Standardising the Printer Fleet to Cut Tickets

Thursday, 15:30, at Acme, a company of about a hundred and forty people in one building and a small depot. Omar has just finished the printer ticket census for the desk and has a second spreadsheet open beside it. That is the list of printers the census made him build because nobody had one. There are forty-eight devices. They are thirty-one different models from six makers. Nineteen of them sit on individual desks and belong, in the sense that matters, to the people whose desks they sit on. The toner cupboard holds forty-four kinds of cartridge, eleven of which fit nothing the company still owns. Omar's manager, reading over his shoulder, asks how it got like this, and the honest answer is that nobody decided anything. Every printer was bought by someone who needed one that week.

A printer fleet is every printing device the desk supports. Standardising it means deciding, in writing, how many models the fleet contains, where devices go, and who may have one. It also means deciding how consumables and repairs are handled and how a new printer is requested. Then the fleet is brought to that standard as devices are replaced. This is the article in this series where the numbers move most. Every model in the fleet is a driver to manage, a set of consumables to stock, a jam procedure to learn and a card to write. Every device on a desk is a device with no queue, no spare and no champion. This article is part of our Printer Support series. It is about choosing the fleet instead of inheriting it.

It covers why variety costs tickets, fewer models, and fewer locations with the printer-to-staff ratio worked through. Personal desk printers and the conversation about retiring them follow, then the managed-print arrangement described generically. Next comes what the standard removed from the queue on a fictional desk, then writing the standard down. A short war story and the fleet standard to copy finish the article, in that order.

Why Variety Costs Tickets

Every model in the fleet is a separate thing the desk has to know, and the knowing is where the hours go. A model has its own driver, which the drivers article wants held once on the server and tested against the documents people actually print. It has its own consumables, which need a shelf, a par level and a reorder route. It has its own paper path, so the jam-clearing steps and the on-printer card differ. It has its own screen, its own error codes, its own way of printing a test page. It also has its own idea of where the bypass tray is. A desk that supports three models carries three of each. A desk that supports thirty-one carries thirty-one, and on any given day knows about four of them well.

Variety also defeats spares. A fleet with three models can hold one spare device, or one of each, and swap a dead printer in an hour. A fleet with thirty-one cannot hold spares for any of them and swaps a dead printer when a new one arrives. That takes days, during which the floor walks to another device and raises tickets about that one being slow. And variety defeats the knowledge base. A jam article with pictures is cheap for one model and a month's work for thirty-one, so it is written for none. The desk answers the same question thirty-one slightly different ways.

The honest edge is that some variety is earned. Examples are a warehouse label printer and a wide-format plotter in a drawing office. Others are a device that prints on card for a marketing team and a receipt printer at a counter. These are specialist devices doing work the general-purpose device cannot. A standard that tries to abolish them will be ignored by the people who need them. The standard names them as exceptions, with an owner each, and holds the rest of the fleet to two or three models. The point is not uniformity for its own sake. It is that every model should be there because someone chose it, which is the one thing Omar's list could not say about any of them.

Fewer Models: Two or Three, Not Thirty

The general-purpose fleet needs two models, possibly three, and the shape is the same on most desks. One is a workgroup mono device, which prints black on plain paper quickly and cheaply. It is what most of the building needs most of the time. Another is a workgroup colour multifunction device, which prints, copies and scans, in colour when asked. It sits where the scanning and the occasional colour job happen. Optionally, add a smaller device for a spot where the workgroup one does not fit, a reception desk or a small branch. Choose it from the same maker's range so that it shares a driver family and, ideally, a cartridge with one of the other two. Pick the models from one maker where possible, so that one family driver covers the fleet and the consumables cupboard has three shelves.

