When Commercial Is Expensive: Seats, Suites and Shelfware

Tuesday, 11:20, at Harbourview Clinic. Farah runs a desk of three that looks after two sites and ninety staff. It also looks after a set of clinical systems that must not go down during surgery hours. Marcus, the practice manager, has walked over from finance with a printout and a question he has clearly been saving. "Why is the help desk the biggest software line we have? Bigger than the clinical records system. For three people." Farah knows the answer, which is that the help desk is not a help desk. It is an ITSM suite, bought two years ago. Harbourview then expected to merge with a second group of clinics and grow to a desk of twelve. A consultant configured the suite over six weeks. It was licensed for fifteen agents and four approvers on a three-year term. On an ordinary Tuesday, Farah, Dev and Priya use it to close about thirty tickets. The merger did not happen. The contract did.

A commercial help desk is expensive when the organisation pays for more than the desk uses. That means more seats than people, more features than habits, more process than the team has, or more years than the need lasts. None of those is the vendor's fault and none of them shows on the invoice as waste. The invoice describes what was bought, not what was used. This article is part of our Open Source vs Commercial Helpdesk Tools series. And it is the companion to The Hidden Costs of Commercial: that article lists where commercial costs hide. This one is about the four situations in which they stop hiding.

It covers why the expensive part is rarely the licence you use, then paying per agent for people who log in twice a year. Next come features nobody uses, the suite bought for a desk of three, and the contract that outlives the need. The arithmetic of over-buying, with a table, and when commercial is genuinely worth it follow. Then come a short war story, the fit check to copy, and right-sizing without ripping everything out, in that order.

The Expensive Part Is Rarely the Licence You Use

The seats the desk uses, for the features it uses, for the years it needs them, are nearly always fair value. A desk of three paying for three seats of a tool it works in all day has little to complain about. What makes commercial expensive is the gap between that and what was bought. Every seat beyond the third and every feature behind a door nobody opens is cost without use. So is every process the suite implements that the clinic does not run. So is every month of the term after the need has changed. The gap is invisible on the invoice because the invoice has no column for "used". It has a column for "quantity", and the quantity was decided on the day of purchase by people imagining a future.

The gap grows in one direction. Seats get added and are rarely removed, because adding is a request and removing is an admission. Features are bought in bundles, so the one you need brings eleven you do not. Suites are sold for the organisation you plan to be, and plans change more often than contracts. And the term, once signed, is the one quantity nobody can adjust at all. Commercial software is rarely expensive on the day it is bought. It is expensive on the day somebody finally compares the two columns, which at Harbourview was a Tuesday, two years in, with a printout.

Paying Per Agent for People Who Log In Twice a Year

An occasional user is someone with a full agent seat who touches the tool a handful of times a year. And most desks have more of them than agents. The department head who approves equipment requests twice a quarter. The clinical lead who wanted to "see what IT is doing" and looked once. The practice manager who signs off on changes. The locum who covered the desk for a fortnight in the summer. The facilities contractor who fixes the door entry system and was given a login so he could update his own tickets. Each has a seat. Each seat costs what Farah's costs. Their combined use of the tool in a year would not fill one of Farah's afternoons.

The sign is the login report. Most tools can show when each agent last signed in. A desk that pulls that list once a quarter will usually find a third of its seats unused since the previous quarter. Usually, a few have never been used at all. The fix depends on the tool and the person: some tools offer a cheaper role for people who only approve or only view. And some let an approver act from an e-mail without a seat. Some do not. Then the choice is between paying for the seat and finding a way for that person to do their two tasks a year without one. For an approver, that is often a reply to a message. Ask what the occasional user actually does. The answer is rarely "works tickets", and "works tickets" is the only thing a full seat is for.

The seat question is a property of the licensing model, not of commercial software as such. Under a per-agent subscription, each occasional user is a line on the invoice to be defended at every renewal. Some self-hosted tools, including QuerySolv, are sold under a licence per installation. Under that model, the occasional user is simply a login to administer. The cost moves from the invoice to the server you run and the accounts you keep tidy. Neither is free of consequence. The difference is whether the consequence is paid in money every month or in administration every quarter. And a desk with many approvers and few agents should notice which of those it has more of.

Features Nobody Uses

Shelfware is software that was paid for and is not used. In a help desk it is usually features rather than whole products. There is the automation engine nobody has written a rule for, or the asset register with eleven entries. There is the satisfaction survey that was switched off after the first complaint. There is the reporting module whose one report is the same one the base product drew. It got there by the demo effect. The demo showed everything. The evaluation checklist scored the tool on how much of everything it had. The desk bought the bundle that scored highest, which was the bundle with the most things in it. Features on a checklist are weightless. Features on an invoice are not.

The cost of shelfware is the bundle it forced. Feature bundles are priced so that the one feature a desk needs sits in the level above the one it would otherwise choose. The whole desk moves up a level to get it. At Harbourview the one feature was approvals, needed because equipment purchases had to be signed off by a department head. The level that had approvals also had change management, asset management, a configuration database and a service catalogue. Every one has been configured by the consultant and used by nobody since. The desk is paying for a filing cabinet with one drawer open.

