IT Department Outsourcing Thailand 2026 | Where to Draw the Line

IT Department Outsourcing Thailand 2026 | Where to Draw the Line

Ask, at any plant in Thailand, who knows the most about the internal systems, and you almost always get a single name back. That person was not hired as an IT specialist. They are a production engineering manager, or the general affairs section chief, or an expatriate who happened to touch servers in a previous job. Most plants that begin looking at IT department outsourcing start moving from the moment that one name comes to mind.

Two questions dominate the early conversation. How much can we hand over, and what will it cost. But something has to be settled before either can be answered. The question is not what goes out, it is what stays in. That may look like the wrong order, but the scope handed to a provider cannot be defined until the retained side is defined. An outsourcing contract signed without that definition produces, on the day of the first serious incident, an exchange that begins and ends with “that is outside our scope”.

This article takes a clear position. What decides the outcome of IT department outsourcing at a Japanese-owned plant in Thailand is not the provider’s technical skill but the boundary design, meaning what goes out and what stays in. As long as the choice is framed as all-in or nothing, the cost discussion never lands. What follows splits the scope into three layers, shows which benefit each layer protects and what disappears the moment a layer is cut, and works the whole thing through on a costed model plant.

What IT Department Outsourcing Is, and What Changes at a Japanese-Owned Plant in Thailand

IT department outsourcing means handing some or all of the work carried by an in-house information systems department to an external provider. The territory is wide. It covers monitoring of servers and network equipment, taking user enquiries, maintaining hardware, managing software licences and coordinating with vendors. IT maintenance outsourcing, managed IT services, IT operations outsourcing and system monitoring outsourcing are all names for parts of the same territory.

The market keeps expanding. The global IT outsourcing (ITO) market is forecast to grow from 6,386.5 hundred million US dollars in 2026 to 7,520.8 hundred million in 2031, a compound annual growth rate of 3.32%. The small and medium enterprise (SME) segment grows slightly faster at 3.96%, which fits the picture that the harder it is to carry an information systems department in house, the more of it moves outside. Note that this is a forecast for the global market. It is not a figure for Thailand, and it is not a figure for Japanese-owned plants operating there. Read it as background on direction of travel, nothing more.

There are four differences between ordinary IT department outsourcing in Japan and the version that applies to a Japanese-owned plant in Thailand.

Of these, the second row is what twists the cost discussion most. An hour of office mail being down and an hour of the production management server being down cannot sit under the same phrase “system failure”. The second one adds up in money for every minute it lasts. That is why monitoring stops being an IT task and becomes a profit and loss item.

The third and fourth rows feed straight into how a provider is selected. Operators on the floor must be able to raise a question in Thai, and the substance of that question must reach head office in Japanese. Not many providers can satisfy both. In practice, the ability to bridge languages matters more than raw technical depth.

The Double Squeeze | Thailand’s IT Talent Shortage and the One-Person IT Department at Head Office

“Why not simply hire an IT person locally” is the obvious alternative, and it always gets raised. There is a structural wall in front of it.

In the survey of business conditions of Japanese companies overseas that JETRO conducted from August to September 2023 for the Asia and Oceania region, Japanese companies in Thailand reporting a shortage of IT personnel such as programmers as either very serious or somewhat serious came to 56.7% in total. The average across the Asia and Oceania region was 58.2%, so Thailand is not an outlier. The accurate reading is that IT talent is short across the whole region.

There is, however, one item where Thailand sits clearly outside the regional pattern. It is the shortage of general managerial staff. Thailand recorded 79.8% against a regional average of 68.8%, which is 11 points higher. The shortage is not only in IT people but in the layer that would evaluate, place and manage them. Put those two findings together and the difficulty of the hiring option comes into focus. Even if you do hire, an appointment solves nothing when no manager on site can judge the quality of that person’s work.

So the answer is to send someone from head office in Japan. Except that side has no slack either. In a survey Metaps ran from September to October 2021 among 514 company employees working in information systems departments, 11.4% fell into the category of a one-person IT department, where a single person effectively is the IT function. That is roughly 59 people out of 514. Among the 27 respondents who said they felt the strain, the leading concerns were unease about security at 59.3% and being too loaded to keep up at 51.9%.

