TL;DR
- Equipment tracking software records what hardware you own, where it is and who has it, and the whole value of it is whether people believe the answer.
- If one person can see every device you own, you do not need software. A spreadsheet they maintain is more accurate than a database nobody updates.
- Tracking works by one of three mechanisms: manual entry, physical tags and scanning, or automatic network discovery. Most fleets need two of the three.
- Discovery finds what is switched on. Tags find what is in a room. Neither finds a laptop in somebody's flat, which is the population that actually goes missing.
- The field that makes a register useful is not the serial number. It is the date somebody last confirmed the record was true.
- Measure accuracy by sampling and publish the number. An unmeasured register gets treated as either perfect or useless, and it is neither.
The Register With 611 Rows and No Confidence
A company had tracking software, a full register, barcode labels on every machine issued since 2024, and an integration pulling device check-ins from its management console. By any description it was doing this properly.
Then a manager asked whether a specific person had been issued a second laptop. It took two days to answer. The register said no. The management console showed two machines checked in under that user's account. One of them had a barcode that did not appear in the register at all, because it had been issued during a period when the person doing the labelling had been on leave and the backlog was entered later from memory.
Nobody had been careless, the software was working, and the answer still took two days, because three systems each held part of the truth and no field anywhere recorded when a human had last confirmed any of it. The register was not wrong so much as unaccountable. You could not tell, looking at a row, whether it described something somebody had verified last month or something typed in two years ago and never revisited.
Best tools for Device Management
That is the actual product requirement in this category, and it is not the one the feature comparisons are about.
When You Don't Actually Need Tracking Software
When the manual way is genuinely fine
Under about fifty devices in one place, with one person who can walk round and see them. A spreadsheet with six columns beats a database here, because its accuracy comes from the maintainer's direct knowledge rather than from a process. Buying software at this scale moves the record further from the person who knows the truth.
When friction starts appearing
The first signal is a question taking longer than it should. How many machines are out of warranty, who has the devices bought in the 2023 refresh, how many laptops are sitting unissued. If answering any of those involves opening three things and cross-referencing, the spreadsheet has stopped being a register and become a list.
When it becomes a liability
The point at which you give an answer and later discover it was wrong. Not a missing device, a wrong statement: telling an auditor you hold 611 devices when you hold 604, or telling a manager somebody has one laptop when they have two. Being wrong confidently is worse than being slow, and it is the thing that destroys the register's standing internally.
The edge case that forces it
An acquisition, which arrives with its own register in its own format, or a refresh, which changes the state of a third of your fleet in a quarter. Both overwhelm a manual process in a single event rather than gradually, and both are foreseeable far enough ahead to prepare.
Five Questions People Ask First
"Is this the same as asset management software?" Broadly the same market, with tracking usually describing the narrower job of knowing where things are and who has them. The wider platforms add procurement, contracts, software licensing and finance. Buy the narrow thing if the narrow thing is your problem, because the wider platforms are configured for a different buyer and cost accordingly.
"Do we need barcode labels?" Only if devices are physically handled by your staff. Labels are excellent in an office with a storeroom and a person who scans things. They are nearly useless for a fleet that ships direct to employees' homes and never passes through your hands, because nobody is ever going to scan them.
"Can it find devices automatically?" Network discovery finds what is switched on and reachable, which is a genuinely useful population and is not the same as the population you own. Discovery will not see a laptop in a drawer, a device with a departed employee, or anything bought and never deployed, and those three categories are exactly where registers go wrong.
"How much should this cost?" Less than people expect for the register alone. There is a capable free and open-source option, several hosted products at tens of dollars a month, and a long tail of enterprise platforms priced by quote. The cost that matters is not the licence, it is whoever maintains the data.
"What should we track beyond laptops?" Start with the devices that hold data and the ones expensive enough to notice, which is usually laptops, phones and a few specialist items. Monitors, docks and keyboards can go in later if they ever matter. A register that tries to track every peripheral from day one is a register nobody finishes populating.
The Three Tracking Mechanisms
Manual entry
Somebody types the record when a device is issued, moved or returned.
