Home › Topics › How to Choose a Network Monitoring Tool › Total Cost of Ownership

Total Cost of Ownership: The Price Tag Is the Small Number


When a monitoring tool gets approved, the conversation is about one number: the license or subscription price on the quote. It is the number on the purchase order, the number finance asks about, and the number everybody remembers. It is also, in most cases, the smallest number you will pay. The larger ones arrive quietly, in hours rather than invoices. They include the week it took to deploy, the evenings spent learning it, and the monthly care it needs to stay accurate. There is also the module you discovered you needed in month four, and the effort of leaving when the time comes. None of them is on the quote, because none of them can be.


This guide is about seeing all of those numbers before you choose. So the choice is made on what a tool will really cost your organisation to own over several years. You will learn the eight cost categories that matter and a worksheet method for estimating them without pretending to a precision you do not have. You will learn how to compare candidates on total cost instead of sticker price. You will also learn how to present the result to management in a way that survives the question "why not the cheaper one?"


In this guide

  • Why the price tag is the small number
  • The eight cost categories
  • The worksheet method: hours times what an hour costs
  • The worksheet template
  • Comparing candidates on total cost, not sticker
  • Where estimates go wrong
  • Presenting total cost to management
  • Where to go next

Why the price tag is the small number


Total cost of ownership, usually shortened to TCO, is the sum of everything a thing costs you. It runs from the day you decide to get it until the day you have fully stopped using it. For a monitoring tool that includes money that leaves the building (licenses, hardware, consultants) and time that leaves your team (deploying, learning, maintaining, integrating, migrating). The time is the part people skip, because it does not appear on any invoice. Another reason is that admitting "this will take me sixty hours to set up" feels like a confession. But your hours are real and finite. And every one spent nursing a monitoring server is one not spent on the project your manager keeps asking about.


The horizon matters too. Looking at one year flatters tools with low entry costs and heavy ongoing care. Looking at three years is the usual compromise. It is long enough for the first big upgrade, the first growth in device count, and the first "we need the add-on module" conversation to appear. But it is short enough that the estimates are still believable. Whatever horizon you pick, use the same one for every candidate.


The eight cost categories


Walk through these in order for every tool on your shortlist. Some will be near zero for a given candidate. Write the zero down anyway, with the reason, because a zero you can explain is evidence and a blank is a guess.


1. License or subscription

The obvious one, but read the quote for what it counts. Some tools charge per device, some per sensor or check, some per interface, some per user of the console, some a flat rate per server. The counting unit decides how the cost behaves when your network grows. That is why the same tool can be cheap for one organisation and painful for its neighbour. Estimate the count for each of your three years, not just the first, and note anything that resets or renegotiates at renewal. The guides under licensing models and hidden costs go deep on how these structures behave. For the worksheet you need only the projected figure and its basis.


2. Hardware or hosting

Where will the central server live, and where will the collectors live? A virtual machine on an existing host costs a share of that host, plus storage that grows with the history you keep. A dedicated box costs the box. A hosted service costs the subscription, plus whatever you need on-site to reach devices that the service cannot see from outside. Do not forget the database. Some tools bring their own; some want a database server you already run and know how to back up. The difference shows up here and in the maintenance category below.


3. Deployment time

Installing the software is the short part. Deployment means discovering devices, fixing the credentials that discovery reveals to be wrong, and adding the devices it did not find. It means setting thresholds, building the dependency tree, configuring notifications, and creating the views the help desk and management asked for. Then it means living through the noisy first weeks while you tune it all. Estimate it in hours of your team's time, and if consultants are involved, in their hours too. A trial on your own network is the only reliable way to size this. The first ten devices tell you what the next two hundred will cost.


4. Training and the learning curve

Formal training, if any, is easy to price. The learning curve is harder and larger. It includes the weeks during which everyone is slower because the tool is new. It includes the questions that go to the one person who understands it, and the alerts that get misread. Count hours for the administrator, and count them again for each help-desk or on-call person who will use the tool. A product that is powerful but unfamiliar can have a learning-curve cost that dwarfs its license. And the best-documented product in the world still costs the hours spent reading the documentation.


5. Ongoing maintenance hours

This is the category most often left blank and most often the largest. Upgrades and the testing around them. Database housekeeping. Re-tuning thresholds as the network changes. Adding the new switch, the new site, the new server. Investigating the collector that stopped collecting. Renewing the certificate on the console. Some tools need an hour a month; some need a day a week; and the difference over three years is enormous. Ask every reference customer one question: "How many hours a month does it take to keep it healthy?" and believe the highest answer you hear. The lowest answer usually comes from someone who has not upgraded yet.