Two caveats belong with that survey. First, it dates from 2021 and is not a current figure. Second, the base sizes are small, at 514 respondents, roughly 59 of them, and 27 within that group, so the percentages should be read as indicative rather than precise. Even so, the finding that a meaningful share of companies run a thin information systems function, and that the people inside a thin function name security and workload as their problems, is hard to ignore for any company with overseas sites. Expecting a department already at capacity on domestic work to also watch a server in Thailand is not realistic.

There is nobody spare in Thailand and nobody spare at head office. That is the double squeeze. A time dimension sits on top of it. Pay and conditions for locally hired staff in Thailand have been improving, and companies have been reported to be reducing the number of Japanese expatriates in favour of local hiring. Expatriate posts are trending down, and the expatriates who remain rotate out within a few years. A design that depends on the one person you have today lasts exactly as long as that person does.

What Actually Happens in a Plant Where IT Rests on One Person

Key-person risk usually gets described as “we are in trouble if they leave”, which pushes the remedy towards the impossible task of making sure they never leave. The problem that actually bites is the invisible dependency that already exists while the person is still sitting at their desk.

Here are the symptoms you see most often in plants in Thailand.

• The server administrator password exists only in a file on that person’s local PC and nowhere shared
• The network diagram has not been updated, and the current cabling exists only in that person’s memory
• Maintenance contract renewal dates are tracked in a personal diary, and nobody else knows when they fall due
• Vendor correspondence arrives at that person’s individual mailbox, so no history is shared
• Requests from the floor go straight to that person’s chat rather than into any system

All five get more likely the more capable the person is. Asking them fixes it immediately, so nobody bothers to write the procedure down. The result is that this individual’s working hours become the plant’s IT availability. If a server dies overnight, nobody knows until morning. If the person is away on a business trip, it stays down.

A condition specific to Thailand compounds this. When the one person is a Japanese expatriate, they will normally rotate out within a few years, and the handover window is rarely longer than a month. A month of handover, conducted with an outdated network diagram, passwords held locally and vendor contacts sitting in a personal mailbox, transfers only surface information. The successor arrives and starts re-investigating everything the predecessor already knew. The cost of that re-investigation appears nowhere in the accounts, but it is certainly incurred.

When the one person is a Thai employee, a different problem shows up. If they move to another company, there may be no handover window at all. And head office has nobody able to verify what that person had been doing.

The remedy for key-person risk is not to add people and not to retain people. It is to build a structure in which information does not accumulate inside one head. Concretely, that means putting passwords, network diagrams, contract dates, vendor contacts and the asset register outside the individual, and keeping them updated. Keeping them updated is exactly the part that never survives in house, which is where outsourcing starts to earn its place.

Cutting the Scope Into Three Layers | Monitoring, Help Desk, Maintenance

Now to the design. If IT department outsourcing is framed as hand it all over or hand over nothing, the quotation comes back either too expensive or too narrow to matter. In practice, cutting the scope into the following three layers is what lets cost and benefit line up against each other.

The three layers stack. Outsourcing layer 2 on its own is technically possible, but a help desk with no monitoring underneath it still depends on users reporting faults before anyone knows there is one. Outsourcing layer 3 alone gives you a tidy register and nobody moving when something stops. The default shape is therefore to build upwards from the bottom.

The point that matters is that each layer protects a different benefit. Deciding that the budget will not stretch to layer 3 is a legitimate decision, but it has to be made in the knowledge that benefit C disappears in its entirety along with it. A later section puts numbers on how much the payback period worsens as layers come off.

One terminology note. These three layers are not the same thing as the familiar level 1, level 2 and level 3 support tiers. Those tiers divide work by difficulty of the issue. These layers divide work by which functions are outsourced. The two get mixed up easily when quotations are compared, so it is worth agreeing with providers that the conversation is about functional scope.

Layer 1 Basic Monitoring | Buying Down the Time Until Someone Notices

When you outsource system monitoring, the thing you are buying is not a monitoring tool. It is the time between something going wrong and somebody noticing. Get that wrong, and monitoring quotations look expensive.

Plant system failures have time windows in which the delay in noticing is structural. Night shifts, weekends and the hours before the shift starts. If the production management server stops on a Saturday evening, nobody knows until whoever comes in on Monday morning notices. Even if the line was not running, the output data that should have been captured in the interval is simply missing.

Here is the minimum that basic monitoring should cover.

The part of monitoring design that pays off in practice is deciding where alerts go and at what hours. If every alert lands on one person’s mobile, then for as long as that person is asleep you are not monitoring at all. Where the work is outsourced, the provider takes first receipt and only items requiring a decision come inside.