Right when volumes are low and the people handling devices are the people maintaining the record. It is more accurate than its reputation when those two groups are the same. It fails the moment they diverge, because anybody updating a record about a device they did not physically handle is working from a message, and messages get missed.
Physical tags and scanning
Barcode or QR labels applied at receipt, scanned at each movement and during audits.
Right for anything with a storeroom and a stock process. The scan is what makes an annual physical audit feasible rather than a fortnight of squinting at serial numbers. It fails for direct-ship fleets, where nothing is ever in your hands, and it fails quietly when labels fall off, which they do.
Network and management discovery
Pulling device identity and check-in data automatically from your management console, directory or network.
Right as a reconciliation source, and it is the single most useful addition to a manual register because it is the only one that updates itself. It fails as a primary source because its population is wrong: it sees managed, powered, connected devices, and it is blind to everything in storage, everything unissued, and everything with somebody who has left. Those are the three categories you most need to know about.
The Field Nobody Adds
Most registers hold what a device is. The ones that get trusted hold when somebody last confirmed it.
A last_verified date, with the method and the person, changes how a register is read. Without it every row looks equally authoritative, which means the register as a whole is only as credible as its worst entry, and nobody can tell which entry that is. With it, a row verified last month is treated as fact and a row last touched in 2023 is treated as a lead.
Four things populate it, in descending order of strength. A physical sighting, which is the strongest and rarest. A scan, during an audit or a movement. A user confirmation, where the holder states they have it, which is weaker than people think and still much better than nothing. And a management check-in, which is automatic and only tells you the device existed and was online, not who has it or where.
Record which of the four produced the date, because they are not equivalent. A register whose rows are all management check-ins is a register that knows its devices are switched on somewhere, which is a different claim from knowing where they are.
Then the useful report falls out for free: everything not verified in the last twelve months, sorted by value. That list is your audit scope, and it is usually a fraction of the fleet rather than all of it.
What the Four Verification Methods Are Actually Worth
The last_verified field is only useful if the method is recorded alongside it, because the four are not interchangeable and treating them as equivalent is how a register looks healthier than it is.
| Method | What it proves | What it does not prove | Effort |
|---|---|---|---|
| Physical sighting | The device exists, in that condition, at that place | Nothing beyond that moment | High, and only feasible for devices you can reach |
| Barcode or QR scan | The labelled device passed a known point | That the label is still on the right machine | Low where there is a storeroom, impossible for direct-ship |
| User confirmation | Somebody says they hold it | That they hold the serial you think they do | Very low, and scales to any geography |
| Management check-in | A device with that identifier was online | Who has it, or where it physically is | None, it is automatic |
So a register whose rows are all management check-ins knows its fleet is powered on somewhere, which is a weaker claim than it looks. And the two cheap methods, user confirmation and check-in, are the only ones that work for a distributed fleet, which means distributed companies are structurally limited to the weaker evidence and should say so rather than pretend otherwise.
The practical combination for a remote fleet is check-in as the continuous signal and an annual user confirmation as the human one. Send everybody a one-line message asking them to confirm the serial on the underside of their machine, which takes each person forty seconds and produces something no automatic source can: a link between a device identifier and a named human who says they have it.
Two cautions on user confirmation. Ask for the serial rather than yes or no, because a yes confirms the person has a laptop and not that they have the laptop in your record. And expect roughly a fifth not to reply, which is itself the finding, since non-responders correlate strongly with the devices that later turn out to be missing.
How to Choose: Five Questions Before You Talk to Any Vendor
Do your devices pass through your hands? This single question decides whether tags are worth anything. Direct-ship fleets should not buy a product whose main differentiator is scanning.
Can it import your existing data without losing it? Every migration drops fields the new system has no column for, and the dropped ones are usually the custom ones that carried your institutional knowledge. Ask to see an import of your actual file, not a sample.
Does it integrate with your management console? Automatic reconciliation is the highest-value integration by a distance, because it is the only one that maintains itself. Ask specifically what it does when the two disagree.