6. Integration work

Connecting the tool to your ticketing system, your chat channels, your paging rota, your documentation, your inventory. The catalogue may say the integration exists. The cost is the hours to make it work with your ticketing system's fields and your escalation rules. Add the hours to fix it when either side changes. Count only the integrations you genuinely intend to build in the horizon, but count all of those honestly.


7. Add-on modules discovered later

Many products sell a core and a set of extras: flow analysis, configuration backup, application monitoring, a higher tier of reporting, more probes, more history. You will not know which extras you need until you have lived with the core for a while. That is exactly when it is hardest to say no. During the evaluation, ask for the full catalogue of extras and their counting units, then mark the ones your requirements list already implies. Those go in the worksheet now, not as a surprise in year two.


8. Exit cost

Every tool is eventually replaced (yes, this one too), and leaving is never free. Exporting history in a usable form, rebuilding thresholds and dependencies elsewhere, retraining everyone, running two systems in parallel for a while, and untangling the integrations you built. Tools with open data formats, documented exports, and standard collection protocols are cheaper to leave; tools that store everything in a proprietary shape are not. The guides under migrating monitoring platforms describe what a migration actually involves. For the worksheet, estimate the hours and put them in year three. Pretending you will never leave is how exit cost becomes a trap rather than a line item.


The worksheet method: hours times what an hour costs


The method rests on one idea: time has a price, and you can estimate it. Every category above is either money out (which you get from quotes) or hours in (which you estimate). To put hours on the same footing as money, multiply them by your organisation's loaded hourly cost. That is the full cost of employing a person for an hour. It includes salary, benefits, overheads, and the equipment they use, divided by the hours they actually work. Finance can give you this figure, or a sensible approximation of it, and it is usually notably higher than the raw hourly wage. You do not need the exact number; you need the same number applied to every candidate.


The steps are simple enough to do on paper:


  • Fix the horizon and the growth assumptions. Three years, and a written guess at device counts, sites, and headcount for each year. The same assumptions apply to every candidate.
  • Gather the money-out figures. Gather license or subscription per year from the quote using your projected counts, hosting or hardware, and formal training. Include consultant time if any, and the add-ons your requirements already imply.
  • Estimate the hours-in figures. Deployment, learning curve per person, maintenance per month, integration work, and exit. Use the trial to measure deployment and maintenance rather than guessing where you can.
  • Convert hours to cost. Hours multiplied by the loaded hourly rate. Keep the hours visible alongside the converted figure; management often finds "two hundred hours" more persuasive than the equivalent sum.
  • Add it up per year, then across the horizon. One total per candidate, with the money-out and hours-in subtotals shown separately.
  • Write down every assumption. A total is only as trustworthy as the sentence that explains each line.

Ongoing maintenance, per year:
  hours per month  x  12  x  loaded hourly cost

Example shape (your figures go here):
  6 hours/month x 12 = 72 hours/year x rate = yearly maintenance cost

Learning curve, per person:
  hours of reduced productivity  x  loaded hourly cost
  (then multiply by the number of people who will use the tool)

Three-year total for one candidate:
  sum of all money-out lines
  + sum of all hours-in lines converted at the loaded rate

The worksheet template


Copy this into a spreadsheet, one sheet per candidate, and fill in the blanks. The "basis" column is not optional; it is where the credibility lives. A worksheet with the basis column blank is a wish with borders.


Cost categoryYear 1Year 2Year 3Basis of estimate
License or subscription   Quote at projected device count per year; renewal terms
Hardware or hosting   VM share, storage growth, database server, probes
Deployment (hours in)   Measured on trial; scaled to full device count
Consultants (money out)   Quoted days, if any
Training and learning curve (hours in)   Per person, times people; new staff in later years
Ongoing maintenance (hours in)   Hours per month from references and trial
Integration work (hours in)   Only integrations planned within the horizon
Add-on modules   Extras already implied by the requirements list
Exit cost (hours in)   Export, rebuild, parallel run; placed in final year
Money-out subtotal    
Hours-in subtotal (hours)    
Hours-in converted at loaded rate   Rate from finance; same for all candidates
Total cost of ownership   Three-year total in the last column

Tip: when you cannot measure a line, estimate a range rather than a single figure — "forty to eighty hours." Carry the range through to the total. A candidate whose total swings wildly with the estimate is telling you something about its predictability, which is itself worth knowing.


Comparing candidates on total cost, not sticker


With a completed worksheet per candidate, the comparison is a single row: three-year total, side by side. But the interesting reading is in the shape of each total, not just its size. Two candidates can arrive at similar totals by completely different routes. One is heavy on subscription and light on hours; the other is the reverse. Which shape suits you depends on whether your organisation is short of money or short of people. A team of one is almost always shorter of people, and should read the hours-in subtotal with particular attention.