There is one more thing. Monitoring thresholds cannot be set perfectly from day one. False positives appear immediately after go-live, and a run of false positives trains people to ignore alerts, at which point monitoring exists on paper only. The contract should therefore state that the first three months are a threshold tuning period with monthly adjustment. Leave that vague and within six months nobody is reading the alerts.

Layer 2 First-Line Triage and Help Desk | Taking IT Support Off the Expatriate Manager

What goes outside in layer 2 is user support. Taking enquiries such as the printer will not print, the handheld will not scan, the shared folder will not open, and performing first-line triage on them.

The reason this layer looks larger in benefit terms than the raw hours suggest is the cost of interruption. When an expatriate manager is pulled out of a meeting or away from a line walk and resolves the issue in ten minutes, the time lost is not ten minutes. Once the time needed to get back into the interrupted task is counted, the felt cost exceeds the handling time. The model plant figures below put the expatriate manager at 32 hours a month of IT support and the administrative staff member at 18 hours a month, and those are counts of hands-on time only.

There are three design points when a help desk is outsourced.

• It must accept enquiries in the local language. A desk that requires Thai operators to write in English or Japanese does not get used. A desk that does not get used means requests keep landing in the expatriate’s chat exactly as before.
• Intake must funnel into one record. Multiple entry points such as phone, chat and mail are fine, but they must all write to one place. Scattered records make it impossible to see what is happening how often, and without that, no permanent fix can be justified.
• Escalation criteria must be agreed in advance. Not the vague standard of anything the provider cannot judge, but concrete acts, such as any operation that updates production data, any decision that incurs licence cost, or any restart that stops the line.

Some things should be deliberately excluded from scope. Questions about how to use the production management system in a business sense, such as why the stock figure for this part has gone negative, are business questions rather than IT questions. Sent to an outsourced help desk, they sit unanswered. Intake can still be a single channel, but business questions must be routed to the internal process owner, and that traffic rule needs to be set at the start.

Layer 3 Scheduled Maintenance and a Single Vendor Contact Point

Layer 3 is the work you do while nothing is broken. Scheduled inspection, patching, restore testing of backups, and tracking of maintenance contract and licence renewal dates.

Because it is work done while nothing is broken, it is the first thing to slide in house. In a plant where IT rests on one person, it is more often not being done at all. What that produces is missed renewals. Equipment fails while a maintenance contract has lapsed and the emergency call-out is charged at a premium. A licence expires and work stops. Losses like these tend to be written off as bad luck, but with no mechanism tracking the dates, they will happen eventually.

The other important function in layer 3 is consolidating vendor contact into one point. A plant IT environment involves several vendors at once.

Scattered contact points mean somebody has to decide who to call every time something breaks. The person making that call is the same one person. The model plant below puts that coordination load at 6 hours a month. Make the provider the single front door and have them dispatch to each vendor, and those 6 hours leave the plant.

How much is included in scheduled maintenance, where the monthly fee stops and individual quotations begin, cannot be compared without knowing how to read the contract. The breakdown of maintenance costs and how to read a quotation is covered in our article on business system maintenance costs. This article assumes only that scheduled maintenance is inside the layer 3 monthly fee.

Hold the IT Asset Register Jointly With the Provider

Within layer 3, the item that works most directly against key-person risk is the IT asset register. Ownership design matters enough here to warrant its own section.

What belongs in the register is more than a list of equipment. Who is using which device, how many licences of which software are installed, how long each maintenance contract runs, who the contact is at each vendor. Only when that operational layer is complete is information genuinely held outside the individual.

Where opinions divide is who holds the register. There are three patterns.

The third is the recommendation. Send the work of updating outside, keep ownership of the register inside. The contract should state that register data can be viewed and exported by the plant at any time during the term, and is handed back in a general-purpose format when the contract ends. Without that clause, the moment you decide to change provider you are starting the same inventory from zero all over again.

Designing the fields and running the first physical inventory is enough material for an article of its own. The practical build is covered in our article on IT asset management for factories, which is the place to go when you actually start. What concerns this article is one narrow question, which is who updates the register and who owns it.

How Far Backup and Recovery Can Be Handed Over

Separately from the three layers, backup and recovery always comes up. It should be handled not as a question of whether it can be outsourced but as a question of where the line of responsibility falls.

