A factory knows exactly what came in the goods-in door and exactly what left the dispatch door. Everything between those two doors is a story people tell each other. This is what the Orbit Suite does to the middle.
A composite discrete manufacturer — an appliance and electronics assembler of the kind NBS already works with. It is a contract assembler, which makes the traceability and labelling pressure real rather than theoretical. The last section re-points it at food, plastics and automotive.
| Part of the plant | What it does | Scale |
|---|---|---|
| Goods-in & components store | Supplier receiving, inspection, quarantine, kitting to lineRoughly 2 400 component SKUs and R14m of stock on hand | 2 400component SKUs |
| Assembly | Two lines, 14 stations, running two shiftsEach unit is serialised and labelled on the line | 1 800units per day |
| Quality | In-line test, end-of-line test, non-conformance and rework loopISO 9001 certified — audited annually | 6QC staff |
| Maintenance | Line equipment, compressors, conveyors, test rigs, tooling | 12maintenance staff |
| Finished goods & dispatch | Pack, palletise, logistics labelling, outbound to three retail DCs and one brand owner | 3 + 1delivery destinations |
| People | 160 production across two shifts, 18 stores and warehouse, 12 maintenance, 14 office, 6 QC | 210headcount |
None of these are failures of effort. They are what happens when the only two points a factory measures precisely are its two doors, and everything between them is reconstructed afterwards from memory and paperwork.
Goods-in is exact and finished goods is exact. In between, where a given order actually is is a question answered by walking the floor and asking.
“The line was down about an hour.” It was four hours twenty across three stoppages, and the same bearing caused two of them — but nothing recorded that, so nothing fixed it.
Templates live on individual machines. Versions drift, an old one gets printed, and a retail DC rejects the whole pallet over a barcode nobody can scan.
A supplier recalls a component lot. You can say which month it went into production. You cannot say which units — so you quarantine far more than you need to.
A wall-to-wall count shuts the store for three days, corrects the book, and is wrong again by February because nothing between counts is measured.
Test records, material certificates and work instructions are filed properly and found slowly. The ISO audit is a fire drill, not a report.
A factory that cannot see its middle cannot improve it. Every continuous-improvement programme, every OEE target and every corrective action depends on a measurement that currently does not exist — so the improvement effort goes into the two places already measured, which are the two places already fine.
This blueprint is not about replacing the systems at the doors. It is about instrumenting the middle so the rest of the management system has something true to work from.
Each stage adds one machine-read event to a place that is currently a handwritten note or nothing at all. Below, what happens on the ground and what carries it.
A supplier truck arrives unannounced at the gate. A storeman counts cartons against a delivery note, signs it, and the GRN is keyed later that afternoon — sometimes the next morning.
Material certificates and certificates of conformity come as paper in the box. Anything on inspection hold is marked with a sticker and a hope.
Ephoro books the driver and vehicle in at the gate by scan, so an unannounced delivery is at least a recorded one. TAG reads tagged pallets through the dock portal, reconciling the delivery against the order before the truck leaves.
Ledger raises the GRN at the dock rather than at a desk, and Vault takes the material certificate against that receipt — so when a customer asks for traceability on a batch, the certificate is attached to it rather than filed near it.
Kits are picked against a printed list. A wrong or short kit is discovered at the station, and somebody walks back to the store while the line waits.
Stock accuracy drifts between quarterly counts, so the system says a part is there when it is not — and the planner finds out at the worst possible moment.
Ledger's picking handheld confirms every line by scan against the bin, so a short or wrong kit is caught in the store rather than at the station.
TAG turns the quarterly wall-to-wall count into a walk with a reader — which means it can be done monthly, or weekly on the high-value bins, and stock accuracy stops being a seasonal event. Flux tracks the reusable totes and carriers themselves, which is usually the first thing a plant loses.
Units move down the line through fourteen stations. Progress is tracked on a whiteboard at the end of the shift. The serial number is printed and applied at the end, from a template on the label PC.
If the template is the wrong version, nobody finds out until the pallet reaches the DC.
Read points at line entry, mid-line and line exit turn WIP into a count that updates itself. Pulse clocks both shifts, so labour cost per unit is a daily number and an absence on the early shift is known before the line is short-handed.
