1 de noviembre de 2026 • Centoffer Editorial • 19 min de lectura
Per-Dispatch vs. Dedicated Onsite IT Engineer: How to Choose the Right Model
Per-Dispatch vs. Dedicated Onsite IT Engineer: How to Choose the Right Model
Every enterprise that runs physical sites eventually faces the same budgeting argument. Finance looks at the field IT line and asks why it keeps changing month to month. Operations points out that a broken point-of-sale terminal, a dead switch in a comms room, or a new starter’s laptop cannot be fixed from a help desk. Someone proposes “just hiring a person for the site.” Someone else points out that the person would sit idle half the week. The real question underneath all of this is onsite IT engineer cost — and the honest answer is that it depends far more on your demand pattern than on any rate card.
There are two basic ways to buy on-site IT labour. You can pay per dispatch, where an engineer is sent to a site for a specific job and you pay for that visit. Or you can pay for a dedicated onsite engineer, a person assigned to your site who is there whether or not anything breaks. Both models are legitimate, both are widely used, and both can be badly overpaid for when they are applied to the wrong site.
This guide sets out what each model really costs, how to calculate the break-even point between them, which hidden costs change the answer, how each model affects SLA performance, and what to put in the contract. It ends with a step-by-step checklist you can run against your own site list.
What is the difference between per-dispatch and a dedicated onsite IT engineer?
Per-dispatch means you pay for on-site work one job at a time. An incident or request is raised, a qualified engineer is assigned, travels to the site, does the work, and the visit is billed. In Centoffer’s glossary, dispatch is simply “assigning and sending a field engineer to perform an on-site job” — the commercial model wraps a price around that act.
A dedicated onsite engineer means you pay for an engineer’s time at a site, usually by the month. The engineer works a fixed schedule at your premises, takes walk-up requests, handles moves and changes, keeps the comms room tidy, and becomes the person who knows where everything is plugged in. You pay for availability, not for tickets.
Between the two sits a spectrum of hybrids — part-time engineers who attend fixed days each week, scheduled “site days” bundled with ad-hoc dispatches, and dedicated engineers who also cover nearby satellite sites. We return to those later, because for most organisations the right answer is a mix.
The essential trade-off is this:
- Per-dispatch converts on-site IT into a variable cost. You pay only when work happens, but every job carries travel time, a minimum charge, and a delay between the call and the engineer’s arrival.
- A dedicated engineer converts on-site IT into a fixed cost. You get immediate presence and environmental knowledge, but you pay for every hour, including the quiet ones.
Neither is inherently cheaper. Which one costs less depends on how much on-site work a given site generates, how predictable it is, and how fast it needs to be done.
How much does an onsite IT engineer cost under each model?
The cost of an onsite IT engineer under each model is made up of different components, and comparing a per-visit price against a monthly salary without breaking both down is the most common mistake buyers make. The table below lists what actually goes into each.
| Cost component | Per-dispatch model | Dedicated onsite engineer |
|---|---|---|
| Labour | Billed per visit or per hour on site, often with a minimum (for example, a first-hour or half-day minimum) | Monthly salary or monthly contract rate, regardless of ticket volume |
| Travel | Usually charged or built into the visit price; rises sharply for remote sites | Zero for the home site; charged if the engineer covers other sites |
| Out-of-hours work | Multipliers for evenings, weekends and public holidays | Overtime or shift allowance if work falls outside contracted hours |
| Statutory employer costs | Borne by the provider and embedded in the rate | Borne by whoever employs the engineer — you or your provider — and passed through |
| Leave and absence cover | Not your problem; the provider sends whoever is available | You need backfill for annual leave, sick leave and training days |
| Management overhead | Low: ticket approval and invoice review | Higher: scheduling, performance management, timesheets, tooling |
| Idle time | None — you pay only for jobs | Every quiet hour is paid for |
| Spare parts and tools | Varies; often billed separately | Varies; a resident engineer may need a small local stock |
Two points stand out. First, the per-dispatch price you see on a rate card is not the cost of a job — the real cost is the visit price plus travel plus any multiplier, multiplied by the number of visits. Second, the dedicated engineer’s salary is not the cost of a dedicated engineer — statutory contributions, leave cover, management and idle time all sit on top of it.
