID THIS QUARTER (three to five, each with the number it moved)
  Action: ______________________  before ____  after ____  verdict: worked / did not / too early
  Action: ______________________  before ____  after ____  verdict: worked / did not / too early
  Action: ______________________  before ____  after ____  verdict: worked / did not / too early

WHAT THE DESK IS CARRYING (risks and waits older than the quarter, no decision yet)
  ______________________________  since ________  next step owned by: ______________
  ______________________________  since ________  next step owned by: ______________

WHAT THE DESK NEEDS NEXT QUARTER (one ask, with a horizon)
  [ ] Nothing beyond the monthly asks
  [ ] Approve: ______________________________ (hours / days / a decision) by ________
  [ ] A conversation between ______________ and ______________ about ______________
  [ ] Agreement in principle to ______________; the case follows at the next review
  The trend it would move, and by roughly how much: ______________________________

NEXT QUARTER'S PLAN (three changes, and the number each should move)
  1. ______________________________  should move ____________ from ____ to about ____
  2. ______________________________  should move ____________ from ____ to about ____
  3. ______________________________  should move ____________ from ____ to about ____

SOURCES
  Drawn from the monthly spreadsheet, rows ________ to ________.
  Definitions unchanged since ________ / changed this quarter: ______________________

Marcus drew his first quarterly page the week after Yusuf's question, from the twelve monthly rows that were already in his spreadsheet. It took him an afternoon, most of which went on the sentences under the charts. The answer to "are we better or worse than a year ago?" turned out to be "busier by a fifth, as fast as we were, and better liked". Those words were at the top of the page, with three charts beneath them that said the same thing in shapes. Yusuf read the sentence, looked at the charts for about a minute, and asked when the sixth person would be needed. That question was the next quarter's page, and Marcus had already started it.

Frequently Asked Questions