When a factory in Thailand starts evaluating a case packing automation robot, the comparison sheet almost always opens with the robot’s payload and cycle time. Yet once the line is running, the place where it actually stops is not the robot. It is the corrugated board upstream, and the way the product is presented in the seconds before it is loaded. This article breaks case packing into 3 processes – case erecting, case loading and case sealing – and works out where investment actually stops the stoppages, and where it does not, with every figure checked against Thai labour costs.
A case packing automation robot can simply move the stoppage somewhere else
In factories looking at end-of-line automation, the first document shared with a supplier is almost always the robot datasheet. Payload, reach, repeatability, and a cycle time expressed as so many pieces per minute. Those 4 specifications get laid out against competing models in a comparison table, and the capital request is built on that table.
A few months after commissioning, though, the conversation on the floor is never about robot speed. It is “some days the case will not stand up straight”, “the product does not arrive in a consistent orientation, so an operator ends up re-arranging it anyway”, and “the first hour of the shift is lost to blank feeding faults”. The place that stops is not the place that was on the comparison table.
This is not an unusual failure. It is structural. The comparison table only covers the robot itself, because the processes immediately before and after it are pushed outside the quotation as “existing equipment” or “by customer”. A process left outside the quotation is also left outside the design. A process that was never designed becomes the rate-limiting step after go-live.
It is worth fixing the scope of this article first. Case packing here means putting product into a corrugated case and sealing it. Stacking the finished cases onto a pallet is a separate line downstream, with a different equipment configuration and a different investment range. The cost structure and BOI treatment on the palletizing side are covered in palletizing robot pricing and investment payback, and the two articles are meant to be read together. If you treat both as a single project, the quotation range widens so far that comparison becomes impossible.
Case packing is not one machine – split it into case erecting, case loading and case sealing
The phrase “case packing” actually covers at least 3 independent processes. Before you decide what to put out to quotation, write those 3 down separately on paper.
| Process | What it does | Main equipment | What happens when it stops |
|---|---|---|---|
| Case erecting | Opens a flat corrugated blank, folds the bottom flaps and fixes them with tape or glue so the case stands as a box | Case erector | Everything downstream idles. Operators end up erecting by hand to keep up |
| Case loading | Puts product into the case in the specified count and orientation | Case packer, articulated robot, side-push machine | Cases back up. Speeding up only this step achieves nothing if neither neighbour can follow |
| Case sealing | Closes the top flaps and seals them with tape or glue | Case sealer | Badly sealed cases move downstream and cause problems in palletizing and transport |
Splitting the job into 3 changes how you buy it. Most enquiries arrive as a single request to “automate case packing”, and the integrator converts that into a turnkey line quotation. Real factories, however, are rarely under equal pressure at all 3 processes. Some cannot keep up with case erecting by hand; others are perfectly happy loading manually but suffer inconsistent seal quality.
Once the 3 processes are separated, a much better question becomes available. How many people are tied to each process today, and which of those people are the ones you actually want to redeploy? Until you can answer both with numbers, no brand and no model selection has any basis behind it.
There is one more thing that only becomes visible after the split. The 3 processes run in series, so the real capability of the line is set by the slowest of the 3. Raising the loading step in the middle to 50 cases per minute changes nothing if case erecting runs at 3 cases per minute – the line still runs at 3 cases per minute. That arithmetic is worked through in detail later.
The line stops at the ends, not in the middle
This is the central argument of the article. Of the 3 processes, the one buyers compare most enthusiastically is case loading in the middle. It is where the robot moves, it looks good on video, and the specifications are easy to line up. The processes that actually stop the line are at the two ends.
The upstream fault line is corrugated quality. An automatic case erector repeats the same sequence over and over – pull one flat blank off the magazine with suction cups, fold it to a fixed angle, and bond it at a fixed position. That sequence assumes the blank’s dimensions, score lines, glue quantity and board stiffness are all as designed. The variation a person absorbs with a small adjustment of the hands is not absorbed by a machine.
The other fault line sits immediately before loading and is the presentation of the supply. A robot cannot pick unless it knows where the workpiece is and which way it is facing. A person reaches out and corrects the orientation no matter how the product arrives on the conveyor; a robot can only grip at a defined position and pose. Vision will let it pick from bulk, but that is a separate investment and it changes the cycle time.
On case erector faults, Combi Packaging Systems states on its own blog that the main cause of jams is not the machine but the corrugated. Its specialists explain that automating packaging raises the bar on the quality of the board you buy. Corrugated that used to pass when the work was manual no longer passes on an automatic machine.
Lantech reports that manual case erecting averages 3 cases per minute, while an automatic case erector runs steadily at up to 30 cases per minute. It also notes that hand-erected cases tend to be skewed and out of square, and that this becomes a source of jams downstream in case packers, robotic loaders and case sealers. In other words, the quality of the case upstream shows up as downtime on the equipment downstream.