Backup splits into three activities. Taking the copy, storing it, and restoring from it. Taking and storing are easy to place outside, and once they are inside the monitoring scope, job success and failure is picked up automatically. Restoring is the problem.

• Success or failure of the copy is machine-detectable. Design an alert when a backup job fails and the provider can act on it.
• Integrity of storage can be outsourced too. Retention generations, geographic separation of copies and media condition checks all sit comfortably in scope.
• Restoring requires an internal decision. Which point in time to roll back to is a business judgement. Rolling back three hours deletes output records that have already been entered. That judgement cannot be handed to a provider.

The workable line is therefore that taking, storing and testing are outsourced while the decision to execute a restore stays inside. The item most often missed here is restore testing. Backups that copy successfully but cannot be restored are not rare. Once a year is enough, but an actual restore exercise should be written into the outsourced scope.

Design that extends to disaster recovery, including how to set targets for which systems come back within how many hours, is covered in our article on backup and disaster recovery for business systems. Knowing roughly where your own recovery targets sit before you sign makes the scope conversation far more concrete.

Drawing the Line Between Head Office Standards and Local Optimisation

IT department outsourcing at a Japanese-owned plant carries one more distinctive issue. The standards set by head office in Japan do not always fit the local reality.

The head office information systems department defines group-wide security policy, identity infrastructure, standard PC builds, data retention periods and so on. That is correct practice in itself. But the local plant has conditions the standard never anticipated. Handheld terminals on the floor run an old operating system that cannot be updated. The line supplied by a local carrier does not work with the head office standard equipment. Thai law requires certain data to be held locally.

Bringing in a provider makes the mismatch visible. The provider moves to fit local reality, head office asks for conformity to the standard, and local management is caught between them.

Avoiding that requires a document, written before the outsourcing contract is signed, setting out which areas are governed by head office standards and which are decided locally.

How head office standards are applied when a system is first deployed to an overseas site is the deployment-phase question, covered in our article on rolling out systems to overseas plants. What this article addresses is the operating phase that follows. Standards agreed at deployment gradually erode on site once operation begins, which is a very common trajectory. The point at which a provider comes in is also a chance to put that erosion back in order.

What Cannot Go Out | Five Things You Keep and Four You Hand Over

Here is the whole discussion reduced to a boundary. What cannot go outside is not work. It is authority and accountability.

Five things stay inside. Two of them belong to head office and three to the local site.

• Head office keeps, first, investment decisions and approval of the IT budget. How much is spent on which system is a management decision. A provider can propose. It cannot decide.
• Head office keeps, second, the setting of global standards. Security policy, identity infrastructure and data retention rules cannot vary by site.
• The local site keeps, first, definition of business requirements and acceptance. Only the site can define what has to work for the business to run, and the decision to accept what a provider has delivered also sits locally.
• The local site keeps, second, approval of data access rights. Deciding who sees which data requires knowledge of personnel records and organisational structure. Hand that approval to a provider and you get accidents such as leavers’ accounts staying active.
• The local site keeps, third, ultimate responsibility for audit and legal compliance. Dealings with Thai authorities and accountability for personal data handling remain with the local entity regardless of what has been outsourced.

Four things can go to the provider.

• The provider takes, first, monitoring and first notification. Keeping eyes on systems around the clock is more realistic from outside.
• The provider takes, second, first-line help desk response and triage. Local-language intake and the initial response.
• The provider takes, third, scheduled maintenance and upkeep of the asset register. Continuous work performed while nothing is broken does not survive in house.
• The provider takes, fourth, day-to-day vendor coordination. Acting as the front door and dispatching to each vendor.

Attach this five-and-four boundary as a schedule to the contract and the argument about whose scope an incident falls into becomes short. Contract instead for “IT operations, all inclusive” without that boundary and every situation requiring a decision turns into a negotiation.

Costing It on a Model Plant

Now to money. First the general market rates, then the model plant.

In Japan, the cost of IT department outsourcing is presented as a general rule of thumb along these lines. A fixed monthly retainer runs roughly 50,000 yen to 300,000 yen, pay-as-you-go pricing runs roughly 5,000 yen to 10,000 yen per incident or per hour, and a dedicated on-site engineer arrangement runs roughly 100,000 yen to 1,000,000 yen per month. These are not figures from any particular survey. They are presented as a general industry rule of thumb. They are also a Japanese domestic rule of thumb and cannot be converted into a Thai quotation. The model plant calculation below is not a conversion of those amounts. It is built independently from Thai conditions, and every figure in it stays in Thai baht (THB).

