Running stores and counting stock
Encoding, receiving, transfers, counts, checkout, returns and locate, on one platform.
See the workflowsRFID software development
Custom RFID software development for the tags, readers and printers you choose
We build the software that sits between your RFID hardware and your stock record, and we integrate it with the tags, readers and printers you already use. You buy the hardware from your own supplier. We make it do the job you bought it for.
Tell us what you have and what it should do
Your tags, readers, printers and the workflows you need. We usually reply within 24 hours.
Tags on the shelf, and the number still does not add up
If two of these are true, the hardware is probably fine
- Stock sent from one store arrives at the next, and the two counts do not agree.
- The handheld finds the tags, but the stock number in your system still gets corrected by hand.
- A new reader vendor means asking whether the app has to be rewritten.
The tags and readers are doing their part. What is missing is the software that matches the reads to the stock record and puts a receiving or transfer mismatch in front of a person. That is the part we build.
Adding RFID to your own product
One app over several reader vendors, with an adapter for each vendor SDK.
See how readers plug inShowing products at a trade show
A tagged product on an RFID reader plate brings up its data sheet and videos.
See the displayTracking samples, assets or work in progress
We can build the same software layer for a lab, a site or a production line.
See what we can buildRead rate and stock accuracy are two different numbers
Read rate is how many tags the reader catches. Stock accuracy is how close your system is to a hand count. Studies put store stock accuracy at about 63% without RFID and about 95% with it. So the software has to reconcile reads against the record, not just collect them. When anyone quotes you an accuracy figure, ask which of the two it is.
Show data table
| Measure | Value | Target | Range |
|---|---|---|---|
| Store stock accuracy without RFID | 63% | 100% | 63% to 100% |
| Store stock accuracy with RFID | 95% | 100% | 63% to 100% |
RFID integration, from your readers to your stock record
Tags and printers
Items are encoded and printed with their tags in store.
Handheld and fixed readers
Your chosen readers pick up the tags in range.
One adapter per reader vendor
Each vendor SDK gets its own adapter, so adding a vendor means adding an adapter, not a new app.
Filter and reconcile
We filter duplicate and stray reads, and match the rest against the stock record before anything acts on it.
Mismatch review
A receiving or transfer mismatch goes to a person, who accepts or corrects it. Stock is not adjusted automatically.
Your stock record
The count your staff and reordering rely on.
On the project we shipped, one Android scanning app runs over three reader vendors' SDKs, each through its own adapter. A new reader vendor means a new adapter, not a new app.
Stock lifecycle software, from label to locate
Our RFID inventory management software development covers seven workflows on one platform, in production. A receiving or transfer mismatch goes to a person before stock changes.
- Encoding and printing
Tags are encoded and printed in store, using the platform's own numbering.
- Receiving
Incoming stock is read in and checked against what was expected. A mismatch goes to a person.
- Store-to-store transfers
Stock read out at one store is read in at the next. A mismatch goes to the same review.
- Stock counts
A handheld sweep is reconciled against the record, so the count reflects what is on the floor.
- Checkout
Items are sold on the platform's own sale screen, which reads the tag.
- Returns
A returned item is read back into stock.
- Locate
Staff find a single tagged item with an RFID reader.
What one shipped project changed
Full stock audit
Before2 to 3 staff over a full weekend
AfterTen minutes
Measured
| A full stock count | Before | After | Measured |
|---|---|---|---|
| Full stock audit | 2 to 3 staff over a full weekend | Ten minutes | Measured |
Measured on the client's own operation. The case study shows how each figure was obtained, without naming the client.
5 working daysMissing or misplaced stock flagged within five working days
Measured
8%Sales up after the scanning apps were merged
Client-reported
40%The client's own scanning software cost down 40%
Client-reported
*Disclaimer: Each figure carries its label. Measured means we saw the number, and client-reported means the client told us and we did not measure it.
Metal, liquids and your own products
Metal reflects the signal and water absorbs it. In a university lab test, a tag on metal lost up to 12 dB of antenna gain, against at most 1 dB on cardboard. There are fixes at the tag, at the read point and in the software. As atlasRFIDstore puts it, "testing is always key". So test the tags on your own products, in your own space, before anyone builds software around them.
Show data table
| Item | Value |
|---|---|
| Cardboard (up to) | 1 dB |
| Water (up to) | 6.4 dB |
| Metal (up to) | 12 dB |
Built for a client's trade show booth
An RFID product display for your trade show booth
Place a tagged product on the plate
The RFID reader plate has its own built-in antenna, so your visitors only have to set the product down.
The plate reads the tag
An industrial-grade Raspberry Pi takes the read.
The device finds the product's media
Data sheets, pictures, test reports and videos are loaded onto the device before the event.
The monitor shows it
An industrial-grade monitor brings it up in under two seconds.
We can build a display like this for your booth. You buy the plate, the Pi and the monitor; we build the software and install it on them. Tell us how many products you show and what each should bring up.