The label stops being a local file. Loftware holds one approved template per product with business rules, versioning and an audit trail, and the print-and-apply head applies it on the line — so the version that prints is the version that was approved, every time. Cipher keeps the station terminals and wrist scanners on the build they are supposed to be on, which is the difference between an app update taking an hour and taking a shift.
End-of-line test results go on a sheet. A failure gets a red sticker and goes to the rework bench, where it joins a pile with no queue and no owner.
Non-conformance reports are written after the fact. Whether the corrective action actually happened is a matter of trust.
A non-conformance becomes an Orion ticket against the unit and the station: owned by a named person, carrying an SLA, costed for the labour and parts the rework consumed, and closed only when somebody closes it.
TAG keeps the quarantine and rework areas as real locations rather than corners of the floor, so held stock is held stock. Vault holds the test records and work instructions the auditor asks for — by revision, not by lever-arch file.
A breakdown is a phone call. A fitter attends, fixes it, and tells somebody. Planned maintenance happens when there is time, which is to say sometimes.
Gauges and torque tools are calibrated on a spreadsheet with reminders that nobody owns. Tooling walks.
Orion is the heaviest-used application in a plant. Every conveyor, compressor, test rig and press is an asset with its own history; breakdowns and planned work are both jobs; labour and spares are costed against the machine that consumed them. Downtime stops being remembered and starts being a number with a cause attached — which is the whole basis of an OEE programme.
Flux holds custody of gauges, jigs, fixtures and torque tools with calibration dates, which is exactly what the ISO auditor comes to look at. Ledger carries the spares as stock, so a critical spare at zero is visible before the machine needs it.
Units are packed, palletised and labelled for the receiving DC's own standard. A wrong logistics label is the single most common reason a load comes back.
The truck leaves with a delivery note. Proof of delivery returns days later, creased, or not at all.
The pallet label is generated from the same approved Loftware template set as the product label, against the destination's own standard. Dock-door readers confirm what physically left against what was meant to leave — the last chance to catch a mis-ship, taken automatically.
Ledger invoices off the dispatch rather than off a note, and Relay carries the run on the driver's phone with proof of delivery signed at the DC and a tracking link the customer can follow.
A unit comes back under warranty with a serial number and no history. A supplier recalls a component lot and you quarantine a whole month of production because you cannot narrow it further.
The brand owner you contract-manufacture for asks for quality data and somebody spends a day building a spreadsheet.
This is the payoff for everything above. Because the serial was bound to the kit, the station, the shift and the test result, “which units contain that component lot” is a query — so the recall is scoped to the units actually affected rather than to the calendar.
Fuse gives the brand owner their own front door for order status, quality data and corrective actions, instead of a monthly spreadsheet. Orion carries the warranty job with the unit's own build history attached, and Vault holds the evidence behind it.
Thirteen applications behind one Orbit sign-in. Ten carry real weight in a plant; three are supporting, and it is worth being straight about which.
Carries a job from logged to closed: scheduling, labour and parts costed against that asset's own history, an SLA the team is held to, and recurring tasks that raise themselves.
Here: The most heavily used application in a plant. Every machine as an asset with its history; breakdowns and planned maintenance as jobs; non-conformances and rework as owned, costed, closable tickets.
UHF at 860–960 MHz. Fixed readers to 33 dBm read 500-plus tags a second at ranges to 20 m; the handheld is a rugged Android terminal with the same range plus 1D/2D barcode.
Here: Dock portals at goods-in and dispatch, read points at line entry, mid-line and exit, quarantine as a real location, and a stocktake that becomes a walk instead of a shutdown.
One set of double-entry books carrying quotes, orders, invoices, debtors, creditors, banking and stock at weighted-average cost, with internal barcodes and a handheld picking app.
Here: GRN at the dock, kit picking confirmed by scan, spares as stock, dispatch-driven invoicing. Read the limits section — there is no bill of materials and no works order.
Equipment booked out against a named person and a place, verified by barcode, with the audit trail kept for the day somebody asks.
Here: Gauges, jigs, fixtures, torque tools and reusable totes — who holds it and when it was last calibrated. In an ISO 9001 plant this is the audit, not housekeeping.
