Security Guard Company Software Problems: The 7 That Cost Real Money, and How to Fix Them

Why does the fill a post screen get built without the money on it?

Because coverage and cost live in different systems and, in most guard companies, in different heads. The dispatcher owns coverage. The controller owns margin. The scheduling tool was designed to answer the dispatcher's question, so it shows availability and hours, and it stops there.

TrackTik will show the schedule and can flag overtime. What it does not know is that this contract is billed at a flat rate, that this officer carries a shift differential, that the client on the next site has an approved overtime clause and this one does not. So the decision gets made at midnight against the only information present, which is who answers the phone.

The fix is a screen, not a report. Rank available officers by true cost to cover, reading hours to date, pay rate, the post's bill rate and the contract's overtime terms, and show the margin impact of each candidate before the assignment is confirmed. Add a hard rule that blocks an assignment which pushes a flat bill post into unbillable overtime, or requires a supervisor override with a logged reason.

Build that first. It is the smallest piece of software in the category that pays for itself, and it is the one that changes behaviour on the night rather than explaining it two weeks later.

What goes wrong when you migrate posts, rates and credentials?

Rates are the trap. In most companies the authoritative bill rate lives in a signed contract, the working bill rate lives in a spreadsheet, and the rate actually invoiced lives in your accounting system. They disagree more often than anyone expects, usually because an annual escalation was applied in one place and not another, or a site was added mid term at a rate somebody agreed verbally.

Migrate the spreadsheet and you have made the disagreement permanent and automated. The first invoice run then produces numbers that are wrong in a new and confident way, which is worse than being wrong in a familiar one.

Reconcile before you migrate. Take every active site and compare the contract rate, the scheduling rate and the last invoice line, and resolve the differences with whoever owns the account. Expect to find revenue you have not been billing, which is usually the first return the project produces.

Credentials need the same treatment. Most companies hold guard cards and certifications in a spreadsheet with an expiry date and nothing else: not the issuing state, not the class, not a copy of the document. Migrate what you can evidence, mark the rest as unverified, and require verification before an officer is eligible for a post that needs that credential. Then run one full pay period in parallel and reconcile both sides before you switch, because a bad first payroll costs more trust than the build costs money.

Why do payroll and accounting integrations break after launch?

Payroll breaks on rules rather than on data transfer. The file format is the easy part. What changes constantly is the logic that turns a clocked hour into a pay line: rounding conventions, grace periods, holiday premium, shift differentials, union step increases that take effect on an anniversary nobody flagged. When that logic lives partly in your system and partly in the payroll provider's configuration, a change in either place silently produces a different answer.

Accounting breaks on coding. Each client wants their own purchase order references and cost centres, and those change when the client reorganises. An invoice export mapped once at go live starts producing rejected invoices six months later, and the fix happens in a spreadsheet because the deadline is today.

Own the rules in one place. A single engine should take a clocked hour to both a pay line and a bill line, applying your rounding, differentials and step tables, and produce the payroll export and the invoice basis from the same source. Then reconcile automatically every period, with discrepancies surfacing as an exception queue during the period rather than as a mystery at close. Confirm before you sign that the developer has shipped working connections to your specific payroll provider and accounting system, because this category lives or dies on those two.

What happens when certified payroll and licensing rules are not covered?

They get treated as a reporting feature and added later, and both of them are gates rather than reports.

On government work under the Service Contract Act or Davis-Bacon, the wage determination and the health and welfare rate by job classification are not a calculation you apply at the end. They determine what an officer must be paid for the hours they worked, so getting them wrong is a compliance finding rather than a rounding error, and it is discovered by an auditor rather than by you. A system that computes ordinary payroll and then produces a certified report from it will produce a certified report that is wrong.

Licensing has the same shape with a sharper edge. A guard card expires on the 14th, the officer works the 15th, 16th and 17th at an armed post, and you have billed a client for an unlicensed officer at an armed site. That is a contract breach, an insurance problem and in some states a regulatory one. Tracking in a generic tool is a date field with a reminder, disconnected from scheduling, so it does not stop the assignment.

