← All posts

An inventory and stock-taking app for warehouses and shops

If stock counts still happen on paper or in a spreadsheet, a mobile app on the warehouse floor can make the count faster and the numbers trustworthy. Here is what that app needs, and how to scope the first version.

Author
Elie Chaaban
Published
Reading time
8 min
Topics
Mobile Apps, Inventory, Warehouses, Planning

The stock figure in the spreadsheet is confident. The shelves are less so. Someone counted at the end of last month, the notes went into a sheet, and since then goods have arrived, gone out and moved, with a few of those movements written down and a few not. When you next need a reliable number, the team books a weekend and starts again with clipboards.

An inventory and stock-taking app is a phone or handheld tool for that work. Staff walk the aisle, scan a barcode, enter a quantity, and the count is already attached to a location and a user. The office sees progress and variances without retyping. That is the job. It is not, in its first version, a complete warehouse management system.

This article covers when a custom app is worth it, what it should include, what can wait, and how I usually build it.

When a stock-taking app is the right tool

The pattern fits when:

If you have a short item list on one shelf and one person who already keeps the sheet honestly, an app may be more than you need. If counts take a team, stall the operation, or produce numbers nobody uses for ordering, it is worth talking about.

A laptop in the office does not solve this. The work happens standing up, often in poor light or a basement, sometimes with a weak signal. That is why this product is usually a mobile app, with a small admin website for the office.

Why the phone's hardware matters here

Barcode and QR scanning is the main reason to be on a phone rather than a clipboard. The camera reads the code, the app opens the right item, and the user types a quantity or confirms a suggested one. When a label is torn, there has to be a fallback: search by name or code, or a short list of items in that location.

Location also matters, in two senses. The app should always know which warehouse, zone, aisle or bin the user is counting, so a quantity is never saved "somewhere". On a large site, a scanned location label identifies the shelf, not GPS. GPS can help pick the site when someone works across several buildings; it should not be the only way you identify a bin.

Then there is connectivity. Warehouses, cold rooms and underground storage are exactly the places where a signal drops. A count entered in a dead zone has to be saved on the device and sent when the phone is back online, without creating duplicate lines. I wrote about this properly in building apps that keep working on unreliable connections. For a stock-taking app, offline support is often a core feature, not a polish item.

What the first version should include

Locations you can walk

Build a simple tree: site, then zone or aisle, then bin or shelf, at whatever depth you already use. Every count and every movement is attached to a location. If you only have "warehouse" and "shop" today, start there. Do not invent a bin system the team will not maintain.

Each location needs a code, a name, and ideally a barcode or QR on a printed label so a user can scan where they are standing.

Items the scanner can find

Your item list is the other foundation: a code, a name, a unit (piece, box, kilo), a barcode where you have one, and a photo when it helps someone confirm they have the right product. If the same product has several barcodes (a single unit and a carton), the app should accept both and still know the unit.

I usually import this list from the spreadsheet or accounting system you already have, rather than asking the team to retype it. Messy codes and duplicates need time. That work belongs in the first version if you want the app trusted on week one.

Count sessions

A count should be a session, not a loose set of numbers. Someone in the office, or a supervisor, starts a count for a site or a set of locations, staff open it on their phones, and the session has a status: in progress, in review, closed.

On the floor the flow is short: scan or select the location, scan the item, enter the quantity, move on. The app should show what has already been counted there so two people do not redo the same shelf without noticing. If a second count of the same item disagrees, the app should flag the variance rather than silently overwrite.

When the session is closed, you have a snapshot: system quantity (if you had one), counted quantity, variance, who counted, and when. That snapshot is what the office reviews.

Adjustments with a reason

Not every change is a formal count. Items are damaged, found, written off, or corrected after a receiving error. A simple adjustment screen, with a reason and a user, keeps those movements out of informal notes. A short list of reasons is enough in version one.

Transfers between locations

If you regularly move stock from the warehouse to the shop, or between sites, a transfer (quantity out of one location, into another, with a status of in transit if needed) belongs in the first version. If you almost never move stock, it can wait.

Receiving, in a small form

Recording an inbound delivery against a supplier document or purchase order, and putting quantities into a location, is often worth including early, because this is how the "system quantity" stays honest between counts. A full purchasing or vendor portal is a different product. The stock app only needs the moment goods hit the floor.

An admin desk for the office

The phone is for walking. The office needs a web admin to manage items, locations and users, to start and close count sessions, to review variances, and to export a report. I go through the general shape of that desk in what your app's admin panel actually needs. For this product, the variance report and the session review are the screens that matter most.

Roles should be tight: counters on the floor should not be able to edit the master item list or close a session they should only submit. Supervisors review. An admin can do both.

What can wait

A lot of warehouse software is built around picking, packing, shipping, wave planning, replenishment and hardware integrations. Most businesses looking for a stock-taking app do not need that in version one. Full pick-and-pack flows, conveyor or laser-hardware integrations, demand forecasting, and multi-company logistics can usually wait.

If you already work with lot numbers or expiry dates, we include them from the start, because adding them later means recounting the truth. If you do not use them today, we do not add them for completeness. The same test applies to serial numbers: include them only if you already cannot ship without them.

Connecting it to the system you already use

The app should not become a second inventory book that drifts away from finance. The usual pattern is:

The first release can start with a controlled import and export, as long as everyone knows which system is the authority for quantities. A live two-way sync is valuable, and it is also a good way to delay the first count. I would rather your team complete a real count in the app, then tighten the sync.

Languages, devices and day-to-day use

If the floor team works in Arabic, English or both, the app has to as well, including a right-to-left layout when Arabic is on. This is much cheaper to do from day one. The same goes for the admin if the office uses a different language.

You do not need a special industrial device to start, as long as the phones are sturdy and the scanners work in your lighting. If you later want dedicated handhelds, a React Native app can still be the client, talking to the same backend. Large buttons, a one-handed flow, and a design that works with gloves are requirements, not decoration. I want to watch a real count, or review a recording of one, before the screens are finished.

How I build it

The work is mobile app development: a React Native app for iOS and Android, a backend in Next.js or .NET Core with Postgres, and an admin website. I deploy that stack with Docker on a Linux server, behind a CI/CD pipeline, and I publish the app to both stores when the users need it there. For an internal team we can distribute it more tightly, but it is still built to App Store and Google Play standards so you are not trapped.

NeonTask is a useful comparison, not a stock-taking app. It is a field-service platform I designed and built for facility-maintenance companies. Technicians use a React Native app for jobs, check-ins, notes and photos, and the office runs the day from a web app. The lesson I would carry here is the same: the people on their feet get a short, reliable mobile flow; the office gets the overview; one backend holds the records.

If your operation also involves technicians visiting sites, a field-service product is a better starting point than this one. If the main pain is the count, stay with a stock-taking first version and do not bolt on job scheduling to get it started.

What to send me

A useful starting note is short: what you sell or store, how many locations and items, how you count today, and which system holds the official numbers. If you can share a sample of the item list and a recent count sheet, even better.

Email hello@eliechaaban.com or use the contact section, and we can walk through a real count on a call and agree a first version your team will use on the floor.

More from the blog

Let's work
together