Staff clock in from the app or a fixed point. The manager's roll call shows who is on site, who is late and what the week is costing.
Here: Two shifts and 160 production staff. Labour cost per unit becomes a daily figure, and a short early shift is known at 06:10 rather than when the line stalls.
A personal drive for everyone and shared drives per team behind the same Orbit sign-in, with links you stay in control of and a restorable trash.
Here: Work instructions by revision, material certificates against the receipt, test records against the unit. The ISO audit becomes a retrieval exercise rather than a search.
A control tower plans and assigns runs on a live map; the driver's phone carries the run sheet, stop check-in and proof of delivery; customers get a public tracking page.
Here: Outbound to three retail DCs and the brand owner, with proof of delivery signed at the gate instead of a creased note returning on Thursday.
Customers log requests, follow progress and see their own sites, assets and contracts. What they log lands in Orion and Ledger rather than in somebody's inbox.
Here: The brand owner's front door: order status, quality data and corrective actions on demand, rather than a spreadsheet somebody builds each month. Often contractual.
Reads the South African driver's licence barcode and the vehicle licence disc at the gate, so a person and a vehicle are signed in by scan in seconds.
Here: Supplier trucks at the receiving dock, contractors on site for a shutdown, and a live answer to “who is in the building” that a fire drill can actually use.
Contacts organised by company, industry and province. Mail built from blocks and rendered server-side, sent cleanly, reported back on.
Here: A supporting role rather than a load-bearing one: supplier notices, product-change notifications, and the recruitment list a two-shift plant runs constantly.
Registration, invitations, a check-in desk a scanner links to by QR, live arrival statistics and walk-in registration at the door.
Here: A narrower fit than in most sectors, and worth saying so. It earns its place on contractor induction and safety days, supplier conferences and factory tours.
Tags every vehicle on a yard so the whole yard is walked in one pass, and flags units standing too long.
Here: The weakest fit of the thirteen on a single-site plant. It becomes relevant where the yard holds trailers. Otherwise leave it out and say so.
Built on Google's Android Management API. A device enrols by scanning a durable QR code and lands on the policy chosen for it — the apps it may run, the settings it may change, and kiosk mode if it should only ever do one job. From the console you can push an app, lock a handset somebody left behind, watch data usage, draw a geofence, and wipe a leaver's device.
Here: The ten rugged handhelds across two shifts and the wrist-worn scanners. Kiosk mode locks a station terminal to the one job it does, and an app update reaches every device on the floor without anybody collecting them in a box first.
Orbit itself is the control plane underneath all thirteen: one sign-in, per-application entitlement, and per-user permissions down to section and action, so a storeman, a line supervisor, a quality manager and the financial director see four different systems.
This is where a manufacturing deployment differs most from every other sector: the hardware is not an accessory to the software, it is half the solution — and it is NBS's own core business rather than something bought in.
| Where | What | Why this one |
|---|---|---|
| Label managementAcross every printer in the plant | Loftware Cloud or NiceLabel | One approved template per product with business rules, versioning and an audit trail. The direct answer to dark spot 03 — the label stops being a file on somebody's PC. Licensed separately from Orbit. |
| On the line | NBS Real-Time Print & Apply, NBS Automatic Flat-Surface Labelling Head, TSC PEX Series print engine (4″/6″) | Prints and applies the approved label at line speed without an operator deciding which template to open. |
| Stores & dispatch | TSC MH Series (4″/6″), TSC MB Series (4″), TSC T800 / Alpha Series RFID | Duty-cycle printers for bin, kit and pallet labels. The RFID models print and encode in one pass, which is what makes a tagged pallet cheap enough to be routine. |
| Date & lot coding | Printronix C160 · C170 Food Grade | Direct marking where a label will not survive. The food-grade unit matters the moment this blueprint is re-pointed at food and beverage. |
| Fixed read points18 points across the plant | Impinj R700, 12 dBi UHF panel antennas, Impinj xSpan, UHF RFID Security & Access Gate | Goods-in dock ×3, store in/out ×2, line entry, mid-line and exit ×4, quarantine ×1, finished goods ×2, dispatch dock doors ×4, asset egress ×2. |
| In people's hands10 terminals | Newland MT93 Megattera Pro UHF, Newland MT67 Sei II / N7 Cachalot Pro II, Newland WD1 II Watch Scanner | The MT93 for RFID work, the mobile computers for store and picking, and the wrist-worn scanners for station work where both hands are already busy — the detail that decides whether line operators actually use it. |
| Production paperworkWhere it still exists | Printronix P8000 line matrix | Only where a dirty, high-duty environment still needs printed paperwork. Worth quoting honestly: the aim of this blueprint is to need less of it. |
All supplied, installed and supported by NBS. Every RFID line above is subject to a site survey before it is quoted — on a plant handling metal chassis or liquid-filled product this is not a formality.