Who will maintain it, and is that in their job description? The most common failure in this category is not a bad tool. It is a good tool with no owner, which decays faster than a spreadsheet because nobody feels responsible for a system.
What does it cost at three times your current device count? Per-asset pricing tiers have cliffs, and a tool chosen at 400 devices can become notably expensive at 1,200 without any change in what you need from it.
The Options
Figures below were read from each vendor's own pricing page, most recently 9 October 2026. The split in this category is clear: register software mostly publishes prices, and anything that also handles physical logistics mostly does not.
Snipe-IT
Best for: teams who want a real register, will self-host, and have no budget.
Why companies choose it: the self-hosted edition is free and open source, and it is a genuine asset register rather than a cut-down one, with check-in and check-out, custom fields, audits and labels. Hosted tiers are published at $39.99 monthly or $399.99 a year for Basic, $99.99 or $999.99 for Small Business, and $249.99 or $2,499.99 for the smaller dedicated plan. Larger dedicated plans are published at $5,000 and $7,500 a year.
Where it struggles: self-hosting is a real commitment, meaning a server, updates and backups owned by somebody. It tracks and does not ship, recover or dispose of anything.
AssetTiger
Best for: small teams wanting a hosted register priced by asset count rather than by user.
Why companies choose it: pricing is published and scales with the fleet, at $20 a month for 500 assets, $40 for 2,500, $75 for 10,000, $140 for 50,000 and $275 for 250,000, with annual billing reducing each to $18, $37, $69, $128 and $252 respectively. Unlimited users on every paid plan, which suits a team where several people update records. An inventory add-on is $15 a month, or $14 annually.
Where it struggles: the 250-asset tier is a 30-day trial rather than the permanent free plan it is often described as, so budget for a paid tier from the start. Like Snipe-IT it is a register only.
Asset Panda
Best for: teams wanting heavy configurability, mobile capture and custom workflows.
Where it struggles: publishes no price, confirmed on its own pricing page 9 October 2026, which is a quote request with a 7-day trial attached. Configurability is also a cost: a highly configured register needs somebody to own the configuration.
Lansweeper
Best for: discovery, meaning finding the devices on your network that never made it into any register.
Why companies choose it: it answers a question the others cannot, which is what exists that you did not know about. As a reconciliation source against a manual register it is genuinely valuable.
Where it struggles: publishes no figures even in a browser. And it is discovery rather than tracking, so it will not tell you who holds a device or where it physically is, only that it appeared on the network.
Freshservice
Best for: teams who want asset tracking attached to a service desk they already run.
Why companies choose it: tiered at $19, $49 and $99 per agent per month, with asset management included at the higher tiers, so the register sits next to the tickets that change it. Its Freddy AI layer is priced separately at $29 per agent per month.
Where it struggles: per-agent pricing means the cost tracks your support team rather than your fleet, which is the wrong variable if you have many devices and few agents.
RemoAsset
Disclosure: RemoAsset is owned by the same people who publish HROpsLab. It appears here because it competes in this category and is assessed against the same criteria as everything else on this page, with its limitations stated in the same detail.
Best for: fleets where the tracking problem is really a logistics problem, meaning devices that ship direct to people and never pass through an office.
Why companies choose it: because it procures and delivers the device, the record is created from the order rather than typed afterwards, which removes the most common source of register error. Returns update the record on receipt for the same reason.
Where it struggles: it publishes no price and requires a demo. It is considerably weaker as a pure register for hardware it did not supply, because then it inherits whatever data you already have. And it is not a certified disposal vendor, so the end of the lifecycle is a separate purchase.