Choose the models on the evidence the desk already holds. The census says which devices generated the fewest tickets per page printed. The device and model hotspots article in the hotspot series shows how to tell a lemon model from a lemon unit. That matters here because standardising on the model that produced half the jam tickets would be an expensive way to make a point. Where the evidence is thin, run two candidate models side by side on the busiest floor for a quarter and count. A trial costs two devices. The wrong standard costs a fleet.

Then bring the fleet to the standard by attrition, not by decree. Every device that dies is replaced with the standard model. No device is thrown away while it works, unless the measuring article has declared it dead on its ticket count. On a typical fleet that means the standard is mostly in place within two or three years and complete within five. The budget conversation each year is "the same number of replacements as last year, all the same model". That is a sentence a finance team can sign. The alternative, replacing forty-eight devices in one project, is a sentence a finance team has heard before and will ask you to repeat slowly.

Fewer Locations: The Printer-to-Staff Ratio

The second lever is how many devices there are at all. And the useful measure is the printer-to-staff ratio: the number of people who share each general-purpose device. Omar's forty-eight devices for a hundred and forty people was about one device for every three people. That ratio produces a large fleet of lightly used devices, each needing a driver, a cartridge and a card. Most of them print a few pages a day. The question the standard asks is how many people can share one well-placed workgroup device before the walk and the wait become a cost of their own.

Here is the worked example at Acme, with every figure invented for the purpose and none of them a benchmark for your building. Omar counted the pages: the whole company printed somewhere between forty and fifty thousand pages a month. Most of those pages came from the finance, sales-admin and operations teams. Most of the nineteen desk devices printed fewer than two hundred pages a month each. He proposed one workgroup device for every twenty to thirty desk-based staff, placed within a short walk of the desks it served. He also proposed one colour multifunction per floor. For the heavy-printing teams he halved the ratio, and for the depot he kept one device because the depot was a separate building. The arithmetic came out at eleven general-purpose devices plus four exceptions, fifteen in all, against forty-eight. Each person's walk to a printer went from a few steps to perhaps twenty, which is the cost the standard accepted on purpose.

The ratio is a range, not a rule, and the range depends on the work. A claims team printing all day needs a device nearer than a software team printing twice a month. A floor with an open plan can share one device across forty desks; a floor of small offices and corridors cannot. Confidential printing means HR, payroll, legal and anyone handling medical or personal records. It needs either its own device in a lockable room or a queue that holds the job. With that queue, the job stays held until the person is standing at the printer and releases it. Most workgroup devices can do this, which removes the main argument for the desk printer in HR. Write the ratio the desk chose, and the reasons, into the standard, and revisit it when the census says the walk has become a ticket.

Personal Desk Printers and the Conversation About Them

The hardest part of standardising is the nineteen devices on desks, because each belongs to a person who asked for it, got it, and has a reason. The reasons are real. Convenience, for someone who prints all day. Confidentiality, for someone whose printouts should not sit in a shared tray. Habit, for someone who had one at their last job. And status, which nobody says aloud and which is the reason the one in the corner office has never jammed and never been counted. The desk's case against them is also real. A desk printer has no queue on the server, no spare, no champion, and a cartridge type nobody else uses. It also has a user who raises a ticket when it fails, because it is still the desk's problem when it does.

The conversation goes better when it offers the need a better answer than the device gave it. Convenience gets a shared device placed within sight of the desk and the defaults set so it just works. Confidentiality gets the held-until-released queue, which is more confidential than a desk printer in an open-plan office whose tray anyone passing can read. Habit gets a fortnight's grace and a visit on the first day. Status gets, in my experience, left until last and handled by the desk lead rather than the technician. It is framed as a building-wide standard that applies to everyone, which it does. The craft of declining a request while solving its need is in Offering an Alternative Instead of a Wall. The desk printer is that article's purest case: nobody wants a printer; they want their pages, privately, now.