Indicative South African market ranges for the categories of software a plant this size is typically already paying for — not quotations, and not attributed to any named vendor.
| What they pay for today | Orbit application | Indicative today, per month |
|---|---|---|
| Mid-market ERP or accounting, about 35 users | LedgerPartly — see the limits section | R14 000 – 32 000 |
| Standalone WMS or stock system | Ledger + TAG | R4 000 – 12 000 |
| Desktop label design licences, per seat | Loftware | R2 500 – 8 000 |
| CMMS or maintenance software | Orion | R3 500 – 9 000 |
| Quality and document control for ISO 9001 | Vault + Orion | R3 000 – 8 000 |
| Clocking terminals and T&A software, 210 staff | Pulse | R4 500 – 11 000 |
| Calibration and gauge management | Flux | R1 500 – 4 000 |
| Cloud file storage, 45 seats | Vault | R4 500 – 9 000 |
| Access control and visitor management | Ephoro | R2 000 – 4 500 |
| Delivery app or third-party proof of delivery | Relay | R2 000 – 5 000 |
| Email campaign tool | Beacon | R900 – 2 500 |
| Customer and brand-owner portalAlmost nobody has one. The cost is the monthly quality pack somebody builds by hand. | Fuse | — |
| Mobile device management, ~14 handhelds, wrist scanners and tablets | Cipher | R600 – 2 500 |
| Software the plant already pays for | R43 000 – 107 500 |
A plant already spends roughly R75 000 a month at the midpoint on twelve systems, none of which join up. The cost of them not joining up is an order of magnitude larger than the fees.
Orbit is quoted per plant once the site survey is done. Do not lead with cost. Nobody re-platforms a factory to shave a licence line, and nobody should — lead with the fact that the middle of the plant is currently unmeasured.
Light the middle
Keep the line running
Close the loop
Stock accuracy, WIP visibility and the label under control. The label work alone usually pays for the phase, because a rejected pallet is expensive and entirely avoidable.
Orbit · Ledger · TAG · Cipher · Loftware
Downtime with a cause attached, maintenance as jobs rather than phone calls, tooling and calibration under custody, and labour cost per unit as a daily number.
Orion · Flux · Pulse
The gate, outbound delivery, the brand owner's portal, controlled documents and the ISO evidence trail. At this point the recall question has a one-minute answer.
Ephoro · Relay · Fuse · Vault · Beacon · Core
Longer than the twelve months a smaller operation needs, and deliberately so. A two-shift plant cannot absorb change at the same rate, and a phase that disrupts the line costs more than it saves.
A model, not a measured result from a client. Output 39 600 units a month at an illustrative R180 contribution per unit. Every assumption is on the page so it can be argued with.
| Where the money is | Assumption | Per month |
|---|---|---|
| Unplanned downtime, with a cause attachedMeasuring stoppages by cause is what finds the repeat offenders | Recovering 2 percentage points of line availability — 792 units at R180Read the caveat below before using this line. | R142 560 |
| DC rejections that stop happeningOne approved template set, printed and applied by machine | 6 pallet rejections a month at R4 500 all-in — return transport, re-labelling, the credit | R27 000 |
| Component shrink and obsolescence found earlyMonthly cycle counts instead of a quarterly wall-to-wall | 0.15% of R14m of component stock | R21 000 |
| Stocktake labourThe count becomes a walk with a reader | Quarterly count cut from 3 days to 1, 18 people at R60/h, spread monthly | R5 760 |
| Modelled monthly benefit | Conservative end of every range above | R196 320 |
Recovered downtime only turns into money if the plant is selling everything it can make. On a capacity-constrained line, two points of availability is two points of output and the figure above holds. On a line running to demand, the same two points show up as overtime avoided and a shorter second shift instead — a real saving, but a much smaller one, and it belongs in a different row.