That causal chain sets the order of investment. Put a fast robot in the middle, feed it skewed cases from one side and disordered product from the other, and the only thing that increases is the number of stoppages and recovery actions. Tidy up both ends first and the middle can often be met with a comparatively modest grade of equipment.
Upstream fault line 1 – corrugated quality decides whether automation is even possible
Corrugated is usually a material that purchasing selects on price. The moment automation is approved, that basis has to change.
5 corrugated factors that stop an automatic case erector
Reorganising the factors Combi lists into a form the buyer can actually verify gives the following.
| Factor | What goes wrong | What the buyer should verify |
|---|---|---|
| Glue squeeze-out | Too much glue during blank manufacture bonds the inner faces, so the suction cups cannot open the blank | The supplier’s control value for glue application and the inspection criterion for squeeze-out |
| Shallow score lines | A shallow score folds incorrectly and the case will not stand square | The specified score depth and whether measured values are recorded per lot |
| Skewed manufacturer’s joint | A joint glued out of alignment produces a case that erects already twisted | The squareness tolerance on the glued joint |
| Recycled content | Recycled fibre is shorter, so the same specification is weaker than virgin board | The recycled ratio of the liner and whether there is an agreement to notify changes |
| Moisture and high humidity | Absorbing moisture in transit or storage causes blank feeding faults | Transport method, warehouse storage conditions, and days from delivery to use |
In Thai factories, humidity and liner substitution are the two that bite
Of the 5 above, the last 2 are the ones that actually bite in Thailand.
Humidity is high all year, rainy season included, and unconditioned raw-material warehouses are common. Corrugated blanks that have absorbed moisture lose their stiffness, and either come off the magazine in twos and threes instead of one at a time, or buckle at the fold line. The variation an operator shrugs off as “the board is soft today” surfaces on an automatic machine as a feeding error.
Liner substitution behaves the same way. Corrugated suppliers change their liner source or recycled ratio while keeping the same part number and the same specification sheet. Manual erecting barely notices. On an automatic case erector the erected angle changes, and the loading position downstream shifts with it.
4 points you can decide tomorrow
There are 4 points on corrugated that the buyer, not the integrator, has to decide.
First, rewrite the incoming specification for corrugated on the assumption of automatic handling. Beyond dimensional tolerance, state the score depth, the squareness of the glued joint, the allowance for glue squeeze-out, the recycled ratio of the liner and the control range for moisture content. A specification sheet with no numbers in it stops functioning as a specification once you automate.
Second, make advance notification mandatory when the liner or the recycled ratio changes. One line added to the contract or the basic trading terms removes one source of unexplained downtime.
Third, define the storage conditions. Set the storage location after delivery, the clearance from the floor, the stack height, the first-in-first-out rule, and an upper limit on how many days after delivery the stock must be consumed. Feeding faults often fall simply by changing where the board sits and how fast it turns over, with no investment in air conditioning at all.
Fourth, use your own corrugated for the equipment acceptance test. If the factory acceptance test at the machine builder is run and passed on clean board supplied by the builder, the line will not come up on your site. Send several lots of your own blanks, ideally including a rainy-season lot, and have the machine run continuously on them. This single point changes the length of the commissioning period.
Upstream fault line 2 – a robot cannot pick what is not presented
The other fault line is the supply of the product that goes into the case.
3 stages of presentation
The investment you need changes completely depending on which of the following 3 stages your supply sits in.
| Stage | Condition of the supply | Equipment required | Difficulty |
|---|---|---|---|
| Stage 1 | Arrives from the upstream process in a single lane at a consistent position and orientation | Positioning stop and simple alignment guides | Low |
| Stage 2 | Orientation is consistent but position varies, or product arrives in multiple lanes | Aligning conveyor, line-up device, vision for position correction | Medium |
| Stage 3 | Supplied in bulk in a tote or container | Bulk feeding unit, flexible feeder, 3D vision | High |
This is the single biggest reason quotations diverge. For the same enquiry to “automate case packing”, a quotation written on the assumption of Stage 1 and one written on the assumption of Stage 3 would be strange to land on the same figure. When the buyer runs a competitive tender without stating the condition of the supply, the cheapest quotation simply looks like the best one. How to read quotations is covered in detail in how to choose a robot system integrator and why quotations diverge.
Where bowl feeders and flexible feeders part company
There are broadly 2 answers to bulk supply at Stage 3.
According to FeedAll, a bowl feeder needs a dedicated bowl, tooling and changeover for each product variant. A vision-based flexible feeder, by contrast, can handle several variants at once, and changeover amounts to selecting a product profile on the touch panel. Parts the robot leaves behind are recirculated and presented again.
On a catalogue this difference looks like a single line about the number of variants supported. On the shop floor it is the daily difference between physically swapping a bowl and tooling and re-adjusting them at every changeover, and an operator selecting a profile on a touch panel. Measure the time the former takes from your own changeover records.