The policy sentence that makes the conversation possible is in the scope, not in the standard. The desk supports printers on the standard, on shared queues, and nothing else. That is a line the scope article shows how to write and publish. Once it is written, a desk printer is an exception with a name, an approver and a review date, like any other. The request for a new one then goes through the same route as a request for a monitor, which the equipment request workflow describes. Retire the existing ones by attrition too, with a note that the next replacement is the shared device. That way, nobody's printer is taken away and everybody's is eventually not replaced. A desk printer that dies of old age causes no meeting.

The Managed-Print Arrangement, in Plain Words

A managed-print arrangement is a contract under which a supplier provides the devices. The supplier delivers the consumables automatically when each device reports that it is running low, and sends an engineer when a device faults. The supplier charges by the page printed rather than by the device bought. This article names no suppliers and quotes no rates; what matters to the desk is what the arrangement takes away and what it leaves. It takes away the consumables cupboard, the reorder route, the engineer call-out and, in most arrangements, the purchase of the device itself. It leaves the queue, the driver, the defaults, the mapping, the user's machine, the card, and the ticket. Every one of those is still the desk's.

Four things belong in the contract from the desk's side, and they are the four that cause tickets when missing. A response time for a faulted device, in business hours, that the desk can repeat to the floor. The single route for raising a fault with the supplier is the desk. That way, users never phone the supplier directly and the desk never discovers a visit after the fact. A replacement clause, so that a device the engineer has visited three times for the same fault is swapped rather than visited a fourth time. And a monthly report of pages and faults per device, which is most of the data the measuring article needs and which the supplier already has. If you run QuerySolv, service-level targets per priority sit on a business-hours schedule. The promise the desk makes on a printer ticket can then match the hours the supplier's contract counts in rather than a different clock. Whatever the tool, make sure the two promises use the same hours, or the desk will be apologising for a breach the supplier does not recognise.

The failure mode of managed print is the two front doors. If the supplier gives users a sticker with its own number, users call it and engineers arrive unannounced. In that case, the desk's tickets stay open with nobody knowing the fault is fixed, and the census misses every one of those faults. The desk is the front door, always. The supplier is a vendor the desk escalates to, with the same handoff discipline as any other vendor ticket. One sticker, with the desk's address on it, and the supplier's reference number written in the desk's ticket, is the whole arrangement working.

What the Standard Removed From the Queue

The standard earns its place by what it takes out of the queue. So measure before and after with the census, by category and by device. Below is Acme's picture, two years into attrition, with the fleet at twenty-two devices on the way to fifteen. Every number is invented for the example. The reductions are uneven, and that is the honest part. Driver and wrong-output tickets fell hardest, because fewer models on central queues removed most of their causes. Consumables fell because the supplier's automatic delivery emptied the cupboard problem. Jams fell less, because jams are mostly paper and rollers and the standard does not change the weather. Scanning rose a little, because the colour multifunction devices now scanned for everyone and more people used them.

Report it in hours, because hours are the currency the desk and the business share. At Acme the drop was roughly twenty to thirty technician hours a quarter. Then there were the hours the floor no longer spent walking to a second device because the first had no toner. Nobody measured that saving, and it was larger. Do not report it as a percentage of tickets and leave it at that. A percentage invites comparison with a desk whose counting rules differ, and the counts are what a reader can check. The line that goes in the monthly report to leadership is one sentence with the two counts and the hours. It is written the way the monthly report they will read describes.

Writing the Standard Down

The standard is one page, and it lives with the desk's other standards, where the scope and the queue standard already are. It names the models, the ratio and its reasons, the placement rules, and the confidential-printing answer. It gives the consumables and repair route, the specialist exceptions with their owners, and the request route for a new device. It also gives the rule for retiring old ones and who approves a change to any of it. Approval matters more than it looks. A standard the desk wrote alone is a preference, and a standard the operations manager signed is a rule. The difference shows the first time someone senior asks for a printer on their desk. Get it signed. Then review it once a year against the census, which is a one-hour meeting. Review it more often if a new building, a merger or a new line of work changes what people print.