Other uses we can build
We have not delivered these three yet. Each would use the same reader adapters, filtering and mismatch review as the stock lifecycle above, and we would scope it with you from the start.
| What is tagged | Where reads happen | What record the reads update | |
|---|---|---|---|
| Laboratory samples | What is taggedSample containers | Where reads happenAt intake, storage and each hand-off between benches | What record the reads updateThe record of where each sample is and who has it |
| Assets | What is taggedEquipment and tools | Where reads happenAt doors, store rooms and handheld checks | What record the reads updateThe asset register and who last had each item |
| Work in progress | What is taggedParts, batches or carriers | Where reads happenAt each station on the line | What record the reads updateThe production record of what is where |
Why pilots stall, and the order we work in
Between 2018 and 2020, retailers moved past pilots. In Accenture's survey, North American retailers only piloting RFID fell from 46% to 8% in two years. When a pilot does stall, Checkpoint Systems, citing Bill McBeath, says the reasons "have nothing to do with the technology itself". One of them is a use case nobody defined. So the scope we write with you starts with the use case: the workflows, the readers and printers, and what the software must do with each read.
Show data table
| Dimension | 2018 | 2020 |
|---|---|---|
| Only piloting | 46% | 8% |
| Fully adopted | 28% | 47% |
How long each step takes is set in your scope.
- You tell us what you have
Your tags, readers, printers and the workflows you need.
- Written scope
The workflows, the readers and printers, and what the software must do with each read.
- Build
Where the scope allows, the first workflows go live before the rest are built.
- Handover
Your code, your hosting and your accounts from day one. If you move the work in-house or to another partner, a written runbook goes with it.
- Support
The support agreed in the scope.
What makes your build bigger or smaller
Industry guides split the cost of an RFID system into start-up and recurring. You buy the hardware and the tags from your own supplier, so they sit on your bill and not on ours. Our part depends on the seven things below.
| Keeps it smaller | Makes it larger | |
|---|---|---|
| Workflows | Keeps it smallerTwo workflows in a first phase | Makes it largerAll seven, from encoding to locate |
| Reader types | Keeps it smallerOne reader vendor, so one adapter | Makes it largerSeveral reader vendors, or fixed readers at doors |
| Catalog and reordering | Keeps it smallerCounts only | Makes it largerA shared product catalog, and reordering from the reconciled count |
| Printing in store | Keeps it smallerTags arrive already encoded | Makes it largerEncoding and printing in store, which is its own cost line |
| Sites | Keeps it smallerOne store or one warehouse | Makes it largerMany sites, each with its own setup and rollout |
| Your products and space | Keeps it smallerCardboard, fabric, open shelving | Makes it largerMetal, liquids or dense racking |
| Start-up and recurring | Keeps it smallerSoftware set-up and read tuning | Makes it largerOngoing support and maintenance |
- Bring these to the enquiryThe tags you use or plan to use
- Bring these to the enquiryYour reader and printer models
- Bring these to the enquiryThe workflows you need first
- Bring these to the enquiryWhere your stock record lives today
- Bring these to the enquiryHow many sites, and what your products are made of
With these, our first reply can be specific.
When you do not need us
We are a good fit when
- Your workflows do not fit a packaged app, or you have outgrown one
- You run more than one reader vendor, or expect to
- You need counts reconciled and mismatches reviewed by a person, not written straight to stock
Go elsewhere when
- A packaged app already counts and prints on your platform with your reader. Buy it.
- You need hardware: go to a hardware supplier, as we do not sell or resell it.
- A trading partner mandates a particular tag encoding standard. Raise it with us before anything else, so we can tell you plainly whether we are the right team.
Questions before you start an RFID software project
How much does an RFID system cost?
It depends on seven drivers: workflows, reader types, catalog and reordering, printing in store, sites, your products and ongoing support. Industry guides split the cost into start-up (hardware, tags, software, installation and tuning) and recurring (support, maintenance, tags and ribbon). You buy the hardware and tags from your supplier. Our part is set in a written scope.
What software will I need?
An adapter for each reader vendor's SDK, a layer that filters duplicate and stray reads, the apps your staff use for each workflow, and the stock record those reads are reconciled against.
Do I need a software engineer on staff?
Not if a packaged app does what you need. If it does not, we build the software and hand it over: your code, your hosting and your accounts, and a written runbook if you move the work elsewhere.
Will RFID work with liquids and metals?
Often, with the right tags, read points and software, but metal and water cost a tag far more signal than cardboard. Only a test on your own products settles it, so run one before the software is built around them.
How long does it take to deploy?
atlasRFIDstore says an all-in-one commercial solution can deploy in a few weeks and a custom one can take 6 to 12 months. GS1 Singapore says a retail pilot typically runs 8 to 12 weeks. Yours is set in the scope, workflow by workflow.
Can one system serve more than one application?
The stock lifecycle we shipped runs seven workflows on one platform: encoding and printing, receiving, transfers, counts, checkout, returns and locate.
Tell us your tags, readers and printers
Send the tags, readers and printers you have, or plan to buy, and the workflows you need first. Add where your stock record lives today, if you know it. We usually reply within 24 hours.
Tirth BodawalaChief Technology Officer, Atyantik Technologies
- We usually reply within 24 hours.
- We do not sell hardware, so our reply is about the software you need.
Where the outside figures come from
- Stock accuracy: RFID Journal, 24 October 2018, via the Auburn RFID Lab.
- Gain lost by material: J. D. Griffin, Georgia Institute of Technology, 2005.
- Piloting and adoption: Accenture via SCDigest, 21 June 2021.
- Pilot stalls: Checkpoint Systems, citing Bill McBeath, 2018.
- Cost and timing: atlasRFIDstore, RedBeam, Vertical Systems, GS1 Singapore.