Audit it the way you audit seats. Once a year, list every feature or module on the invoice. Write beside each the name of the person who used it last and roughly when. Mark the ones with no name. Then ask of each unmarked feature whether its absence would change anything the desk does. That is the test that Evaluation Criteria That Matter applies before purchase. It applies just as well after. A feature nobody would miss is a line to negotiate away at renewal or a reason to drop a level. A feature nobody would miss that cannot be removed from the bundle is a reason to look at the bundle itself.

The Suite Bought for a Desk of Three

An ITSM suite is a product that implements many of the formal processes of IT service management at once. These cover incidents, requests, problems, changes, assets, a configuration database and a service catalogue. Each has its own module, its own records and its own expectation that someone owns it. Suites are built for organisations with a team per process, and they are excellent for those organisations. A desk of three does not have a team per process. It has three people and thirty tickets a day. A suite asks each of those people to be a change manager, an asset manager and a catalogue owner in the gaps between tickets. The gaps are where the suite lives, and the gaps are not large.

The expense has three parts, and the licence is the smallest. The first is configuration: a suite arrives as possibilities rather than defaults. Turning possibilities into a working desk takes a consultant's weeks or an administrator's months. Every one of those is paid for whether or not the processes configured are ever run. The second is upkeep: every configured module needs maintaining through every upgrade. And a module nobody uses still has to be kept from breaking the ones they do. The third is weight: a suite makes the simple ticket slower. The form has fields the processes need, and the categorisation has levels the reports need. The close requires a step the change module needs. A desk of three pays that weight on all thirty tickets a day. The processes that drown a small team, and the handful worth keeping anyway, are set out in the article on ITIL processes that drown a small team. A suite installs all of them by default. The drowning is optional but has to be declined module by module.

The suite is usually bought for the organisation someone expected to exist. Harbourview's was bought for a merged desk of twelve across five sites, with a change process the regulator would want to see. Those were reasonable things to expect at the time. Expectations are a fine basis for a plan and a poor one for a three-year licence. The honest question at purchase was not "will we need this" but "what does it cost if we do not". Nobody asked it, because the demo had been excellent and the consultant was already booked. Suites are not sold to desks of three. They are sold to desks of three that have been told they will be twelve.

The Contract That Outlives the Need

The term is the one quantity that cannot be audited away, and it is where commercial becomes most expensive when the organisation changes shape. A three-year term may have been signed for a merger that did not happen. A five-year term may have been signed before the desk was outsourced. A term may run on after the department shrank from six to two. Or it may run on after the company was itself acquired and moved onto the parent's tool. Each case leaves an organisation paying for a desk it no longer has. Multi-year terms are sold with a discount, and the discount is real. What it buys is the vendor's certainty, and what it costs is yours. A fixed rate is worth something. The right to stop is worth something too, and it is the thing the discount is for.

The sign is any sentence beginning "we'll have grown into it by then". Growth is a forecast, and the desk is being asked to sign a contract on it. The defence is to size the term to the certainty rather than to the plan. Choose a one-year term while the merger is a possibility, a longer one when it is a fact. Buy seats for the desk you have. Get a written right to add more at the same rate when the desk you expect arrives. Where the vendor allows it, get a clause that lets the seat count fall at each anniversary rather than only rise. Vendors will not always agree. Knowing which clauses you asked for and were refused is itself information about the vendor, and it belongs in the decision record.

When the need has already outlived the contract, the options are narrower but not empty. Read the agreement for a reduction clause or a termination fee that may be less than the remaining term. Or look for an option to convert unused seats into something the organisation does use. Ask, in writing, and be plain about why. Vendors prefer a smaller customer to a departed one and will often meet a reasonable request halfway. And count the months left, because a contract with eight months to run is not a crisis. It is a renewal checklist with a long lead time. The preparation for it is the same as for any other.

The Arithmetic of Over-buying

Put bought against used on one page, in counts, and the gap becomes something a practice manager can read in a minute. The table below is Harbourview's, as a worked example, with the invoice in one column and the desk's actual use in the next. The last column is the fix available at the next renewal or, for the term, the conversation to have before it. Counts rather than money, because counts survive a change in the fee and because nobody argues with a count of three.

Read the gap column and resist the urge to feel foolish, because every line was reasonable when it was bought. The point of the table is the last column. The point of the last column is that nearly every gap closes at a renewal if someone arrives with the counts. A desk that takes this table into its renewal is negotiating. A desk that takes the invoice is approving.

When Commercial Is Genuinely Worth It

Fairness requires the other column, because many desks that pay for commercial software are getting exactly what they pay for. A desk of three with no server and no administrator time is right to buy hosting, patching, backups and support rather than spend evenings on them. The free-tool worksheet in this series shows how many evenings that is. A clinic may have a regulator who wants to see an audit trail and a support contract. In that case, it is right to buy a tool that provides both and to pay for the vendor's name on the paperwork. A desk whose three people are fully occupied is right to pay for a vendor's hours instead of its own. And a tool that fits is simply a cost: three seats, the features the desk uses, a term the organisation is certain of. Costs are fine. There is one disclosure, since this paragraph speaks well of paid software. The publisher of this library sells a commercial, self-hosted help desk, QuerySolv. So weigh the paragraph with that in mind and check it against your own counts.