Why statutory on-costs matter more than buyers expect
When a buyer compares a dedicated engineer against per-dispatch pricing, the salary figure is usually the only number they have. But in most Asia-Pacific markets the employer carries mandatory contributions on top of wages, and those differ by country. A few examples, all from primary government or statutory sources:
- Singapore. From 1 January 2025, the total CPF contribution rate for Singapore Citizens and permanent residents (from their third year of PR status) aged 55 and below, earning above S$750 a month, is 37% of ordinary wages, of which the employee pays 20% — leaving 17% payable by the employer (CPF Board, CPF contribution rates from 1 Jan 2025). This applies to citizen and PR employees; foreign employees on work passes follow different rules.
- Australia. The superannuation guarantee rate rose to 12% on 1 July 2025, applying to all salary and wages paid to eligible workers on and after that date (Australian Taxation Office, 2025).
- The Philippines. Private-sector employers must pay rank-and-file employees a 13th-month payment of at least one-twelfth of the total basic salary earned in the calendar year, no later than 24 December (Department of Labor and Employment, Labor Advisory No. 16, series of 2025, as summarised by PMAP). One-twelfth of annual basic pay is equivalent to roughly 8.3% on top of twelve months’ basic salary — our arithmetic, not a published figure.
These are not the whole picture — health insurance, other social-security schemes, allowances and benefits vary by country and by employer — but they show why the salary is a floor, not the cost. If your dedicated engineer is employed by a provider, these costs are embedded in the monthly rate; if you hire directly, they land on your payroll.
Leave cover is a real cost, not a footnote
A dedicated engineer is not at your site every working day of the year. In Singapore, the Employment Act sets a statutory minimum of seven days of annual leave in the first year of service, rising by one day for each additional year up to 14 days (Ministry of Manpower, written parliamentary answer, January 2026). Many employers offer more. Add public holidays, sick days and training, and there are several weeks a year when the site has no resident engineer.
During those weeks you either accept a gap — which is exactly the risk the dedicated model was meant to remove — or you pay for backfill. A provider that offers a dedicated engineer should state, in writing, who covers absences, how quickly a replacement arrives, and whether the replacement is billed on top.
What hidden costs change the onsite IT engineer cost?
Beyond the line items in the table, several costs rarely appear in a proposal but change the outcome of the comparison.
On the per-dispatch side:
- Repeat visits. If the first engineer arrives without the right part, the right skills, or site access, a second visit is billed. A low first-time-fix rate can quietly double the effective cost per resolved issue.
- Coordination time. Someone on your side has to raise the job, arrange site access, brief the engineer and confirm completion. For a handful of jobs a month this is trivial; at high volume it becomes a part-time role in its own right.
- Business downtime while waiting. A four-hour response window means four hours of a broken till, a dead meeting-room screen, or a warehouse scanner that does not work. The IT invoice does not show that cost, but the business feels it.
- Minimum charges on small jobs. Swapping a keyboard takes ten minutes; a first-hour minimum bills an hour. Lots of small jobs make per-dispatch look expensive per minute of real work.
On the dedicated side:
- Idle capacity. The engineer is paid for quiet afternoons. Some of that time can be filled with proactive work — asset audits, cable management, preparing devices for new starters — but only if someone plans it.
- Skill concentration risk. One engineer cannot be expert in everything. When a job needs network, storage or vendor-specific skills the resident does not have, you dispatch a specialist anyway — and pay both.
- Key-person dependency. Over time, the resident engineer accumulates undocumented knowledge. When they resign, that knowledge leaves with them unless documentation was contractually required.
- Management and HR effort. Timesheets, performance reviews, holiday approvals and replacement hiring take time, whether that time is yours or your provider’s.
A fair comparison prices all of these, not only the headline rate.
