P/N
Minimum
Location
Order
Supplier
For small manufacturers
Spend your day making things, with fewer interruptions to chase supplies. Start with one part and one Arda card. When stock runs low, scan to start the reorder and build from there.
P/N
6 months
to pay off the rollout
Zero
new hires or new machines
A precision shop small enough that the owner still ran the ordering. Throughput went up, the rollout paid for itself inside six months, and the purchasing job did not grow with it.
Austere MFG · precision shop, small team
Arda on the floor
Give every bench a clear next step.
The adhesive in the cabinet. The washers in the bin. Put the refill instructions with the parts and let the team scan them from a phone.
- 01
Set your reorder point.
Set the minimum and the order quantity. For this epoxy: reorder at two cartridges and buy six at a time.
01 / Two-part epoxyView screen - 02
Scan when stock reaches the minimum.
Scan the QR code with your phone while you are standing at the shelf. Gloves stay on and work keeps moving.
02 / M8 stainless-steel washersView screen
Never stock out, without hiring anyone to prevent it
A stockout is a reorder that fired too late, or never fired at all. In a shop of 10 to 75 people that reorder usually depends on one person having a spare five minutes, and that person is also quoting the next job. Arda makes the trigger physical: the card on the bin is the reorder, sized to your supplier's real delivery days, so replenishment starts the moment stock runs low. It does not need anyone's attention to work, which is the only reason it is still working three months in.
The system that survives is the one that does not need you.
Nobody was hired to do the ordering, and it shows on the weeks you are busiest
Twenty-eight people. There is no ERP, because the last quote came back at more than the shop spends on material in a quarter, and there is a spreadsheet that was accurate in March. Ordering is not anybody's job, so it lands on whoever is closest to the shelf and least able to stop: the owner between quotes, or the lead between setups. It has worked for eleven years. It fails about four times a year, always in the same week, always the week the shop is full.
Purchasing is the one job nobody was hired for
At thirty people every hire has to pay for itself on the floor, so the ordering gets absorbed by whoever already has the most to do. That leaves a purchasing function with no owner, no hours allocated to it, and no cover for the week that person is out.
It fails on your best weeks, not your worst
Reordering happens in the gaps, and the gaps close the moment the shop fills up. So the shortage lands on the week you are running two shifts and quoting three jobs, which is the week it costs the most and the week nobody has ten minutes to prevent it.
Anything that needs a champion dies when you get busy
The last system did not fail on features. It failed because it needed somebody to feed it, and the only person who could was also running the floor. At this size the question is never whether the software can do the job. It is whether it still runs in the month you stop looking at it.
How does inventory replenishment work in a small manufacturer using Arda?
Arda's inventory replenishment software for small manufacturers starts with one item, not with an implementation. You print a card, put it on the bin, and the first time somebody runs low they scan it with the phone already in their pocket. That fires the order at the quantity and the supplier already written on the card. It is the same inventory replenishment software Arda runs inside large plants, sized so a shop of fifteen can start it on a Tuesday afternoon for nothing.
One part, then the cell, then everything you stock
Nobody at this size can stop for a rollout, so Arda is built to be adopted in pieces you can afford to abandon. Each stage below is a real stopping point. If the first one does not earn the second, you have lost an afternoon and no money. Stage 1, one part, today, $0: you card the single item that stopped your last job, or the one you have paid to expedite twice this year. Whoever reaches for it scans it, no account, no training, no phone but their own. It costs nothing: the Free plan is $0 a month, with unlimited cards and 10 orders a month. What you learn is whether the scan actually happens on a busy day, which is the only thing worth testing first. Stage 2, one cell or one bench, this month: you card everything the people around that first card reach for, so one area of the floor stops running dry entirely. The operators and the shop lead scan, still nothing to log into. Cards still cost nothing, however many you add; it is the ordering that grows, so once you are past ten orders a month, Core is $79 a month for a hundred. What you learn is what you actually consume, written down, for the first time. Stage 3, everything you stock, when it has earned it: you card the rest of the shelf, on-BOM and off, including the stockroom nobody has been through since the move. Everybody scans, including whoever is covering while the owner is out. Growth is $299 a month for unlimited orders, two seats and support; Pro is $599 a month and adds four seats, integrations and automated workflows; Enterprise is talk to us. What you learn is what purchasing has been costing you in hours, now that it is no longer invisible. There is no stage where somebody sits down and enters your whole item list. The list builds itself, one card at a time, in the order the shop actually needs it.
- Stage 1: one part, today, $0
- Stage 2: one cell or one bench, this month
- Stage 3: everything you stock, when it has earned it
Scan the card when a bin runs low
Every stocked item gets a durable card with a QR code, starting with the one part that stopped your last job. Whoever is standing there scans it with any phone, no extra hardware, no login friction.
The order fires automatically
The reorder goes out at the card's quantity to the supplier already on the card. Nobody types a purchase order, and it does not wait for the one person who usually does the ordering to find five minutes.
See every card, pulled to restocked
Pulled, on order, restocked, in one view of the whole shelf. That is the part that matters: it keeps running in the month nobody opens it.
Each card encodes a proven loop sized to your real consumption: reorder point, the two-bin system, min/max levels, and safety stock. The math lives on the card, not in someone's head.
Hex bolt M8×40, zinc
Reorder qty
600 units
Lead time
4 days
Supplier
Fastenal
Safety stock
150 units
Scanned 10:02
PO sent 10:02
Restocked
Start with the one part that stalled your last job
Card it today, scan it tomorrow, and see the reorder fire without anyone being asked to remember. Expand across the floor when it has proven itself.