The dividing line is not the number of variants but the number of changeovers per day. With 10 variants but only one changeover a month, a bowl feeder is enough. With 3 variants changed over 4 times a day, bowl feeder setup time eats the benefit of automating. Before you go out to quotation, count the changeovers per day from the last 3 months of production records. That one number very nearly decides the configuration of the supply side.
How to think about the cost of vision, and how far bulk picking is genuinely practical, are set out in robot vision implementation cost and payback.
Gripper selection follows the presentation of the supply, not the shape of the product
Start gripper selection from “what is the product” and you will almost certainly have to select again. The same product needs a different gripper when it is presented differently.
Presentation decides the gripper
At Stage 1, where product arrives in a single lane at a consistent position and orientation, a simple vacuum pad or parallel gripper is enough. Better still, several pieces can be gripped at once, so a full case load can go in with a single motion. Cycle time is set by the number of robot motions, so how many pieces you can grip at once is what sets the speed.
At Stage 3, in bulk, the pose of each piece is different every time and simultaneous gripping is impossible. Pieces are taken one at a time, re-oriented, arranged and then loaded. The same robot and the same product now need a different number of units and a different cycle time.
The order in which to decide
Decide the gripper in the following order.
First, decide whether the presentation will be Stage 1, Stage 2 or Stage 3. That is an investment decision about the feeding equipment, and it is the premise for the gripper.
Second, decide how many pieces are gripped per motion at that presentation. This is what fixes the required cycle time.
Third, decide the gripping method. Vacuum suits rigid items with a flat face and does not suit porous cartonboard or soft bags. A parallel gripper is reliable where the shape is consistent but needs clearance from neighbouring pieces. Pinching or scooping works on irregular items but gives up positional accuracy.
Fourth, decide the variant coverage. Can one gripper handle every variant, will you split grippers by variant family, or will a tool changer swap them automatically? This is where the cost moves. Split grippers by variant family and you add design cost and tooling cost for each one, plus storage space and a changeover step. When you request quotations, make suppliers break gripper cost out into a unit price and a quantity.
A common failure
If gripper prototyping proceeds as “send us a product sample and we will build one”, the assumption about presentation is left entirely to the integrator. Integrators normally assume the most favourable case, Stage 1. The result is a gripper that picked perfectly in the trial and then drops pieces constantly the moment it is connected to the real supply on site. When you send samples, always send a video of how those samples arrive with them.
Bagged and irregular products – where the difficulty starts
When what goes into the case is a bag rather than a rigid item, difficulty rises in steps. It is worth separating out exactly where it starts to rise.
| Condition | Effect on difficulty | Direction of the solution |
|---|---|---|
| Solid contents, tightly filled bag | Close to a rigid body and easy to handle | Vacuum gripping usually works |
| Powder or granular contents that flow | The centre of gravity shifts the instant it is gripped and the pose collapses | Scooping, or pushing the bag in while it stays lying flat |
| Low fill ratio with a large seal margin | No reliable vacuum surface, and the excess film snags | Flatten the shape on a conveyor before feeding |
| Bags must stand upright in the case | They will not stand alone, so they fall over and block the next one | Guides inside the case, or loading several bags at once |
| Slippery film surface | Position shifts during transport and alignment is lost | Review conveying speed, add guides, add pose detection |
The difficulty of bagged product comes down not to the bag itself but to whether the pose can be maintained. Put the other way round, build something on the supply side that maintains the pose and a simple gripper will do. Here too, the investment belongs at the upstream end rather than in the middle.
One more point. In factories handling bags, the change with the best return is usually not the choice of robot but a review of the bag specification. Raising the fill ratio to cut the seal margin, or changing the slip characteristics of the film, can drop the required grade of equipment by a full step. Packaging changes need to be agreed with quality assurance, but they are cheaper than capital equipment and the lead time is shorter.
Break the cost into 5 layers
Before comparing quotations, separate the cost into layers. Without layers, you cannot tell whether a cheap quotation is cheap or simply narrow.
The 5 layers
| Layer | Content | How easily it drops out of a quotation |
|---|---|---|
| Layer 1 | Upstream corrugated measures. Revised incoming specification, supplier selection, storage conditions | Almost always out of scope. Left as the buyer’s job |
| Layer 2 | Feeding and alignment. Aligning conveyors, feeders, bulk feeding units, vision for position correction | Swings to the cheap side whenever the presentation is not stated |
| Layer 3 | Main equipment. Case erector, case packer or robot, case sealer | The one layer always included. Catalogue prices exist |
| Layer 4 | Grippers and tooling. Gripping tools, variant-specific tooling, tool changers | Moves a great deal with the assumed number of variants |
| Layer 5 | Installation, safety and commissioning. Guarding, interlocks, risk assessment, delivery and installation, trial runs, operator training, local support | Frequently contracted with the installation and support terms left vague |
Published prices exist only for Layer 3
This is the important part. Almost all publicly available price information refers to the main equipment in Layer 3.