The Comparison
| Tool | Publishes a price | Free option | Discovery | Handles physical logistics |
|---|---|---|---|---|
| Snipe-IT | Yes | Yes, self-hosted | No | No |
| AssetTiger | Yes | No, 30-day trial only | No | No |
| Asset Panda | No | No | No | No |
| Lansweeper | No | Free tier exists | Yes, this is the product | No |
| Freshservice | Yes, per agent | No | Limited | No |
| RemoAsset | No, demo required | No | No | Yes |
The Decision Table
| Situation | Scale | Setup | Primary Pain | Recommended Starting Point |
|---|---|---|---|---|
| One office, one person knows the fleet | Under 50 | Spreadsheet, six columns | Nothing structural | Keep the spreadsheet. Add a last-verified date |
| Register exists, nobody trusts it | Any | Sampling plus a verified-date field | Every row looks equally authoritative | Add the field and sample before changing tools |
| Office with a storeroom and stock movement | 50 to 500 | Hosted register plus labels and scanning | Manual entry lags physical movement | AssetTiger or Snipe-IT, with labels applied at receipt |
| No budget, technical team available | Any | Self-hosted register | Cost | Snipe-IT self-hosted, genuinely free and genuinely capable |
| Devices ship direct to homes, never seen | Any | Platform that creates the record at order | Labels and scanning are worthless here | A logistics platform, since tags cannot help |
| Suspect devices exist that are not recorded | Any | Discovery tool as a reconciliation source | Unknown unknowns | Lansweeper or equivalent, reconciled into the register |
| Asset data needed next to the service desk | 500 plus | Service platform with assets included | Two systems disagreeing | Freshservice, costed per agent rather than per device |
What to Actually Record
Registers fail in two directions: too few fields to answer a question, and so many that nobody fills them in. The workable set is smaller than most templates suggest.
The six that earn their place. Serial number, which is the only identifier that survives a reimage and a rename. Model, so you can group and compare. Assigned person. Location, meaning the country at minimum. Issue date, which drives refresh and residual value. And the verified date with its method.
Three more if they apply to you. Purchase cost, if finance needs the register for fixed assets. Warranty expiry, if you actually act on it rather than just recording it. And the data obligation state, meaning whether a device that held company data has an evidenced wipe, which is the field a security review asks for and almost nobody has.
What to leave out at the start. MAC addresses, detailed specifications, accessory inventories and anything your management console already holds and can be joined on demand. Duplicating console data into a register means maintaining it twice and watching the two disagree, which actively reduces trust.
One field that is not obvious and pays for itself: the asset's current state as a short controlled list rather than free text. In use, in stock, in repair, awaiting disposal, disposed, unaccounted. Free text here produces eleven spellings of the same state within a year and makes every report unreliable. A controlled list of six values is the difference between a register you can query and one you can only read.
The test for adding any field is whether somebody will act differently because of it. A field nobody acts on is a field nobody maintains, and an unmaintained field is worse than an absent one because it looks like information.
One more principle worth holding against vendor demos: every field you add increases the chance that any given row is partly wrong, because completeness and accuracy pull against each other. A six-field register that is 95 per cent accurate answers more questions usefully than a twenty-field register at 60 per cent, even though the second looks far more impressive in a demonstration. Start narrow, prove the discipline holds, and add fields only when a real question has gone unanswered twice. Most registers that collapse under their own weight did so because somebody specified them in a planning meeting rather than growing them from questions people actually asked.
Running the Reconciliation
A register becomes trustworthy through reconciliation rather than through data entry, and there are three sources worth checking it against. None of them is complete, which is the point: the gaps between them are the findings.
Your management console. Devices checked in within the last thirty days. Anything in this list and not in the register is a device you own and never recorded. Anything in the register as active and absent here for ninety days is either in storage, broken, or with somebody who left.
Your finance or purchase records. Every device bought in the period. Anything purchased and never registered is the supplier-to-register handoff failing, and it is the most common single source of missing rows.
Your HR leaver list. Everybody who left in the period. Cross-referenced against assignments, this produces the list of devices that should have come back, which is usually the most uncomfortable of the three outputs and the most actionable.
Run all three quarterly, and record the count of discrepancies found by each. That number tells you which process is leaking, which is more useful than the total. A quarter where finance reconciliation produces fourteen unregistered devices and the leaver list produces one is telling you to fix procurement, not recovery.