Here are the model plant assumptions.

The outsourcing fees are as follows.

The monthly increment column is the additional amount incurred when that layer is added. The actual payment when the scope runs through layer 2 is 30,000 THB a month, and through layer 3 it is 44,000 THB a month.

The benefits break down as follows.

Benefit A, avoided lost profit through early detection of a stoppage. Stoppages traceable to the core server or the network that affected the line ran at 5 a year. Introducing monitoring shortens the time between the fault occurring and someone noticing by an average of 1.5 hours per incident. Applying the hourly lost profit of 24,000 THB to the resulting 7.5 hours a year gives 180,000 THB a year. What is being counted here is only the delay in noticing, not the recovery work itself.

Benefit B, reduced IT support hours. The expatriate manager spends 32 hours a month, or 384 hours a year, on IT support, which at 1,150 THB an hour is 441,600 THB. The administrative staff member spends 18 hours a month, or 216 hours a year, which at 300 THB an hour is 64,800 THB. Of the combined 506,400 THB, the share that can be transferred to a help desk is taken as 65%. The remaining 35% is enquiries that require a business judgement plus physical work that only someone on site can do. Benefit B therefore comes to 329,160 THB a year. Note that the 32 hours and the 18 hours exclude time spent coordinating with vendors, which is counted separately under benefit C. That definition exists to avoid double counting.

Benefit C, removal of key-person risk and prevention of contract and asset losses. Emergency call-outs and premium charges caused by missed maintenance and licence renewals run at 3 a year at 38,000 THB each, or 114,000 THB. Duplicate purchasing and replacement of assets whose location is unknown, both caused by handovers that could not be completed, run at 74,000 THB a year. Coordination time absorbed by the expatriate manager because vendor contact is scattered runs at 6 hours a month, or 72 hours a year, which at 1,150 THB an hour is 82,800 THB. The total is 270,800 THB a year.

Pulling it together gives the following.

What this table shows is that narrowing the scope makes it cheaper and makes the payback slower. Dropping layer 3 frees up 168,000 THB a year in fees, but benefit C of 270,800 THB disappears in full, so annual net benefit falls from 251,960 to 149,160 and payback worsens from 1.27 years to 1.47 years. Drop layer 2 as well and run layer 1 alone, and benefit B of 329,160 THB goes too. Annual net benefit shrinks to 36,000 and payback becomes 3.61 years.

Reducing the number of layers because of a budget constraint is a defensible decision. But it is not a decision to make the arrangement cheaper. It is a decision to slow the payback in exchange for a smaller cash outflow this year. Those are two different things.

Sensitivity | When Help Desk Adoption Slips

The calculation above contains one assumption that fails more often than any other in practice. It is help desk adoption. Build the desk, and if the floor carries on asking the expatriate manager directly, benefit B never materialises.

So adoption becomes a variable. The critical point here is that the factor applies to benefit B only.

• Benefit A, early detection through monitoring, occurs automatically as long as the monitoring agents and the alert path are running. It has nothing to do with whether users go to the desk. The factor is not applied to it.
• Benefit C, upkeep of the register, tracking of contract dates and consolidation of vendor contact, is performed as back-office work by the provider. It is unrelated to adoption on the floor. The factor is not applied to it either.

Even with adoption halved, payback is 3.66 years, which stays inside the range where the investment still stands up. That is because monitoring and the register do not depend on adoption. As a reference point, it is almost the same level as the 3.61 years produced by narrowing the scope to basic monitoring alone. In other words, the worst case degrades only to roughly the same place as the narrowest possible engagement. Read the other way, work that raises adoption is work that pulls payback back towards 1.27 years without spending anything more.

One warning belongs here. The factor must not be applied uniformly across all benefits. Apply 50% adoption to the full annual benefit of 779,960 THB and you get 389,980 THB, which falls below the annual fee of 528,000 THB and produces a net benefit of minus 138,020 THB. That produces the conclusion that outsourcing does not pay, and that conclusion is a calculation error. Adoption measures whether users on the floor go to the desk. It has no bearing on whether a monitoring agent runs at night, and none on whether the provider tracks a contract renewal date. Whenever a factor is introduced into a sensitivity analysis, check every time which benefits actually have that factor in their conditions of occurrence.