HIJ Packing Machine puts case packing machinery between USD 12,000 for a semi-automatic case sealer and USD 85,000 for a fully automatic side-push case packer. Converted at roughly THB 32 to the dollar, that is THB 384,000 to THB 2,720,000.
Motion Controls Robotics, a system integrator, puts a robotic case packing line designed for about 20 cases per minute at USD 150,000 to over USD 300,000. At the same rate, that is THB 4,800,000 to THB 9,600,000. The same source puts a manual-changeover case erector at USD 35,000 to over USD 65,000, an automatic random case erector at over USD 150,000, and a complete end-of-line package at USD 750,000 or more in some cases. A downstream palletizing line runs USD 200,000 to over USD 400,000.
The sources differ, so these are not strictly like-for-like configurations, but the order of magnitude carries a meaningful message. Against USD 85,000 for a single fully automatic case packer, the line as a whole costs USD 150,000 to over USD 300,000. That works out to everything other than that single case packer accounting for roughly 43% to 72% of the investment. Inside that 43% to 72% sit the case erector and the case sealer, which belong to the same Layer 3, together with Layer 2 feeding and alignment, Layer 4 grippers and tooling, and Layer 5 installation, safety and commissioning. A manual-changeover case erector on its own costs USD 35,000 to over USD 65,000, so a substantial share of that gap is taken up by the other two Layer 3 machines. Layer 1, the corrugated measures, is not in the integrator’s scope at all, so it is not even inside that gap. It has to be budgeted separately as a buyer cost.
Put differently, a capital request built on the catalogue price of a single case packer captures only about 30% to 60% of the real investment. And when an integrator’s turnkey quotation looks expensive, most of the difference is not margin – it is the two ends and the installation.
Variant flexibility is what pushes the price of the main equipment up
The price spread on case erectors shows this structure most clearly. A manual-changeover case erector costs USD 35,000 to over USD 65,000, while an automatic random case erector costs over USD 150,000. As a multiple, that is roughly 2.3 to 4.3 times.
What the difference buys is not speed but the ability to change case size without anyone touching the machine. Choose a random erector “to be ready for more variants in future” without having counted how many times a day the case size actually changes, and that difference sits on the balance sheet without ever being recovered.
Speed is set by the cases per minute of the slowest process, not the catalogue
The speed printed in a catalogue is what that machine can achieve alone under ideal conditions. A line does not behave that way.
Speed range by process
The 2 case erecting rows come from Lantech; the other 3 rows are figures published by HIJ Packing Machine.
| Process and type | Speed | Source |
|---|---|---|
| Manual case erecting | 3 cases per minute on average | Lantech |
| Automatic case erector | Steady at up to 30 cases per minute | Lantech |
| Semi-automatic case packing machine | 15 to 25 cases per minute | HIJ Packing Machine |
| Combined erect, load and seal machine | 30 to 50 cases per minute | HIJ Packing Machine |
| Servo-driven side-push type | 40 to 60 cases per minute | HIJ Packing Machine |
An automatic case erector is roughly 10 times manual speed. What that factor of 10 does to the line as a whole is worth working out.
The rate-limiting arithmetic
Suppose you install a servo-driven side-push packer at 50 cases per minute and leave case erecting manual at 3 cases per minute for now. The line then outputs 3 cases per minute. Utilisation of the packer is 3 divided by 50, which is 6%. You are running a 50 cases per minute machine at 6%.
Now put the same money into the automatic case erector first. Case erecting rises to 30 cases per minute and the constraint moves to the next slowest process. If manual loading cannot keep up, that becomes the new constraint. Constraints do not disappear, they move. Predicting where they will move to is what staged investment planning actually is.
How to estimate effective speed
Do not convert a catalogue figure straight into daily output. Work out the cases that will actually flow in a day as follows.
Start from the operating hours in a day and subtract the downtime consumed by product changeovers, to get net operating minutes. Then multiply those net minutes by the catalogue speed of the rate-limiting process, by availability, and by yield. That is the number of cases you can actually ship in a day. Availability here must cover only the stoppages other than changeover, because changeover has already been subtracted. Do not deduct it twice. Availability, yield and changeover time are all worthless if you borrow another company’s averages. Measure them on your own current line for 1 week. Compare equipment on cases per minute alone, without those 3 measured values, and the discussion after installation will be about why the line does not deliver what the catalogue promised.
Building the payback case for a Thai factory
From here the payback is worked through with Thai labour costs. Everything below is this article’s own estimate, and different assumptions produce a different conclusion. Every assumption is stated.
Assumptions
Assume a factory in greater Bangkok. According to JETRO, the minimum wage in Bangkok is THB 400 per day, effective 1 July 2025. The previous rate was THB 372, and the revision is reported to cover roughly 700,000 workers.
The real cost of one operator is not the minimum wage itself. It includes social security, overtime premiums, bonuses, transport allowance and canteen subsidy. This article uses an on-cost of 40%, which puts the real cost at THB 560 per day.