Make Free Kanban CardsFree plan, no credit card
Small manufacturers that stopped doing their own reordering
Shops in the ten to seventy-five range that started with a handful of cards instead of an ERP quote. None of them added a purchasing hire to get these numbers.
0
stockouts
5x
high-value builds
Fat Fender Garage
Parts manufacturer and custom vehicle shop that had already failed two ERP implementations before it tried cards. Stockouts eliminated across everything they stock, and an order backlog of 365+ days pulled down to 20.
Read the Fat Fender case study →5×
production
80%
less inventory held
Rossmonster
Overland vehicle builder that grew straight through the size where most shops stall out. Cards went on 100% of supplies inside a year, critical components came down from 25+ to 5, and production is five times what it was on that fifth of the stock.
Read the Rossmonster case study →2x
throughput
10 min
a day ordering, from 2–3 hours a week
Assembly Industrial
Fabricator of custom trucks and metalwork. Weekly stockouts gone, and the floor went from four builds at a time to eight in six months, without a bigger shop or a bigger team.
Read the Assembly Industrial case study →Arda vs ERP vs spreadsheets at 10 to 75 people
The three real options for a shop this size, judged on what each one costs you in money and in attention before it does anything useful.
FeatureArdaERP / MRPSpreadsheet / memory
Signal to reorderScan a card, order fires automaticallyManual entry, then a planning runSomeone remembers to look
Covers every stocked itemYes, on-BOM and off, same cardOn-BOM only, misses the restOnly if updated by hand
Shop-floor complianceHigh, scanning is the easy pathLow, the floor skips data entryLow, it rests on one person
Setup timeOne part, one afternoonMonths, plus cleaning up your item list firstFast, and it breaks as you grow
Works with your existing ERPYes, splices in, and works fine if you have noneIt becomes the system of recordDisconnected from everything
Cost to get started$0 on the free plan, unlimited cardsLicense plus implementation, before the first orderFree, until one shortage costs you a week
Frequently asked questions
We have outgrown spreadsheets. What comes next, if we are not ready for an ERP?
Put the replenishment loop on cards first and leave the ERP question for later. A spreadsheet fails at this size for one reason: it depends on somebody updating it, and the weeks it most needs updating are the weeks nobody has the time. An ERP or MRP solves a different problem, which is jobs, routing and costing, and it asks for an implementation, a clean item list and a person to own it. That last one is the resource a 10 to 75 person manufacturer does not have. Cards need none of it. You card one item, the scan fires the order, and the consumption history builds itself from real pulls. If you buy an ERP two years from now, the cards keep running alongside it and the history comes with you.
Do we actually need an ERP yet, or just inventory control?
Probably not yet, if the thing you are actually trying to fix is running out of stuff. An ERP is a system of record for jobs, routing, costing and finance, so buying one to stop shortages means paying for the other nine tenths of it and then finding somebody to staff it. It is worth pricing that honestly before you commit. As of July 2026, Katana starts at $299 a month but puts manufacturing management and the shop floor app in paid add-ons, so a manufacturer is at $498 before onboarding, which Katana lists at $2,000 and marks optional. Cin7's $349 entry tier does not include MRP at all, so a manufacturer starts at $599. Fishbowl states on its own pricing page that an implementation package is required, and does not publish what it costs. NetSuite publishes no price, and its own implementation partners put a small manufacturer's first year in the tens of thousands. None of those products is badly built. They are simply shaped for a company that already employs somebody whose job includes inventory. If what actually stops you is the item nobody was watching, fix the replenishment loop first, for $0, and see whether the ERP question survives the quarter. Side-by-side breakdowns are in the Arda Comparison Hub.
Will this still be running in six months, when we get busy?
Yes, because it needs less of your attention when you are busy, not more. Every system that has failed in a shop this size failed the same way: it needed somebody to feed it, and that somebody was the owner or the shop lead. Scanning the card is faster than any of the alternatives for the person standing at the bin, including walking to the office to tell somebody it is low, so it holds on the days when nothing optional gets done. There is no weekly maintenance, no reconciliation and no admin role to fill. The busy month is when the loop does the most work, not when it quietly stops.
How long does setup take, and who has to run it?
One item takes an afternoon and one person. You pick the item, set the quantity it should reorder at, print the card and put it on the bin. Arda sizes the first cards with you during onboarding using your real burn rate and your supplier's real lead time, so nobody has to derive those numbers alone. After that there is nothing to administer. There are no accounts to provision for the floor, because scanning is free and nobody logs in to scan. There is no item list to load, because cards get added one at a time in the order the shop needs them. And there is no integration to schedule, because Arda works with what you already have and works without it.
We do not have an ERP. We have QuickBooks and a spreadsheet. Does that work?
Yes, and it is the most common starting point at this size. Arda does not need an ERP underneath it, because the reorder signal is the physical card at the bin rather than a number sitting in a system. If you run QuickBooks, Arda connects to it, along with NetSuite, Odoo, Xero, Shopify, Stripe, Airtable, Make and Zapier, plus a REST API. If you run none of them, the cards work exactly the same, and the consumption history they build is the data you would have had to produce by hand to configure an ERP later anyway.
What does Arda cost for a small manufacturer, and can we start for nothing?
Create unlimited items and cards on Free, without Order Queue access. Solo includes one user and up to 10 orders a month after its 7-day trial. Team and Pro are priced per Power User and include unlimited scanning for your team. Compare current plans and billing.
Ten cards, one afternoon, and nobody has to remember
Austere MFG grew throughput without adding the purchasing hire that growth normally forces, and the rollout paid for itself inside six months. Start with the one item that stopped your last job, on us, and let the shelf reorder itself.
Free plan, unlimited cards, no ERP rollout.