The levers that raise adoption sit outside the cost model. Accept enquiries in the local language, keep the time from intake to first response short, and be rigorous about redirecting requests to the desk when the expatriate manager receives one directly. The third is the strongest. If the people who designed the process grant themselves exceptions, the floor reverts to the old route.

Selecting a Provider and Running the Transition

Selection and transition realistically run in five stages.

• Stage one, mapping the current state (two to four weeks). Build a list of servers, network equipment, PCs and contracts. It does not matter if the list is incomplete at this point. The purpose is to record the unknowns as unknowns.
• Stage two, provisional scope decision (one to two weeks). Document how far up the three layers you intend to go, and who holds each of the five items that stay inside. This is the heart of this article.
• Stage three, obtaining and comparing quotations (two to four weeks). Hand the same scope definition to several providers. Ask for a price for “the lot” without a scope definition and the quotations that come back cannot be compared.
• Stage four, transition (four to eight weeks). Monitoring configuration, publicising the help desk and the first asset inventory run in parallel. During this period the current internal owner and the provider are both operating.
• Stage five, embedding (three months). Threshold tuning, correction of enquiry routes, and settling the format of the monthly report.

Here is what to look at when comparing quotations.

The part of the transition that deserves the most care is the parallel running in stage four. If the current internal owner is too busy to hand over completely during that window, the provider starts operating on incomplete information. The parallel period requires a deliberate reduction in that person’s normal workload. Cut that corner and the transition itself reproduces the key-person problem it was meant to solve.

Common Failure Patterns

• Contracting for “IT operations, all inclusive” without defining scope. Every incident turns into a scope negotiation. Simply writing the five retained items and the four transferred items into a schedule to the contract removes almost all of it.
• Installing monitoring but sending alerts to one person. Nights and weekends are unwatched, so the result is no different from before monitoring existed. Unless the provider holds first receipt, benefit A does not occur.
• Building a help desk while the expatriate manager keeps taking requests directly. Adoption never rises and benefit B alone drops out of the calculation. The most common version of this is that the person who built the desk is the one creating the exceptions.
• Letting the provider hold the asset register alone. Updates continue, but nothing is left in your hands at the end of the contract, and changing provider becomes effectively impossible.
• Starting outsourcing without reconciling head office standards against local reality. Every time the provider adapts to local conditions, head office raises an objection, and local management is caught in the middle.
• Applying the sensitivity factor uniformly to all benefits. As the previous section shows, the conclusion inverts. A decision to abandon outsourcing then comes out of an arithmetic error.
• Comparing costs on the monthly fee alone. Without the initial cost, the work that will be separately quoted and the cost of exiting the contract, the real total cannot be compared.

Summary | You Are Buying a Boundary, Not Headcount

What you buy with IT department outsourcing is not headcount. It is a boundary between what goes outside and what stays inside, and a mechanism that keeps that boundary intact.

Japanese-owned plants in Thailand sit inside a double squeeze, with an IT talent shortage locally and a thin information systems function at head office. The JETRO survey put the IT talent shortage in Thailand at 56.7% and the shortage of general managerial staff at 79.8%, the latter 11 points above the regional average. On the head office side, a 2021 survey found that a meaningful share of companies effectively run their IT function on one person. Neither side has anyone to spare, and expatriates rotate out within a few years.

Under those conditions, a design that gives everything to the one person who knows the systems lasts only as long as that person stays. Equally, outsourcing everything does not empty the organisation, because investment decisions, data access approvals and audit responsibility cannot leave. The design question is therefore not whether to outsource, but how far up the three layers to go and how to retain the five authorities.

On the model plant, outsourcing through layer 3 produced an annual net benefit of 251,960 THB and paid back an initial cost of 320,000 THB in 1.27 years. Drop layer 3 and it becomes 1.47 years. Run layer 1 alone and it becomes 3.61 years. The relationship that cheaper is slower is built into the layer structure. Even if help desk adoption falls to 50%, payback stops at 3.66 years, because the benefits from monitoring and from the register do not depend on adoption.

One last observation. Outsourcing is also an exercise in making visible what was never visible before. The absence of a network diagram, passwords held locally, contract dates nobody knows, all surface during the preparation stage without fail. At that moment, half the problem is already solved.

If you are working through where to draw the line for plant IT in Thailand, or simply trying to establish what is missing from your current arrangement, that is a good stage at which to talk. It does not matter if IT still rests on one person, or if the scope you would hand over is not yet decided. We are happy to start from building the list of what you actually have, so please get in touch through our contact form.