Operation is 26 days a month, 2 shifts of 8 hours. Annual cost per person is THB 560 multiplied by 26 days and by 12 months, which is THB 174,720.
Assume that case erecting, loading and sealing at the end of the line occupy 3 people per shift across 2 shifts, so 6 people in total. Within each shift, 1 person is on case erecting and 2 on loading, with sealing covered by the loading operators. Annual labour cost for those 6 people is THB 1,048,320.
Effective line speed today is 3 cases per minute, set by manual case erecting as the constraint. Monthly output from that is roughly 75,000 cases. Multiplying 3 cases per minute by 60 minutes, 8 hours, 2 shifts and 26 days gives 74,880 cases. Assume this factory’s current capacity and its order intake are broadly in balance. On that basis, the estimate puts no value for extra output in the numerator – whether more output can be sold is determined by demand, not by equipment.
The exchange rate used is roughly THB 32 to the dollar.
The people you can remove, and the people you cannot
Do not assume all 6 people disappear. Work remains after automation.
| Work that remains | Reason |
|---|---|
| Replenishing corrugated blanks | Filling the magazine is usually left outside the automation scope |
| Replenishing and changing tape and glue | Someone has to watch consumable levels |
| Pallet changes and moving finished cases away | In a configuration that stops short of palletizing, a person stays |
| Fault recovery | Jams, feeding faults and dropped product |
| Quality checks | Seal condition, count of contents, label verification |
This estimate assumes 4 of the 6 are removed and 2 remain, 1 per shift. Annual saving is THB 174,720 multiplied by 4 people, which is THB 698,880.
Investment and payback period
The investment figures come from the integrator estimate quoted earlier. Against USD 150,000 to over USD 300,000 for a robotic case packing line, the best case is taken as USD 150,000 and the worst case as USD 300,000. In baht that is THB 4,800,000 and THB 9,600,000.
Annual maintenance, consumables and power are assumed at 5% of the investment per year. That is THB 240,000 in the best case and THB 480,000 in the worst case.
| Item | Best | Worst |
|---|---|---|
| Investment | THB 4,800,000 | THB 9,600,000 |
| Annual labour saving | THB 698,880 | THB 698,880 |
| Annual maintenance and running cost | THB 240,000 | THB 480,000 |
| Annual net benefit | THB 458,880 | THB 218,880 |
| Payback period | 10.5 years | 43.9 years |
The arithmetic is as follows. Best case, 698,880 minus 240,000 gives THB 458,880, and 4,800,000 divided by 458,880 gives 10.5 years. Worst case, 698,880 minus 480,000 gives THB 218,880, and 9,600,000 divided by 218,880 gives 43.9 years.
Note that this investment figure is an integrator estimate for about 20 cases per minute. The factory in this estimate currently runs 3 cases per minute at 74,880 cases a month, so a 20 cases per minute line is heavily over-specified. Buying surplus capacity is itself one reason the payback period is so long. In practice you would fix the required cases per minute first and re-quote a configuration matched to that speed.
With labour saving alone in the numerator, the investment does not pay back in Thailand. That is the conclusion of this estimate. At THB 400 per day, 4 people at the end of the line come to only THB 698,880 a year, and the numerator is simply too small against a multi-million baht investment.
How to read the vendor claims of 85% reduction and 1 to 2 year payback
HIJ Packing Machine states that full automation can cut end-of-line labour by up to 85% and pay back in 1 to 2 years for medium-volume production. That is the vendor’s statement, not an assertion by this article. What matters is working backwards to the conditions under which those numbers hold.
Take the 85% first. Applying 85% to the 6 people in this estimate gives 5.1 people. This estimate removes 4, which is about 67%. The 1.1 person gap is the people kept for blank replenishment, pallet changes and fault recovery. Achieving 85% means bringing automatic blank feeding, a direct connection to palletizing, and automatic fault recovery into scope. That is a much broader investment, not a conversation about a single USD 85,000 machine.
Now the 1 to 2 year payback. Recovering USD 85,000, that is THB 2,720,000, in 1 year requires an annual saving of THB 2,720,000. Divided by THB 174,720 per person from this estimate, that is about 15.6 people. For a 2 year payback the annual figure is THB 1,360,000, or about 7.8 people. This is a rough back-calculation that ignores maintenance cost; deduct maintenance and the required headcount rises further.
In other words, the vendor’s claim holds for a factory with 8 to 16 people permanently stationed at the end of the line, or a country where labour costs several times Thai levels. It is not a figure to apply to a 3 person per shift line in Thailand. Doing this back-calculation before the capital request keeps a supplier’s payback figure from being carried into your own board paper unchanged.
For reference, if you assumed that removing 4 people held true against the machine-only price of THB 2,720,000, the net benefit after 5% maintenance of THB 136,000 would be THB 562,880 and payback would be 4.8 years. But that assumption does not hold. The machine-only price excludes Layers 1, 2, 4 and 5, and it leaves feeding and case erecting manual. Combining a machine price with a line-level benefit is precisely the calculation that overlooks both ends.