How do you calculate the break-even point between per-dispatch and a dedicated engineer?
The break-even point is the monthly number of visits above which a dedicated engineer costs less than paying per dispatch. The calculation needs only two inputs, both of which you can get from quotes and your own ticket history:
- Fully loaded monthly cost of a dedicated engineer (C). Salary or contract rate, plus statutory employer contributions, plus a share of leave cover, plus management overhead, plus any tools or travel.
- All-in cost of one dispatch (D). The visit price at your typical job length, plus travel, plus the average after-hours multiplier across your real mix of jobs, plus an allowance for repeat visits.
Break-even visits per month = C ÷ D.
If a site regularly needs more visits than C ÷ D, a dedicated engineer is likely cheaper. If it needs fewer, per-dispatch is likely cheaper.
A worked example (illustrative numbers, not market rates)
The figures below are invented round numbers chosen to show the method. They are not Centoffer prices or market benchmarks — substitute your own quotes.
Suppose a provider quotes a dedicated engineer for a regional office at a fully loaded 6,000 currency units per month, including statutory costs and leave backfill. The same provider quotes per-dispatch work at 200 units for the first two hours on site, and you estimate travel and occasional after-hours work add an average of 50 units per visit. Your all-in dispatch cost D is therefore 250 units.
- Break-even = 6,000 ÷ 250 = 24 visits per month.
With roughly 21–22 working days in a typical month, that means the site would need slightly more than one visit per working day before the dedicated engineer becomes cheaper on cost alone.
Now look at your ticket history. If the site generated 9 on-site visits last month, per-dispatch costs about 2,250 units — well under the dedicated option. If it generated 35, per-dispatch costs about 8,750 units and the dedicated engineer saves money.
Adjust for what the formula leaves out
The raw break-even is a starting point, not a verdict. Three adjustments usually matter:
- Jobs versus visits. A resident engineer can handle six small requests in a day that would each have been a separate dispatch. If many of your tickets are short, count tickets that need hands on site, not visits, when comparing.
- Value of speed. If a delayed fix has a measurable business cost — lost sales at a store, an idle production line, a trading desk without screens — add the expected downtime cost of waiting for a dispatch to D. That can move the break-even sharply downward.
- Proactive work. If a dedicated engineer would also carry out asset audits, preventive maintenance or onboarding preparation that you currently pay for separately, subtract that value from C.
Run the calculation per site, using at least three to six months of history so a single unusual month does not drive the decision.
When is per-dispatch the right model?
Per-dispatch is the right model when on-site demand at a site is irregular, low, or spread across many locations, and when a response measured in hours rather than minutes is acceptable to the business. It suits:
- Branch networks and retail estates where each location needs on-site help a few times a month, not a few times a day.
- Remote or secondary-city sites where hiring a resident engineer would be hard and leave them idle much of the time.
- Project peaks such as a hardware refresh, an office move, or a network cut-over, where you need many engineers for a short period and then none.
- Specialist work — structured cabling, wireless surveys, data-centre smart hands — where you want the right specialist for each job rather than a generalist who is always there.
- New markets where you do not yet know how much on-site demand the site will generate. Starting with per-dispatch gives you the ticket data to decide later.
The main risks with per-dispatch are response speed and consistency. Both can be managed contractually, which is why the SLA terms matter as much as the price.
When does a dedicated onsite engineer pay off?
A dedicated onsite engineer pays off when a single site produces steady, daily on-site work, or when the business needs someone physically present within minutes rather than hours. Typical cases:
- Headquarters and large offices with hundreds of users, frequent joiners and leavers, meeting rooms that must work, and executives who expect walk-up support.
- Operational sites where downtime is expensive — factories, distribution centres, hospitals, trading floors — where waiting hours for an engineer is not acceptable.
- Environments with heavy access controls where every new visitor needs induction, escorting or security clearance. Repeated dispatches of different engineers into such a site add friction a resident engineer avoids.
