The Quarterly Hotspot Review: One Hotspot a Quarter, Written Down

The first Thursday of the quarter, 14:00, in the small meeting room at Westmoor District Council. Ruth runs the six-person desk and this is the sixth time she has held this meeting. Dele, who ran the hunt last week, has the top-five list. Clara, the head of corporate services and the person the desk reports to, has a coffee and the one-page pack. On the wall is a sheet Ruth printed in the first quarter and has added to five times: the kill record, five lines. Each is a hotspot the desk found, got a decision on, fixed and watched cool. Ruth is about to tell Clara that the top of this quarter's list is the housing repairs application. It was the top of last quarter's list and the one before. It is still there because the housing manager has not yet said yes or no. Clara writes the housing manager's name down. That is the meeting working.

The quarterly hotspot review is a short, fixed meeting in which the desk shows the organisation what its tickets say about the organisation's own things. The desk chooses one source to remove in the coming quarter, checks whether the last one actually cooled, and adds the result to a running record. The review is the ritual that turns the previous five articles from a one-off exercise into a habit, and it exists because hotspots regrow. A desk that finds and fixes one hotspot has had a good month. A desk that finds and fixes one every quarter, for years, is a different kind of desk. This article is part of our Hotspot Analysis series. It is the last in it, because the review is where the other five come together.

It covers why a ritual rather than a project, who is in the room, and the agenda in five parts. Then come building the top-five list and choosing one hotspot a quarter. The kill record, the refusals and known costs, and keeping the review honest follow. The article ends with a short war story and the review sheet to copy, in that order.

Why a Ritual and Not a Project

A project finds hotspots once, and the organisation keeps making new ones. Every purchase is a possible lemon, every new application a possible by-design ticket machine, every reorganisation a new process with a new gap in it. The hunt that found model B at Meridian Parts would, two years later, find something else. The only question is whether anyone runs it. A quarterly rhythm fits the problem. A quarter is long enough for a process hotspot to show its cycle three times and for a fix to be seen to cool. It is short enough that a new hotspot costs one quarter before it is noticed rather than two years. Once a quarter, every quarter, with the date in the diary before the previous meeting ends.

The review also solves a problem the hunt cannot, which is that the desk's evidence about the organisation's things is read by nobody in the organisation. The desk knows the repairs application is the top source of tickets in the council. The housing manager, who could change it, has never been in a room where that was said with a number attached. The review is that room. It is the one place where a count of tickets becomes a decision by someone who does not work on the desk. It works because it happens on a date rather than when someone finally gets round to raising it. Things raised when someone gets round to it are raised in the middle of a bad week, which is the worst possible moment to ask anyone for anything.

Keep it short enough that it survives. Forty-five minutes is the length this series recommends. A review that takes two hours will be skipped in the first busy quarter and never restarted. The agenda below fits forty-five minutes because most of the work, the hunt and the cards, was done the week before. The meeting reads results rather than producing them. The weekly problem hour described in Problem Management in an Hour a Week is the review's smaller sibling. It handles the repeating symptom, week by week. The quarterly review handles the source, quarter by quarter, and the two feed each other.

Who Is in the Room

Three people, and a fourth by invitation. The desk lead chairs. The technician who ran the quarter's hunt presents the list, because they read the tickets and can answer "what is actually in that line?" without guessing. The manager the desk reports to attends because the review's output is a decision that usually needs someone with authority over other departments. Another reason is that the kill record is the single best evidence of the desk's value that this manager will ever be handed. The fourth seat is for the owner of the hotspot chosen this quarter, invited for the last fifteen minutes. That way, the ask is made in the room with the count on the table and the manager present. That last arrangement is why the review gets decisions that e-mails do not.

It is not a committee, and the moment it becomes one it stops working. No standing invitation to every department head, no slide deck, no minutes beyond the sheet at the end of this article, no action log with twenty-three items. A committee meets to discuss. The review meets to pick one thing and check one thing. The discipline of one is what keeps it to forty-five minutes. (Committees breed. Leave one unattended in a council and by spring it has a sub-group.)

The Agenda in Five Parts

The agenda is the same every quarter, in the same order. The table below gives each part its minutes, its input and its output. Part one checks last quarter's fix against the cooling rule from Killing the Source, Not the Symptoms. That rule means the same cut, the same period length, and a rate at or near the peers. A second period is still to come or now complete. Part two reads this quarter's top five. Part three chooses one. Part four reads the refusals and the known costs. That is the list of hotspots the organisation has decided to live with and what each costs it per year. Part five updates the record, in the room, so that the record is never a job for later.