Frequently Asked Questions

What is IT department outsourcing?

It is the practice of handing some or all of the work carried by an in-house information systems department to an external provider. The scope covers monitoring of servers and networks, handling user enquiries, hardware maintenance, licence management and vendor coordination. IT maintenance outsourcing, managed IT services and system monitoring outsourcing all refer to parts of the same territory. The important point is that outsourcing the work does not move investment decisions, approval of data access rights or responsibility for audit and compliance. Those stay inside.

What does IT department outsourcing cost for a plant in Thailand?

It depends on the scope. On the model plant used in this article, a maker of automotive interior components in Rayong Province with 480 employees and 210 PCs, layer 1 basic monitoring alone is 12,000 THB a month, adding first-line triage and the help desk to reach layer 2 is 30,000 THB a month, and adding scheduled maintenance, the asset register and the vendor contact point to reach layer 3 is 44,000 THB a month. The initial cost through layer 3 is estimated at 320,000 THB. As a separate reference, a fixed monthly retainer in Japan is presented as a general rule of thumb at roughly 50,000 yen to 300,000 yen, but that is a Japanese domestic figure and cannot be converted into a Thai quotation.

If IT maintenance is outsourced, what is left inside?

Authority and accountability. There are five items. Head office keeps investment decisions and approval of the IT budget, plus the setting of global standards such as security policy and identity infrastructure. The local site keeps definition of business requirements and acceptance, approval of data access rights, and ultimate responsibility for audit and legal compliance. Work can go outside. These five cannot. Writing the boundary into a schedule to the contract keeps the scope argument short when an incident occurs.

How much visibility does outsourced system monitoring actually give you?

Basic monitoring tells you that a device or service has stopped, that a resource has crossed a threshold, and that a backup job has failed. Typical targets are process availability on the production management server, free disk space, port link status on switches, availability of wireless access points and reachability of the line. What it does not cover is business data being wrong, such as a stock figure going negative. That belongs to business-side checking. Note also that unless the provider holds first receipt of alerts, nights and weekends are effectively unmonitored.

We have run for years with no IT staff at all. Is it too late to start?

It is not. The first task is precisely to build the list of what you have. Spend two to four weeks listing servers, network equipment, PCs and contracts, and at this stage record the unknowns as unknowns. Completeness is not required. Next, make a provisional decision on how far up the three layers to go, then hand the same scope definition to several providers and collect quotations. Without fixing the scope definition first, the assumptions behind each quotation differ by provider and no comparison is possible. Starting with layer 1 basic monitoring only and adding layers later is also a perfectly workable path.

References

1. JETRO survey on business conditions of Japanese companies overseas, FY2023, Asia and Oceania edition, Thailand section (conducted August to September 2023)

Source for the IT talent shortage in Thailand at 56.7%, the shortage of general managerial staff at 79.8%, and the regional averages of 58.2% and 68.8%. The 1.5 points cited in this article is 58.2 minus 56.7, and the 11 points is the gap between 79.8 and 68.8.

JETRO regional analysis report

2. Metaps survey on the one-person IT department (conducted September to October 2021 among 514 company employees working in information systems departments)

Source for the one-person IT department share of 11.4% (roughly 59 people out of 514), and for unease about security at 59.3% and workload at 51.9% among the issues raised. Please note that this dates from 2021 and that the base sizes are small.

PR TIMES press release

3. Nikkei Business article on pay for locally hired staff in Thailand and the shift to local hiring

Reference for the trend that pay and conditions for locally hired staff in Thailand have improved and that some companies are reducing Japanese expatriate numbers in favour of local hiring. This article uses it only for the direction of the trend and cites no figures from it.

Nikkei Business feature

4. Market size and growth rate of the IT outsourcing (ITO) market

Source for the forecast of 6,386.5 hundred million US dollars in 2026 rising to 7,520.8 hundred million in 2031, a CAGR of 3.32%, with the SME segment at a CAGR of 3.96%. This is a forecast for the global market and not a figure for Thailand.

GII market research report

5. Explainer page on the cost of IT department outsourcing

Source for the rule of thumb of 50,000 yen to 300,000 yen for a fixed monthly retainer, 5,000 yen to 10,000 yen for pay-as-you-go pricing, and 100,000 yen to 1,000,000 yen for a dedicated on-site engineer arrangement. These are not figures from a particular survey. The page presents them as a general industry rule of thumb.

Upgrade guide to IT department outsourcing