- Sites with constant change, such as heavy IMAC volume (installs, moves, adds and changes), where the work itself fills the week.
Even here, the dedicated engineer does not remove the need for dispatch. Specialist tasks, leave cover and demand peaks still call for additional engineers, so the best dedicated arrangements come with an agreed overflow path.
What about a hybrid model?
For most multi-site organisations the cheapest arrangement is neither pure model, but a deliberate mix chosen site by site. Common hybrid patterns include:
- Hub-and-spoke. A dedicated engineer at the largest site also covers nearby smaller sites on scheduled days; anything further away goes to per-dispatch.
- Part-time resident. An engineer attends a mid-sized site two or three fixed days a week. Urgent issues on other days are dispatched. This captures much of the “known face, known environment” benefit at a fraction of the fixed cost.
- Scheduled site days plus dispatch. Routine work — new starters, moves, small fixes — is batched into a fortnightly or monthly site visit, while genuine incidents are dispatched as they arise. Batching cuts the number of minimum charges.
- Dedicated core, dispatched overflow. A resident engineer handles everything they can; specialist tasks, after-hours work and leave cover are dispatched automatically under the same SLA.
The organisational risk with hybrids is fragmentation: different contracts, different SLA definitions and different reporting for each model. The way to avoid that is to buy all three from a platform or provider that measures them on the same scorecard, so you can see when a per-dispatch site has grown into a dedicated one — or the reverse.
Matching the model to the site: a quick reference
| Site profile | Typical on-site demand | Speed requirement | Usually best fit |
|---|---|---|---|
| Large HQ or campus | Daily, steady | Minutes | Dedicated engineer, with dispatch for specialists and leave cover |
| Factory, warehouse, hospital | Daily or near-daily, spiky | Minutes to an hour | Dedicated or part-time resident, plus guaranteed overflow |
| Mid-sized regional office | Several times a week | Same day | Part-time resident or scheduled site days plus dispatch |
| Retail store or branch | A few times a month | Same or next business day | Per-dispatch |
| Remote or secondary-city site | Occasional | Next business day acceptable | Per-dispatch |
| Project (move, refresh, roll-out) | Intense, then zero | Planned | Per-dispatch at scale, scheduled in advance |
Treat this as a starting point; your own ticket data should override it.
How does the choice affect SLAs and service quality?
The engagement model changes which parts of an SLA matter most. A service level agreement (SLA) for per-dispatch work has to measure more clocks than one for a resident engineer, because travel sits between the call and the fix.
For per-dispatch, a meaningful SLA should separately define:
- Dispatch confirmation time — how quickly an engineer is assigned and confirmed.
- Arrival time — how quickly the engineer is on site, ideally tiered by location (metro, secondary city, remote).
- Resolution time — how quickly the issue is fixed, tiered by severity.
- First-time-fix rate — the share of jobs completed on the first visit, which directly controls your hidden cost of repeat visits.
- Proof of delivery — timestamped, geotagged evidence that the engineer was on site and what was done, so billed visits can be verified.
Our guide to IT SLA management in Singapore explains why separating response from resolution, and adding penalty tiers, matters so much.
For a dedicated engineer, travel disappears from the equation at the home site, so the SLA shifts toward:
- Attendance and coverage — contracted hours actually worked, verified through timesheets or attendance logs.
- Backfill time — how quickly a qualified replacement arrives when the resident is on leave or sick.
- Ticket resolution — the same severity-based resolution targets you would set for any provider.
- Documentation — runbooks, asset records and site diagrams kept current, so knowledge does not sit only in one person’s head.
- Customer satisfaction — a short survey at ticket closure, because a resident engineer’s influence on how staff experience IT is large.
A useful rule: if you cannot measure a model, you cannot compare it. Whatever mix you choose, insist that every site reports the same core metrics in the same format.
What should the contract say for each model?
Contracts for onsite IT engineers fail most often on definitions. The following clauses close the gaps that cause disputes.
For per-dispatch agreements:
- Visit pricing structure stated precisely: minimum billable time, increments after the minimum, what counts as travel, and the exact hours that trigger after-hours multipliers.