Hold the order even when part one is dull. A quarter in which the last fix cooled cleanly takes two minutes to report. It is the most important two minutes of the meeting, because the manager sees that the desk's predictions come true. A quarter in which the last fix did not cool takes the full ten and is more important still. The review is where "the fix did not work and here is what we think happened" gets said to the person who approved it. The people who proposed it say it, before anyone else notices. A desk that only reports its successes at the review will be believed about neither.

Building the Top-Five List

The list comes from the quarter's hunt worksheet, described in Finding Hotspots in the Ticket Data, and from the definition cards it produced. It is ranked by desk hours per year rather than by ticket count. A source producing forty quick tickets a month and one producing twelve slow ones may cost the same hours, and hours are what the organisation is paying for. Each line carries the source, its kind, tickets per month, and hands-on hours per ticket as a range from a sample. It carries desk hours per year as a range, and the rate against its peers. Five lines is enough. The sixth entry on a steep list is rarely a decision anyone needs to make this quarter. A list of fifteen tells the manager that nothing on it is urgent. If you run QuerySolv, the quarter's volume report by department is the one page worth stapling behind the list. It is what the manager will ask for when a group hotspot comes up, and it takes a minute to pull.

Mark each line as new or carried, and say how many quarters it has been carried. A list that is mostly carried is normal and is not a failure. The desk can kill one hotspot a quarter, so the second, third and fourth on last quarter's list are expected to reappear, each one place higher. A line that has been carried for three quarters without a decision is a different thing. The marking makes it visible without anyone having to raise it as a complaint. Westmoor's repairs application had a "carried, 3" against it, and that small annotation did more than any sentence Ruth could have said to it. A line that is new is news, and gets a minute of its own. Where did it come from? Is it a hotspot or an incident that has not been taken out of the count?

Be suspicious of a list that changes completely between quarters. Hotspots are durable; that is the point of them. If the top five is five new names, one of these things happened. The hunt's method changed, the labels drifted, or the quarter contained a large incident that was not removed before the rates were read. Go back to the worksheet and check the filter lines before presenting the list. The familiar list is the honest one, which is uncomfortable. It is also why the review needs a manager in the room. A familiar list with no decisions against it is a list of things the organisation has chosen. The choosing should be done on purpose.

One Hotspot a Quarter

Choose one, because a desk of one to ten people can finish one in a quarter and cannot finish three. A hotspot half-fixed is a hotspot. The fix work from the previous article is a quarter's worth of attention around the ordinary work of the desk. That includes the case, the owner conversation, the change, the pilot, the roll-out and the verification. Starting three means finishing none and reporting progress on all of them, which is the condition most improvement programmes reach by their second quarter and never leave. Pick one, finish it, prove it cooled, and the kill record grows by one line every three months. Over three years that is a dozen sources removed, which is more than most desks remove in a decade.

Choose the top of the list, unless there is a reason written on the sheet. The reasons that justify going lower are few and specific. The owner of the top hotspot may have refused, with the refusal still recent. In that case, the second is chosen and the first stays on the refusals register. Or the second may have a cheap redesign available, while the first needs a purchase that will not be decided this quarter. Then the second is done now and the first's case is prepared for the budget round. "The second is more interesting" is not a reason, and neither is "the first is hard". The written reason is what stops the list from being a menu.

The one standing exception is a hotspot whose fix is a single setting or a sentence to an owner, costing an hour and removing a stream of tickets. Do that as well, not instead, and record it as a kill with its hours. Oakhurst Care Homes' rota setting, from the fourth article in this series, was that kind of fix. A review that finds one of these should take it in the same breath as choosing the quarter's hotspot. The alternative is to leave a known hour's fix for three months in the name of discipline. That is not discipline. It is a queue.

The Kill Record

The kill record is the running register of every hotspot the desk has removed. It records the hotspot's name, the quarter it was found, the quarter it cooled, and the rate before and after. It records the fix used, the owner who decided, and the desk hours per year it no longer costs, as a range. It is one page for years and it is the most persuasive document the desk owns. Every line on it is a thing that used to cost the organisation time and does not any more, with a date and a number against it. Keep it where the manager sees it, bring it to every review, and add to it in the room. A kill record maintained for two years separates a desk that is "quite good, I think" from one that has removed avoidable work. The latter has removed four hundred hours a year and can show you which four hundred.