The expensive cases in this article are not commercial software failing. They are fit failing. The right tool for a desk of twelve is bought by a desk of three. The right bundle for a team that runs change management is bought by a team that does not. The right term for a certain future is bought on an uncertain one. Fit is checked with a count, and a count takes an afternoon. Vendor optimism is a renewable resource; the budget is not.

A Short War Story: The Suite at Harbourview Clinic

Farah had inherited the suite from her predecessor, Rob, who had chosen it in good faith. In the year Rob chose it, Harbourview expected to merge with a neighbouring group and grow to a desk of twelve across five sites. Rob had signed a three-year agreement for fifteen agents, four approvers and the level that included change, asset and catalogue modules. The agreement included six weeks of a consultant's time to configure them. The merger collapsed four months after the consultant left. Rob moved on the following spring. For two years the three remaining agents closed their thirty tickets a day through forms designed for a change process nobody ran. Meanwhile, the invoice arrived each quarter at the size of a desk four times theirs. Farah's count, made in an afternoon after Marcus's printout, showed twelve seats unused, two of them belonging to people who had left. It also showed four approver seats doing work two e-mails a month could do. Four modules had no record raised since the consultant's last day. The vendor agreed to a reduction for the final year in exchange for a renewal at the smaller size. By Marcus's arithmetic, the saving over the remaining term and the next was more than the clinic spent on its backup system. The lesson Farah wrote on the sheet she now fills in every autumn was simple. Buy the desk you have, write down the right to grow, and never let a forecast sign a contract.

The Fit Check

Run this once a year, a quarter before any renewal, and the first time this week if you have never done it. Each section compares what is paid for with what is used, in counts, and ends with the fix. It is deliberately short, because the point is to produce a page a practice manager can read, not a report.

THE FIT CHECK — ONE COMMERCIAL TOOL, ONE PAGE
Tool: ______________   Term ends: ________   Notice date: ________
Checked by: ______________   on: ________   Desk size today: ____ people

SEATS
  Agent seats paid for: ____     People who worked a ticket last quarter: ____
  Seats held by leavers: ____    Seats used fewer than four times last year: ____
  Approver or viewer seats paid for: ____   Approvals last quarter: ____
  Could approvals be done by e-mail reply or a cheaper role?  [ ] yes  [ ] no
  Seats to request at renewal: ____

FEATURES AND MODULES
  Module/feature __________________  last used by ________  on ________  keep [ ] drop [ ]
  Module/feature __________________  last used by ________  on ________  keep [ ] drop [ ]
  Module/feature __________________  last used by ________  on ________  keep [ ] drop [ ]
  Module/feature __________________  last used by ________  on ________  keep [ ] drop [ ]
  Does the bundle level depend on one feature?  which: ______________
  Could that feature be met more simply?  how: ______________________

PROCESSES
  Processes the suite implements: ____   Processes the desk actually runs: ____
  Processes with a named owner who ran them last quarter: ____
  Weight added to an ordinary ticket by unused processes (fields, steps): ____

TERM
  Years signed: ____   Years of certain need at signing: ____   Reason for the gap: __________
  Has the organisation changed shape since signing?  [ ] yes  how: ______________
  Reduction, conversion or exit clause available?  [ ] yes  [ ] no  [ ] not asked yet
  Preferred term at next renewal: ____   Right to add seats at the same rate: [ ] requested

RESULT
  Gaps found: ____   Closable at renewal: ____   Needing a conversation now: ____
  Taken to: ______________  on: ________   Next fit check: ________

Farah's first fit check filled every line in the gaps column and most of the fixes. Her second, a year later, had three seats, no modules, and a one-year term with a written right to grow. It took eleven minutes, most of which was finding the pen.

Right-Sizing Without Ripping Out

Over-buying is rarely fixed by replacing the tool, and the instinct to rip it out is usually the expense talking. A desk paying for twelve unused seats has a seat problem, which a renewal fixes. A desk paying for a bundle it does not use has a level problem, which a renewal fixes. A desk may be paying for a suite's processes it does not run. It can switch the modules off, simplify the forms, and run the suite as the help desk it also is. That takes an afternoon of settings rather than a migration. The tool the desk uses every day is often fine. What was wrong was the quantity, and the quantity is negotiable in a way the habits of three people are not.

Replace the tool only when the fit check shows the gap is the tool itself. The smallest bundle may still be built for a desk three times yours. The simplest ticket may still carry the weight of processes you have switched off. Or the vendor's smallest term may be longer than your certainty. Then the decision is a category decision, and it belongs in Deciding for a Small Team. It needs an exit plan and a migration done at the desk's pace rather than the invoice's. Either way, the conversation with whoever holds the budget is the same one. It goes better as one page of counts than as a complaint. How to brief management in that shape is in Explaining Technical Things to Management. Marcus did not want to know why the help desk was the biggest line. He wanted to know what it would be next year, and Farah's answer fitted on the back of his printout.

Frequently Asked Questions