- Location tiers listing which cities or postcodes are covered at standard rates and response times, and how remote sites are priced.
- Repeat-visit rule — whether a return visit caused by a missing part or missing skill is billable.
- Cancellation terms — what you pay if a job is cancelled after the engineer has been dispatched.
- Evidence requirement — no invoice line without matching proof of delivery.
For dedicated engineer agreements:
- Scope of duties — what the engineer does and does not do, so the role does not quietly expand into unrelated work.
- Contracted hours and attendance evidence — including how overtime is approved and billed.
- Absence and backfill — who provides cover, how fast, at what cost, and with what minimum skill level.
- Replacement process — how a poorly performing engineer is replaced, and the handover period.
- Knowledge transfer — a documented handover pack whenever the engineer changes, owned by you, not the provider.
- Overflow — how additional engineers are dispatched when demand exceeds one person’s capacity, and on what pricing.
For both: severity definitions tied to your own critical systems, service credits for missed targets, a monthly report, and a quarterly review in which the model itself — not only performance — is on the agenda.
A step-by-step decision checklist
Use this sequence to decide the model for every site in your estate:
- List every site with its user count, device count, opening hours and business criticality.
- Pull six months of ticket history and isolate tickets that genuinely needed someone physically on site.
- Estimate the business cost of waiting at each site — even a rough figure for lost sales or idle staff per hour is better than none.
- Get two quotes per model: a fully loaded monthly price for a dedicated engineer and an all-in per-visit price including travel and after-hours rules.
- Calculate break-even (C ÷ D) for each site and compare it with the site’s real visit or ticket volume.
- Adjust for speed and proactive work, using the business cost of waiting and the value of planned maintenance.
- Assign a model to each site — dedicated, part-time resident, scheduled days plus dispatch, or pure per-dispatch.
- Standardise the SLA and reporting across all models so performance is comparable.
- Set review triggers: for example, re-evaluate any per-dispatch site that exceeds its break-even volume for two consecutive months, and any dedicated site whose volume falls well below it.
- Review quarterly, because headcount changes, office moves and new systems move sites across the line.
This process turns a recurring argument between finance and operations into a repeatable decision based on numbers both sides can see.
Common mistakes to avoid
Even with a good framework, a few mistakes recur often enough to be worth naming:
- Comparing a day rate with an hourly rate. Always normalise both models to a cost per month per site.
- Ignoring travel for remote sites. A low visit price can be overwhelmed by travel charges once you leave the main city.
- Assuming a resident engineer covers everything. Specialist work and absence cover still require dispatch; budget for it.
- Deciding once for the whole estate. A single model applied across all sites will overpay somewhere.
- Buying the cheapest per-visit price without a first-time-fix commitment. Repeat visits erase the saving.
- Skipping proof of delivery. Without evidence, you cannot verify what you were billed for under either model.
How Centoffer supports both engagement models
Centoffer is a global IT field services platform — Services OS™ — that connects enterprises and IT service providers with vetted field engineers, operated by Brocent Asia Cloud Service Pte. Ltd. in Singapore. For buyers working through the per-dispatch versus dedicated decision, three parts of the platform are directly relevant:
- Transparent pay-per-delivery pricing for dispatched work. Jobs are matched to vetted engineers and providers, so you pay for delivered work rather than idle time between tickets.
- Dedicated and part-time engineer management. The provider portal manages engineers placed on site, with attendance logs, timesheet approval and FTE/PTE classification consolidated in one place — so dedicated and part-time placements are measured, not assumed.
- One SLA view and proof of delivery across both. Orders flow through one platform with live SLA tracking and proof of delivery, and billing is consolidated, so you can see per site whether a dispatch-heavy location has crossed its break-even point.
If you are weighing the models for a multi-site estate, see how Centoffer’s global IT field services work across countries, or talk to our team and bring your site list and six months of ticket history — the break-even calculation is the fastest way to a decision both finance and operations will accept.