Apply a sensitivity factor only to effects that depend on availability
When testing the sensitivity of the payback, it is common to model a case with lower availability. Here, do not apply the factor uniformly to every effect.
The labour saving of THB 698,880 does not depend on availability. Operators are assigned to a shift, and wages accrue while the line is stopped as well. Redeploy those 4 people and the labour that disappears is the same whether availability is 0.9 or 0.7.
Extra output, overtime reduction and scrap reduction, on the other hand, are proportional to availability. If availability falls from 0.9 to 0.7, they become 0.7 divided by 0.9, or about 0.78 times.
| Effect | Depends on availability | Treatment at availability 0.9 to 0.7 |
|---|---|---|
| Labour saving of THB 698,880 | No | Stays at THB 698,880 |
| Extra output, overtime reduction, scrap reduction | Yes | About 0.78 times |
Apply 0.78 to the labour saving by mistake and THB 698,880 becomes THB 545,126. Net benefit after THB 240,000 of maintenance is THB 305,126, and payback becomes 4,800,000 divided by 305,126, which is 15.7 years. The correct answer is 10.5 years, so the error overstates the payback period by 5.2 years. An attempt to be conservative ends up pushing the investment decision the wrong way.
Add to the numerator, or reduce the denominator
Since labour alone does not pay the investment back, the only options are to add to the numerator or reduce the denominator. There are 4 realistic choices.
First, increase the headcount in scope. Move to 3 shift operation, share one robot across multiple lines, or bundle case packing and palletizing into a single investment decision while keeping them separate projects. The configuration when palletizing is included is covered in palletizing robot pricing and investment payback.
Second, narrow the scope. Instead of a full line, automate only case erecting, the rate-limiting step. A manual-changeover case erector is USD 35,000 to over USD 65,000, that is THB 1,120,000 to THB 2,080,000. Against the best case full line at THB 4,800,000, the denominator falls to somewhere between one quarter and under half. But the numerator shrinks at the same time. In this estimate only 1 person per shift, 2 across the 2 shifts, is tied to case erecting, and because blank replenishment and jam recovery remain, the saving does not even reach 2 people. A smaller denominator does not automatically bring payback closer, so recalculate with both the narrowed denominator and the narrowed numerator before deciding.
Third, add downtime and quality to the numerator using your own measurements. If skew in hand-erected cases is causing jams downstream, that downtime has a monetary value. There is no published primary source for the size of the reduction, so this article’s estimate excludes the effect. If you include it, use your own measured values. Borrow someone else’s averages and only the numerator inflates, conveniently.
Fourth, reduce the denominator with investment incentives. That is covered in the next section.
One more point. Design beyond the end of the line, taking in how finished cases move around the factory, and the scope of the investment changes again. The connection to in-plant material movement is set out in in-plant logistics improvement runs in 3 layers.
BOI and the denominator of the investment
Thailand offers investment incentives for automation and robot adoption. Where they apply, the denominator discussed in the previous section falls and the payback arithmetic changes.
This article does not go into announcement numbers or detailed eligibility. The scheme is revised over time, and eligibility depends on the nature of the business and the timing of the application. Primary sources can be checked from the list of BOI-related laws and announcements for Thailand maintained by JETRO. Eligibility and the practicalities of the application are treated in more depth in palletizing robot pricing and investment payback.
Only the practical point needs stating here. Confirm eligibility before you freeze the equipment specification. Eligibility conditions can put constraints on the scope of qualifying equipment or on the timing of installation. Confirm after the specification and the contract are locked and you can no longer make the changes those conditions require.
Product changeover – automation creates setup work that did not exist before
The assumption that automation reduces setup work is only half right. Some setup work does disappear, but there is setup work that only comes into existence once you automate.
| Setup work created by automation | Content |
|---|---|
| Case size changeover | On a manual-changeover machine an operator adjusts guide widths and fold positions. Automatic on a random erector |
| Gripper change | Where grippers are split by variant family, a physical swap and a datum check are required |
| Robot program change | Loading pattern, number of layers, placement positions |
| Supply-side variant change | Physical swap of dedicated bowl and tooling on a bowl feeder, or selecting a product profile on the touch panel of a flexible feeder |
| Supply guide width adjustment | Guides and stop positions on the aligning conveyor |
| First-piece check | An operator opens the first case after changeover and verifies count and orientation |
On a manual line most of this was absorbed by an operator simply holding the product differently. Once automated, it is booked as machine setup time. A 15 minute changeover 4 times a day is 1 hour, and 26 hours over 26 days in a month. Those 26 hours are downtime that did not exist before automation.