Westmoor's record after six quarters is this article's fictional example. It listed the planning department's large-format scanner, a lemon replaced, at forty to sixty hours a year. The leisure centres' till application's by-design access requests, a redesign to a standard role, were ninety to a hundred and thirty. The car park payment machines' failing modems, a bad batch replaced under warranty, were thirty to fifty. The month-end finance reports, a redesign of three roles and a report that runs itself, were seventy to a hundred. The new-starter process, a single onboarding form with a lead time, was a hundred and twenty to two hundred. Those five lines meant three hundred and fifty to five hundred and forty desk hours a year, which is most of a technician. The hotspots were found in the desk's own tickets and removed without a single new system. Clara had the total in her head by the third quarter, which is the point.

Follow the honesty rules, because the record is only persuasive while it is believed. Hours as ranges, from samples, never a point. A hotspot counted once, however many quarters it took. "Cooled" only after two periods at the peer rate, and a line that regresses is reopened and marked, not quietly deleted. The hours are desk hours; requester hours, where known, go in a separate column and are never added into the desk's total. And put no currency figures anywhere on it, for the reason the whole of this group of articles keeps repeating. The desk knows hours, and the moment it guesses at money the record becomes a debate about the guess. The desk owes its leadership wider quarterly reporting, with a one-page format and business language. Where the kill record meets that reporting is the subject of The Quarterly Review in the reporting series. The kill record is one line of that page, and often the one that gets read.

The Refusals and the Known Costs

The second register is shorter and less pleasant, and the review is incomplete without it. It lists the hotspots whose fix was asked for and declined, with the date and the reason given. It lists the hours per year the organisation is choosing to spend by leaving them, and the date they will be raised again. It is written in flat language. "Housing repairs application: standard role for new repairs officers requested in the spring review; declined by the housing manager pending the system replacement project; cost eighty to a hundred and twenty desk hours a year; revisit at the autumn review." No adjectives. The register is a statement of what the organisation decided, kept where the organisation can see it. Its whole force comes from being accurate and dull.

Read it aloud at every review, every line, in five minutes. Most quarters nothing changes, and the reading takes two minutes and reminds the manager what the organisation is paying for and why. Some quarters something has changed. The project that was the reason for the refusal has slipped a year, or a new manager owns the application, or the hours have grown. In those cases, the line is moved back onto the top-five list with its case already written. Refusals revisited this way are approved more often than refusals argued at the time, because nobody is being asked to reverse themselves. They are being asked whether a decision from last year still holds, which is an easier question to say no to.

Keeping the Review Honest

The review is a list of things, not a list of people, and it has to be defended against becoming either kind of scoreboard. It is not a league table of departments. A group hotspot is a thing hotspot wearing a department's name, and the review names the thing. It is not a measure of technicians. The hunt does not cut by who resolved what. If a per-technician figure ever appears in the room, it is for a coaching conversation elsewhere, never for ranking. It must never come from fewer than a few dozen tickets per person in the quarter. The moment a department or a person feels judged by the list, the tickets that feed it start arriving by phone instead. The review then loses its data. Keep the names on the list to models, applications, processes, rooms and groups, and the people will keep telling you the truth.

Keep the method fixed, and say so when it is not. Use the same export, the same filter lines, the same labels, the same period length, every quarter. Where the desk improves the method, as it will when the form gains a device field, mark the quarter where the change happened. In that case, expect the list to shift that once. A review that cannot explain why this quarter's list differs from last quarter's has lost its most important property. That is the manager's ability to trust the comparison without checking it.

Finally, do not find a hotspot because the meeting needs one. A desk that has killed its top five over two years will reach a quarter where the list is flat. Every source is near the peer rate, with nothing worth a quarter's work. Say so, in ten minutes, and go back to the queue. That quarter is the desk's best result, not an empty meeting. The manager who hears "we have nothing above the line this quarter" from a desk with a six-line kill record on the wall will believe it. The desk has shown it knows the difference. Hotspots return. The review will be there when they do.

A Short War Story: The Quiet Year at Thistle Brewing

Thistle Brewing ran a desk of three across a brewery, a bottling plant and a chain of twelve pubs. In one good year its desk lead, Callum, found and removed four hotspots without ever calling them that. A lemon handheld in the pubs was replaced. The stock system's by-design access requests were redesigned to a standard role. A cold-store office that cooked its two computers was re-equipped with sealed units. The monthly duty-rota publication had produced a burst of tickets every month until the publisher's settings were changed. Ticket volume fell by about a third over the year. The desk was quietly proud, and nothing was written down beyond the tickets themselves. At the following year's budget round, the finance director noticed that tickets had fallen by a third and that the desk still had three people. The director proposed reducing it to two on the grounds that the work had evidently gone down. Callum, asked to respond, could say that the desk had made the tickets go away. But he could not, in the meeting, show which tickets, when, or what each fix had saved. The evidence was scattered across a year of closed tickets and his memory. The reduction went ahead. A kill record with four lines on it could have been kept on one page and shown at four quarterly reviews. That would have made the same fall in volume the desk's strongest argument instead of its weakest. A hotspot killed and not recorded is a budget cut waiting for a finance director to find it. The wider case for why a quieter queue is not a smaller job is in The Headcount Case. Callum would have liked to have read it the week before.