Establish which of the two the client is in the first meeting. Presenting the capacity-constrained number to a demand-constrained plant is the fastest way to lose the room, and it is an easy question to ask.
R196 320 a month at the conservative end of every assumption. Set it against the monthly and once-off figures quoted in your proposal. A plant carries a larger implementation and training cost than a smaller operation, so it takes longer to earn back — the trade is that the benefit concentrates in a single lever, and the lever is a big one.
Note what is not claimed: no revenue growth, no headcount reduction, no yield improvement and no saving on the software licences themselves. Everything above comes from measuring things that are currently estimated.
The first of these decides the deal, and it should be said in the first meeting rather than discovered in the third.
Ledger costs and moves stock, but it does not explode a BOM, schedule production or run MRP. A plant that needs those keeps its ERP, and Orbit sits alongside it rather than replacing it — which changes the pitch but not the value, because everything in this blueprint lives in the middle of the plant, not in the planning layer. It is a strength when volunteered and a problem when found.
Orbit records what people and read points tell it. It does not connect to line controllers, so cycle times and counts come from RFID events and operator scans rather than from the machine itself. For most plants that is enough to go from nothing to measured; where an MES already exists, Orbit complements it rather than competing.
Metal chassis and liquid-filled product detune UHF tags, and a plant is full of both. On-metal tags, antenna placement and read-zone shielding are a design exercise — which is why NBS quotes fixed readers only after a site survey. A quote given without one is a guess.
Label management is a third-party product that NBS resells, implements and supports. It is priced on printer and user count, not included in the Orbit subscription, and it should appear as its own line in any proposal. It is also the single highest-leverage item in this blueprint.
Android allows one enterprise mobility manager per device, and switching it is not a settings change — every handset and tablet has to be factory-reset and re-enrolled. Plan it as a scheduled exercise with the fleet in a room, not as a background migration. NBS has done exactly this moving its own fleet off Hexnode, which is why the line is in the plan rather than discovered in week three.
This blueprint is built on discrete assembly. The seven stages hold across most of manufacturing — what changes is which stage carries the weight, and which hardware line matters.
| Sector | What changes | Where the weight moves |
|---|---|---|
| Food & beverage | Date and lot coding becomes the critical print rather than the label, allergen and batch traceability tighten, and the cold chain runs to the delivery point. HACCP records join the ISO evidence in Vault. | Printronix C170 Food Grade coding, Relay for cold-chain sequencing, and stage 7 traceability becomes the primary sale rather than the payoff. |
| Plastics & packaging | Tools and moulds become the central asset — shot counts, service intervals and where a mould physically is. Changeover time, not breakdown, is the dominant downtime lever. Regrind and scrap need their own stock treatment. | Flux and Orion carry far more weight: the mould is an asset with a shot-count-driven service schedule. Stage 5 becomes the lead stage. |
| Automotive tier supplier | Serialisation and documentation dominate. Label compliance runs to the customer's own standard rather than a retail one, and the customer portal stops being a nice-to-have and becomes contractual. | Loftware and Fuse move to the front. Stage 3 serialisation and stage 7 traceability are the whole conversation; the label standard is set by the customer, not by you. |
| Job shop · low volume, high mix | WIP visibility matters more than throughput, because the question is “where is that order” rather than “how many did we make”. Downtime matters less; setup and scheduling matter more. | Stage 2 and stage 3 read points carry it, and the absence of a works order in Ledger becomes a harder objection — qualify this one carefully before pitching. |
All thirteen applications behind one Orbit sign-in — click a planet to explore what each one does.
Enterprise label management — one approved template, business rules, versioning and an audit trail.
Readers, antennas, gates, handhelds, printers and tags — the layer this blueprint stands on.
The same suite across a road freight operator — the order, the depot, the road, the yard and the POD-to-cash cycle.
The same suite across a livestock agency and red-meat abattoir, from farm gate to butchery.