A fourth source is worth adding once the first three are routine: your own shipping or courier records. Anything despatched to an employee address in the period should have a matching assignment, and the mismatches are usually replacements issued in a hurry where nobody closed the record on the machine being replaced. That particular gap is how one person ends up holding two laptops with only one of them recorded, which is exactly the question that took two days to answer in the example above.
Migrating Without Losing What You Know
Most companies arrive at tracking software already holding data, and the migration is where institutional knowledge quietly disappears.
The columns that get dropped are the ones that mattered. A spreadsheet accumulates custom columns over years: a note saying a machine has a flaky port, a flag that one laptop belongs to a specific project, a comment that somebody is leaving in March. New systems have no field for any of this, the import maps what it recognises, and the rest is left behind in a file nobody opens again.
So before importing, go through the columns and decide for each one whether it is reference data, which the new system should hold in a custom field, or working notes, which should be preserved as a free-text note against the asset. Doing this costs an hour and it is the difference between a migration and a reset.
Import into a staging state rather than straight into the live register. Bring everything in marked unverified, then promote rows as they are confirmed, rather than importing 600 rows that all appear authoritative on day one. This is the single best opportunity you will ever have to establish the verified-date discipline, because every row genuinely is unverified at that moment and nobody can argue with it.
Keep the old file, read-only, for a year. You will need it, usually for a question about something that happened before the migration, and the instinct to delete it once the new system is live is strong and wrong. Mark it clearly as superseded with the date, because the failure mode here is somebody finding it in eighteen months and updating it instead of the real system.
And do not migrate the devices you cannot account for. The temptation is to bring across every row so the totals match. A register that starts with 40 rows you know to be wrong is a register that starts without credibility. Import what you can stand behind, keep the rest in a separate list titled unaccounted, and work that list down as a project with an owner.
What Getting This Wrong Costs
The direct cost is licences and labels, and in this category it is genuinely small. The free option is real and the hosted ones are tens of dollars a month.
The second cost is the decisions. Refresh budgets built from an age field that does not match the machines people hold. Warranty renewals bought for devices already disposed of. Insurance valued on a list that overstates the fleet by 7 per cent. None of these appear as a tracking failure; they appear as variance.
The third is the one that ends the programme. A register that gives a wrong answer once gets double-checked forever afterwards, and the double-checking is the real cost: every question routed through it now takes two days instead of two minutes, which is exactly the opening example. Credibility is the asset here, and it is lost in a single confident wrong answer rather than gradually.
The fourth cost falls on the people who have to work around an untrusted register, and it is larger than it looks. Teams build private spreadsheets. A manager keeps a list of their own team's devices because the central one was wrong once. Somebody in finance maintains a parallel fixed-asset sheet. Each of these is individually reasonable and collectively corrosive, because now there are five registers, they disagree, and reconciling them is a job nobody owns. The shadow registers are also invisible until somebody leaves and their spreadsheet goes with them.
So the question worth asking of any tracking tool, before any feature comparison, is how it will let somebody tell a true row from a stale one.
Who Maintains It, and the Job That Does Not Exist
Every failure in this category eventually reduces to the same thing: the register has a tool, a schema and no owner. It is worth being specific about what that ownership actually involves, because "somebody should keep it updated" is not a job description and does not survive a busy quarter.
The role is small and it is not zero. For a few hundred devices it is roughly half a day a month: running the three reconciliations, chasing the discrepancies each one produces, and sampling twenty rows. That is a real commitment and it is much less than people fear, which matters because the usual reason it goes unassigned is an assumption that it is a large job.
It should sit with whoever handles devices, not whoever is good at spreadsheets. The most common structural mistake is giving the register to an analyst who never touches hardware, because they are then accountable for the accuracy of records created by other people's actions. Give it to the person who receives returns and issues machines, and the record changes when the thing moves.
Name a deputy. A single owner means the register stops being maintained during annual leave and parental leave, and three months of drift takes considerably longer than three months to repair.
Give them the authority to refuse. The person maintaining the register needs to be able to say that a device cannot be issued until it is recorded, which is a small piece of authority that prevents the single largest source of error. Without it they are documenting other people's shortcuts, and the shortcuts always win because the person taking them has a deadline and the person recording has a preference.