The Review Sheet

The sheet below is the forty-five minutes on one page, filled in during the meeting and filed with the quarter's hunt worksheet and cards. Part one is last quarter's kill, with the cooling rule's boxes. Part two is the top five, ranked by hours, with the new or carried marks. Part three is the one chosen, with the reason if it is not the top line, and the ask in one sentence. Part four is the refusals register, read aloud. Part five is the record update and the date of the next review. The kill record itself is a separate page that grows by a line a quarter. Its column headings are given at the bottom of the sheet so that the two documents agree. Fill in hours as ranges, name a person for every decision, and never leave the room without the next date.

QUARTERLY HOTSPOT REVIEW — ONE PAGE
Desk: ______________   Quarter reviewed: ____________   Review date: ____________
Present: chair ______________  hunt presenter ______________  manager ______________
Hotspot owner invited: ______________ (for part 3)

1. LAST QUARTER'S KILL (10 minutes)
   Hotspot: ______________________   Fix used: [ ] replace [ ] retrain [ ] redesign [ ] retire [ ] automate
   Rate before: ______   Rate after, period 1: ______   period 2: ______   Peer rate: ______
   Control source and its change: ______________________
   Tickets moved to another label or channel? ______________________
   Status: [ ] cooled, add to kill record  [ ] still watching, second period due ____________
           [ ] not cooled; what we think happened: ______________________________________

2. THIS QUARTER'S TOP FIVE, RANKED BY DESK HOURS PER YEAR (10 minutes)
   #  Source                        Kind      Tickets/month  Hours/ticket   Desk hours/year   Rate vs peers  New / carried (quarters)
   1  __________________________    _______   ______         ____ to ____   ______ to ______  ____ x         __________
   2  __________________________    _______   ______         ____ to ____   ______ to ______  ____ x         __________
   3  __________________________    _______   ______         ____ to ____   ______ to ______  ____ x         __________
   4  __________________________    _______   ______         ____ to ____   ______ to ______  ____ x         __________
   5  __________________________    _______   ______         ____ to ____   ______ to ______  ____ x         __________
   Method unchanged from last quarter? [ ] yes  [ ] no, what changed: ______________________
   Known incidents removed before rates were read: ______________________

3. THE ONE FOR THIS QUARTER (10 minutes)
   Chosen: ______________________   Top of the list? [ ] yes  [ ] no, because: ______________________
   Owner outside the desk: ______________   In the room? [ ] yes  [ ] no, to be asked by ____________
   Options asked in order (retire / redesign / automate / retrain / replace): ______________________
   The ask, in one sentence: _______________________________________________________________
   Expected tickets per month after: ______   To be counted on: ____________ and ____________
   Quick kill taken as well (one-hour fix)? ______________________   Hours/year: ______ to ______

4. REFUSALS AND KNOWN COSTS (5 minutes)
   Hotspot                      Asked on     Declined by      Reason                 Desk hours/year    Revisit on
   _________________________    __________   ______________   ___________________    ______ to ______   __________
   _________________________    __________   ______________   ___________________    ______ to ______   __________
   _________________________    __________   ______________   ___________________    ______ to ______   __________
   Moved back onto the list this quarter: ______________________

5. THE RECORD (10 minutes)
   [ ] Kill record updated in the room       [ ] Refusals register updated
   [ ] Cards and hunt worksheet filed at: ______________________
   [ ] Next review booked for: ____________   Hunt to be run the week of: ____________
   Kill record total, desk hours per year freed to date: ______ to ______

KILL RECORD COLUMN HEADINGS (separate page, one line per hotspot)
   Hotspot | Kind | Quarter found | Quarter cooled | Rate before | Rate after | Fix used |
   Decided by | Desk hours/year freed (range) | Requester hours/year, if known | Notes

Ruth's sixth review ran to forty minutes. The planning scanner, last quarter's kill, had completed its second period at the fleet rate and went onto the record as line six. The repairs application stayed at the top with "carried, 3" against it. Clara, rather than Ruth, asked for the housing manager to attend the last quarter of an hour of the next review. That was a sentence Ruth had been trying not to say for nine months. Dele had run the hunt and had never seen the record before. Dele read the total at the bottom twice and asked whether it was really all from the tickets. It was. It always is.

Frequently Asked Questions