So in a factory with frequent changeovers, think in this order. First, count the changeovers per day from the last 3 months of production records. Next, identify which combinations of variants account for the most changeovers. Then examine whether the high-frequency combinations can be consolidated onto one gripper and one case size. Solving it through production planning and packaging consolidation is often cheaper than solving it with equipment. Equipment selection comes after that.
9 items to settle on paper before you go to quotation
Write out the following 9 items yourself before requesting quotations. Run a competitive tender with these blank and each supplier will quote against a different set of assumptions, so the prices cannot be compared at all.
| # | What to settle | How to settle it |
|---|---|---|
| 1 | Which of case erecting, loading and sealing is in scope | Write the 3 processes separately and mark each as this project, next project, or not at all |
| 2 | Whether presentation is Stage 1, Stage 2 or Stage 3 | Film the actual supply and attach the video to the request for quotation |
| 3 | Target variants and their mix | State what share of volume the top variants represent. Full variant coverage is considered last |
| 4 | Changeovers per day | Count them from the last 3 months of records. The feeder and gripper configuration is decided here |
| 5 | Rate-limiting process and target cases per minute | Measure current effective speed and write in one line which process goes to what speed |
| 6 | Incoming specification for corrugated | State score depth, squareness, glue squeeze-out allowance, recycled ratio and moisture control range as numbers |
| 7 | Headcount removed and headcount retained | Name the work the retained people will do. Leave this vague and the benefit gets inflated |
| 8 | Acceptance criteria | State how many hours of continuous running on your own corrugated and your own product constitutes a pass |
| 9 | Local support terms | Attendance, training, spare parts stock, response time. Confirm whether there is a support base inside Thailand |
Item 8 is the one most often skipped. Sign a contract without acceptance criteria and every problem during commissioning gets classified as a local operating issue. Put the conditions for continuous running on your own corrugated, your own product and your own changeover frequency into the contract.
Frequently asked questions
How much does a case packing automation robot cost?
Published price information puts the machine alone at USD 12,000 to USD 85,000, which at roughly THB 32 to the dollar is THB 384,000 to THB 2,720,000. Note that the lower bound of USD 12,000 is a semi-automatic case sealer while the upper bound of USD 85,000 is a fully automatic side-push case packer, so the same range contains two quite different things. That is main equipment only. An integrator estimate for a robotic case packing line at about 20 cases per minute is USD 150,000 to over USD 300,000, or THB 4,800,000 to THB 9,600,000. Everything other than that single case packer, the case erector and the case sealer included, accounts for roughly 43% to 72%. A complete end-of-line package can reach USD 750,000 or more.
Can we install just the case erector first?
You can, and in many factories that is the sensible move. Manual case erecting averages 3 cases per minute while an automatic case erector runs steadily at up to 30 cases per minute. Where case erecting is the constraint, fixing that one process changes the output of the whole line. A manual-changeover case erector at USD 35,000 to over USD 65,000 gives a far smaller denominator than a full line. Remember, though, that you can only remove the people tied to case erecting, so the numerator shrinks as well. And the constraint does not disappear, it moves to the next process, so work out in advance where it will move to.
Is automating bagged product harder than boxed product?
It depends on the conditions. A tightly filled bag with solid contents behaves almost like a rigid body, but powder or granular contents flow, so the centre of gravity shifts the instant the bag is gripped and the pose collapses. A bag with a low fill ratio and a large seal margin is also harder, because there is no reliable vacuum surface. The difficulty comes down to whether the pose can be maintained rather than to the bag itself, so if you build something on the supply side that maintains the pose, a simple gripper is often enough. Reviewing the fill ratio or the film specification can be a cheaper solution than capital equipment.
Can a collaborative robot do case packing?
Sometimes yes and sometimes no. The answer comes from the required cycle time and the presentation of the supply, not from the type of robot. A collaborative robot may allow you to omit guarding, but coexisting with people places limits on its operating speed. The higher your target cases per minute, therefore, the worse the fit, and at some point the speed limit itself becomes the constraint. Settle item 5 above, the rate-limiting process and the target cases per minute, as numbers first, then require candidate suppliers to state the effective speed after risk assessment. Compare speeds with the safety functions active rather than catalogue maximums, and whether a collaborative robot is sufficient answers itself.
Do we have to change our corrugated?
You do not need to change the part number, but you do need to add control items to the incoming specification. The main cause of jams on an automatic case erector is the corrugated rather than the machine, with glue squeeze-out, shallow score lines, a skewed manufacturer’s joint, strength loss from recycled fibre, and feeding faults from moisture or high humidity all cited. Automating packaging raises the bar on corrugated quality by a step, as Combi puts it. Start by writing score depth, joint squareness, glue squeeze-out allowance, recycled ratio and moisture control range into the specification as numbers, and by agreeing advance notification when the liner changes.
Should we automate through to palletizing at the same time?
It helps in the sense that it enlarges the numerator, but design it as a separate project. Case packing and palletizing use different equipment configurations, and even in integrator estimates a palletizing line is quoted separately at USD 200,000 to over USD 400,000. Combined into a complete end-of-line package it can reach USD 750,000 or more. The realistic sequence is to fix the constraint and the presentation on the case packing side first, and connect the downstream process after that.