A Short War Story: The Depot Printers at Bellweather Logistics

Bellweather Logistics ran eight depots, and for years each depot manager had bought printers from whichever shop was nearest when one died. So by the time a technician called Nadia inherited the desk the company owned around forty models. The desk kept a drawer of fifty-odd cartridge types, most of them for one device each. On a Friday in the busiest month, the dispatch printer at the largest depot failed. There was no spare, no matching cartridge for the nearest alternative and no engineer until Tuesday. The depot's drivers' manifests were handwritten for a day and a half. The depot manager estimated the cost at a full day of two office staff and some lost deliveries nobody could quite count. Nadia used the incident to get a two-model standard signed and bought three spare devices to hold at the hub. She moved consumables to automatic delivery and replaced depot printers only as they died. Two years later a depot printer failed on a Friday again and a spare was on a van within the hour. Nobody at the depot found that remarkable, because they had forgotten the first Friday. The lesson in her standard's first line was that a fleet nobody chose has no spares, by definition, because nobody knew what to buy.

The Fleet Standard

The template below is Acme's standard, cleaned up for reuse. Section A is the models, B the ratio and placement, and C the confidential-printing answer. D is the consumables and repair route, E the exceptions with owners, F the request and retirement rules, and G the signature and review. Fill in your own models and your own ratio with its reasons. The numbers in the worked example above are Acme's, and your building, your teams and your printing are different. The standard is short on purpose. A fleet standard that lists every setting is a configuration document. And that belongs in the queue standard; this page says what the fleet is and who decided.

THE PRINTER FLEET STANDARD — ONE PAGE
Owner: ______________________   Approved by: ______________________   Date: ____________   Next review: ____________

A. THE MODELS (general purpose; from one family where possible)
   Workgroup mono:              ______________________   Shared driver family: ______________________
   Workgroup colour multifunction: ___________________   Cartridges shared with: ______________________
   Small / branch device:       ______________________   (optional; same family)
   No other general-purpose model is bought. Replacement is by attrition unless a device is declared dead on its count.

B. RATIO AND PLACEMENT
   One shared device per ______ to ______ desk-based staff, within a short walk; reasons: ______________________
   Heavy-printing teams (______________________): one per ______ to ______
   One colour multifunction per floor / area: yes / no       Separate buildings: one device minimum each
   Devices are placed away from external doors and damp; paper is stored where it is used.

C. CONFIDENTIAL PRINTING
   Departments with confidential output: ______________________
   Answer: held-until-released queue on the shared device / dedicated device in a lockable room (pick per team)

D. CONSUMABLES AND REPAIR
   Route: desk-held stock with par levels / managed-print supplier with automatic delivery (pick one)
   Supplier response time (business hours): ______   Single front door for faults: the desk, always
   Replacement clause after ______ visits for the same fault: yes / no      Monthly pages-and-faults report: yes / no

E. SPECIALIST EXCEPTIONS (one line each)
   Device ______________  Purpose ______________________  Owner ______________  Review ________
   Device ______________  Purpose ______________________  Owner ______________  Review ________

F. REQUESTS AND RETIREMENT
   A new or replacement printer is requested through: ______________________ (the equipment request route)
   Personal desk printers: not supported; existing ones are not replaced; exceptions in section E with an approver.
   A device is retired when it fails and is replaced with a section A model, or when its ticket count declares it dead.
   Spares held: ______ mono, ______ colour, at ______________________

G. CHANGES TO THIS STANDARD
   Proposed by ______________  Agreed by ______________  Date ________  What changed: ______________________

Omar's standard went to the operations manager with the census stapled to it. It came back signed in a week, mostly because the toner cupboard had been photographed. Two years on, Acme has twenty-two devices from three models and three shelves in the cupboard. Acme has a corner office whose printer died of old age last spring and was, to nobody's great surprise, not replaced.

Frequently Asked Questions