Protect the half day. The reconciliation work is the first thing dropped when something urgent arrives, and because nothing breaks when it is skipped, one skipped month becomes four. Put it in a calendar with a named owner rather than treating it as something that happens when there is time, since there is never time and the decay is invisible until somebody asks a question you cannot answer.
And publish the accuracy figure somewhere other people see, quarterly. It converts an invisible maintenance task into a number with a trend, which is the only thing that reliably keeps it resourced.
When You're Ready to Move Beyond the Spreadsheet
The honest position is that spreadsheets fail later than vendors suggest and for a different reason than they claim. They do not fail because they lack features. They fail when the person maintaining the sheet stops being the person who sees the devices, at which point every entry becomes second-hand and nobody can tell which entries are current.
That transition is about visibility rather than headcount, which is why a 90-person company across four countries usually has a worse register than a 300-person company in one building. If your devices are in places your maintainer cannot walk to, you have already crossed it.
The sequence that works costs almost nothing and does not start with buying anything. Add a last-verified date with a method to whatever you have now. Sample twenty rows, verify them physically or against a management check-in, and write the accuracy percentage down. Reconcile against your management console, your purchase records and your leaver list once, and count the discrepancies from each. Then go and look at tools, knowing which of your three processes is leaking and what your accuracy actually is, which turns a feature comparison into a specific question about your own problem.
Frequently Asked Questions
What is equipment tracking software?
It is software that records what hardware an organisation owns, where each item is and who currently holds it, usually with a history of movements and an ability to audit the list against reality. The feature lists across the category look similar, and the thing that separates a useful implementation from a useless one is not features but whether the data is trusted, which comes down to whether anybody can tell a recently verified record from one typed in two years ago.
Do we need barcode labels and scanning?
Only if your devices physically pass through your team's hands at some point. Labels and scanning are excellent where there is a storeroom, a receiving process and somebody who scans items in and out, because they make a physical audit feasible rather than a fortnight of reading serial numbers. For a fleet that ships direct from a supplier to an employee's home and never returns to an office, nobody will ever scan the label, so that capability is worth nothing to you.
Can tracking software find devices automatically?
Network and management discovery can find devices that are switched on and reachable, which is useful and is a different population from the one you own. Discovery is blind to anything in a drawer, anything bought and never deployed, and anything held by somebody who has left, and those three categories are precisely where registers go wrong. Use discovery as a reconciliation source against a maintained register rather than as the register itself.
What is the most important field in an asset register?
The date somebody last confirmed the record was true, together with how they confirmed it. Without it every row appears equally authoritative, so the credibility of the whole register is set by its worst entry and nobody can identify which that is. With it, you can treat a recently verified row as fact and a stale one as a lead, and you get a useful report for free: everything unverified in twelve months, which becomes your audit scope.
Is there a genuinely free option?
Yes. Snipe-IT's self-hosted edition is free and open source and is a real asset register rather than a limited one, with check-in and check-out, custom fields, audits and label support. The cost is self-hosting, meaning a server, updates and backups that somebody has to own. Be careful with other products described as free, since AssetTiger's widely cited 250-asset tier is a 30-day trial on its own pricing page rather than a permanent free plan.
How do we check whether our register is accurate?
Sample rather than count. Pick twenty rows at random, not by judgement, and verify each one by physical sighting or a management check-in, then express the result as a percentage and write it down with the date. Correcting those twenty rows is worth very little on its own, since the remaining rows have the same error rate, so the value of the exercise is in identifying which process is producing errors and fixing that instead.
What should we reconcile the register against?
Three sources, quarterly. Your management console, where anything checked in but unregistered is a device you own and never recorded. Your purchase records, where anything bought and never registered shows the procurement handoff failing, which is the most common cause of missing rows. And your leaver list, cross-referenced against assignments, which produces the devices that should have come back. Record how many discrepancies each source produced, because that tells you which process to fix.
HROpsLab takes no vendor money and publishes no paid placements, which is why the owner's own product is listed above with its three limitations rather than its three best features.