Make both structural. Wage determinations, health and welfare rates, holiday premium rules and union step tables become first class rules that drive pay and bill together. Credentials become a gate: each post carries the licences, certifications and training hours it requires, each officer carries current credentials with expiry dates, and anyone lapsed or inside the warning window drops off the eligible list with the reason shown.

Should you build custom or configure what you already own?

Under roughly a hundred officers on mostly standard commercial contracts, straight time hourly billing, no certified payroll and no union agreement, stay bought. TrackTik plus a competent controller and a payroll processor is genuinely the right answer at that stage, and building would waste money you need for recruiting.

WinTeam and Celayix exist for good reasons and are solid choices for standard back office scheduling, payroll and billing. Many companies never need more than one of them configured properly, and the honest question is not which tool is best in general but whether your specific rules fit inside a packaged tool's switches.

Keep what works even when you build. TrackTik is strong for scheduling, post orders and guard tours, so if that layer serves you, leave it and build the margin and compliance layer alongside it. Absorbing it later is a decision you can make once the new system has proved itself.

Build when the signals stack up: past a couple of hundred officers, spreadsheets that have become load bearing systems only one person understands, per guard licence fees that now rival a developer's time, government or union contracts that packaged payroll cannot express, and margin leaking in the gap between three tools that will not talk to each other. The tell is simple. When your advantage lives in how you schedule, bill and prove coverage, and the packaged tool forces you to run that advantage in a spreadsheet, the spreadsheet is the product you should own.

How do hidden costs get into the quote?

Guard company quotes go wrong in six predictable places, and each one is a real body of logic rather than a checkbox.

• Certified payroll is quoted as a report. Wage determinations and health and welfare rates by job classification drive pay itself, which is a different piece of work.
• Union rules are quoted as configuration. Step tables, seniority and premium rules encoded from an actual agreement are a rules engine.
• Multi state licensing is treated as one credential list. Classes, renewal rules and armed requirements differ by state and each variation is its own logic.
• Offline first mobile is assumed to be an online app with retries. Dead zone posts need local capture with device timestamps and geostamps that survive a delayed sync.
• White labelled client portals are quoted as a theme. Per client report formats and branded coverage evidence are content and configuration work that recurs with every large account.
• Integrations are counted as two. Payroll, accounting and sometimes access control or alarm systems are each their own project with their own testing.

The honest bands from Digital Heroes delivery experience are $60,000 to $130,000 over 12 to 16 weeks for a focused first release such as margin aware dispatch with a clean payroll export, and $150,000 to $400,000 phased over 6 to 12 months for a full platform covering scheduling, offline guard tours, credential gating, certified payroll, client portals and accounting integrations.

What separates a build that works from one that fails here?

Whether the developer can model the domain before they quote. Ask a candidate to whiteboard how bill rate, pay rate, differentials and overtime relate to a single clocked hour, and how a post's required credentials gate an assignment. If they treat officers as generic field workers, keep looking, because everything expensive in this category lives in those two relationships.

Second, offline tour capture has to be real. Scans, photographs and daily activity reports captured on the device, timestamped and geostamped locally, syncing when a signal returns. A basement checkpoint that cannot reach a tower is otherwise indistinguishable from a checkpoint that was never walked, and that difference is what a renewal conversation turns on when a client emails on Monday about a 3am scan.

Third, service level credits belong next to the operational data that triggers them. A missed post the system already recorded should propose the contract credit on the next invoice, rather than living in an email thread until the client remembers it and you issue a manual adjustment.

Fourth, phase it so value arrives early. Ship coverage and dispatch first because that is where daily margin leaks, then certified payroll, credential gating and client portals. You should not be waiting a year for the first useful thing.

Fifth, get ownership in writing before work starts: the repository, the deployment and a documented data model in your company's name, with continuity of support as a written term. At Digital Heroes the client owns all of it from the first commit. Leaving per guard licensing only makes sense if what replaces it is an asset you control.