Summary
Automating case packing is not about buying one machine. It is about designing a chain of 3 processes – case erecting, case loading and case sealing. Buyers compare the process in the middle, where the robot moves, but the line stops at the two ends.
The upstream fault line is corrugated quality. Glue squeeze-out, shallow score lines, a skewed manufacturer’s joint, strength loss from recycled fibre, moisture and high humidity all produce jams on an automatic case erector. Skew in hand-erected cases shows up as downtime on the case packer and the case sealer downstream. Manual case erecting averages 3 cases per minute; an automatic case erector reaches 30 cases per minute.
The other fault line is the presentation of the supply. Aligned single-lane supply, varying position, or bulk. Which of those 3 stages you are in decides the equipment, the gripper and the investment. The dividing line between a bowl feeder and a flexible feeder is not the number of variants but the number of changeovers per day.
Cost splits into 5 layers. Published prices exist only for Layer 3, the main equipment. Against USD 85,000 for a fully automatic case packer on its own, the line is USD 150,000 to over USD 300,000, so everything other than that single case packer, the case erector and the case sealer included, is roughly 43% to 72%. On speed, a 50 cases per minute packer sitting behind manual case erecting at 3 cases per minute gives a line running at 3 cases per minute and a packer utilisation of 6%.
Payback in Thailand came out harsh in this article’s estimate. On a Bangkok minimum wage of THB 400 per day plus 40% for statutory benefits and allowances, giving THB 560 per day, at 26 days a month with 3 people per shift across 2 shifts for 6 in total, removing 4 people saves THB 698,880 a year. Against an investment of THB 4,800,000 to THB 9,600,000 with maintenance at 5% a year, payback is 10.5 to 43.9 years. With labour alone in the numerator, it does not pay back.
The vendor claim of 85% reduction and 1 to 2 year payback, worked backwards, assumes 8 to 16 people permanently stationed at the end of the line, or a country where labour costs several times Thai levels. When you run a sensitivity case, do not apply the factor to the labour saving, which does not depend on availability. Apply 0.78 uniformly by mistake and payback stretches from 10.5 years to 15.7 years, overstating it by 5.2 years.
So what to decide tomorrow is not a model of equipment. Measure which of the 3 processes is the constraint. Fix whether presentation is Stage 1, Stage 2 or Stage 3 using video. Put numbers into the incoming specification for corrugated. And count the changeovers per day. Request quotations once those 4 are filled in, and every supplier’s price will finally sit on the same footing.
Even at the stage where you have not yet decided which process to automate first, it is perfectly reasonable to engage us just to measure the constraint and separate out the presentation of the supply. TOMAS TECH supports Japanese manufacturers in Thailand end to end, from shop-floor measurement and process design through technical advice on integrator selection to commissioning. There are cases where downtime falls simply by revising the incoming specification for corrugated, so please get in touch through our contact page even for a discussion that does not assume capital investment.
References
- Most Troubles with Automated Box Forming Aren’t the Machine, It’s the Corrugated — Combi Packaging Systems The main cause of case erector jams lies with the corrugated rather than the machine, covering glue squeeze-out, shallow score lines, skewed joints, strength loss from recycled fibre, moisture and high humidity
- Packaging Line Downtime and the Case Erector — Lantech Manual case erecting averages 3 cases per minute, an automatic case erector reaches 30 cases per minute, and skew in hand-erected cases causes jams downstream
- Manual vs. Automated Case Erecting — Lantech When to use manual case erecting and when to automate it
- Case Packing Machines — HIJ Packing Machine The USD 12,000 to USD 85,000 price band, speed bands of 15 to 25, 30 to 50 and 40 to 60 cases per minute, and the vendor statement of 85% labour reduction with 1 to 2 year payback
- How Much Is a Pick, Pack, Pal Robot Cell Going to Cost — Motion Controls Robotics A robotic case packing line at about 20 cases per minute for USD 150,000 to over USD 300,000, a palletizing line at USD 200,000 to over USD 400,000, a manual-changeover case erector at USD 35,000 to over USD 65,000, an automatic random case erector at over USD 150,000, and a complete end-of-line package at USD 750,000 or more
- Flexible Feeding Systems — FeedAll Variant-specific tooling and changeover on bowl feeders, versus product profile switching and recirculation on vision-based flexible feeders
- Business brief on Thailand’s minimum wage increase — JETRO The Bangkok minimum wage of THB 400 per day, the previous rate of THB 372, effective 1 July 2025, covering roughly 700,000 workers
- List of BOI-related laws and announcements for Thailand — JETRO Where to find primary sources on Thailand’s investment promotion scheme
- Thailand food market environment survey, March 2026 — JETRO Background material on Thai market conditions for companies considering case packing automation in the food sector