The honest part, said plainly: the cheapest tool to buy can be the most expensive tool to own. Consider a product with no license fee at all that needs a day a week of a skilled person's care. It also needs a second server for the database and a consultant to make the alerting behave. That product is not free. A product with a noticeable subscription can be the cheaper owner. That holds if it deploys in an afternoon, needs an hour a month, and the help desk understands it without a course. Neither of those descriptions belongs to any category of tool in particular; there are expensive-to-own products at every price point and cheap-to-own ones too. The worksheet exists precisely because you cannot tell which is which from the quote.


Tollgate Foods replaced its monitoring with a tool that had no license fee. The approval went through in a week because there was nothing to approve. The administrator who set it up spent most of a month on it. Then he spent about a day a week keeping the collectors alive, the database trimmed, and the alert rules from drifting back into noise. Nobody counted the hours, because hours do not arrive as invoices. When he left, the tool went with him, in the sense that nobody else could log in and do anything useful. The replacement project started from a blank worksheet. The finance summary for those three years read "software: nil".


One more comparison is worth doing before you finish: the cost of the incumbent, if you have one. Fill in a worksheet for the tool you already run, using the horizon from today. It often turns out that staying is not the free option it appears to be. And that number belongs in the management conversation alongside the candidates.


Where estimates go wrong


Some errors show up in nearly every first attempt at this worksheet, and knowing them in advance is most of the cure.


  • Optimistic hours. Deployment estimates made before the trial are routinely far too low. Measure the first ten devices, then scale, then add a margin for the devices that turn out to be odd.
  • Counting only the administrator. The learning curve applies to everyone who opens the console. Five help-desk technicians each losing a few hours a week for a month is a large number that never appears if you only count yourself.
  • Forgetting the second year. Year one has the deployment; year two has the first major upgrade, the first growth in device count, and the renewal. Year three has the exit if you are being honest. A worksheet that only looks like year one, repeated, has missed the point.
  • Treating "included" as free. A feature that is included but takes forty hours to configure costs forty hours. The worksheet tracks effort, not entitlement.
  • Ignoring exit cost because leaving feels far away. It is far away right up until the day it is not. The tools that make it cheap are the ones that were chosen with it in mind.
  • Using different assumptions per candidate. If one candidate is costed at today's device count and another at next year's, the comparison is meaningless. Fix the assumptions first and apply them to everyone.

Presenting total cost to management


Management does not want the worksheet. They want one page that answers three questions. What will each option cost us over the horizon? Why is your recommendation not the cheapest quote? How sure are you? Build the page in that order. Start with a small table: one row per candidate (including "keep the current tool" if that is a real option). Show money-out, hours-in converted, and the three-year total. Keep the raw hours visible in a column. A manager who shrugs at a currency figure will often sit up at "that is five weeks of the network administrator's time every year." Nobody shrugs at five weeks.


Then explain the shape. If your recommendation has a higher quote and a lower total, say exactly where the difference comes from. For example: "it deploys in days rather than weeks, and references report an hour or two of care a month rather than a day a week." If your recommendation is also the cheapest quote, say so briefly and move on; nobody needs convincing of that. Where a candidate's total depends heavily on a single estimate, show the range and what drives it. "If maintenance turns out to be double our estimate, the ranking does not change" is a powerful sentence. And "if it doubles, the ranking flips" is an honest one that tells management where the risk lives.


List the assumptions in a short block at the bottom: the horizon, the growth figures, the loaded rate, and where each hours estimate came from. This is not bureaucracy; it is what lets the same page be re-read at renewal to see which guesses were right. Finally, resist the urge to pad the argument with the cost of outages the tool would have prevented. It is a real benefit, but it is a different conversation. And a total-cost page that stays strictly about cost is more credible than one that reaches for justification.


Where to go next


Total cost of ownership is not a trick to make expensive tools look reasonable or free tools look expensive. It is a way of writing down what you already suspect: the hours matter as much as the money. It puts that in a form that lets you compare fairly and explain the result. Fill the worksheet for every shortlisted candidate. Feed the totals into the cost row of your scoring sheet, and keep the pages for renewal. The price on the quote will still be the number everybody remembers. Your job is to make sure it is not the only one written down. The other guides in How to Choose a Network Monitoring Tool cover the requirements that come before this and the demo glitter that inflates the wrong lines. They also cover the playbook that puts it all on a timeline for a small team.


Put this into practice with SysPeep Monitor

The deployment and maintenance lines on this worksheet can only be measured, not guessed. And the fairest way to measure them is a trial on your own network. SysPeep Monitor is one candidate you can measure that way. It is a Windows-hosted monitoring service that can use a database you already run. It discovers devices automatically and offers a free trial. So your hours-in estimate can come from the clock rather than the brochure.

Download Free TrialLearn About SysPeep