Shelves, vans and counts
Inventory App or Spreadsheet? A Small Business Test With a Barcode
A spreadsheet is an excellent inventory system for one honest person with a good memory. The trouble starts with the second person, or the second van. Until then, an inventory app is often a costly way to get the same wrong number on a nicer screen.
November 19, 2026 · 10 min read
The short answer
A small business should keep inventory in a spreadsheet while one person counts a modest list of items in one place. Move to an app when several people change stock, stock lives in several places, or running out costs a job. Try a packaged tool first. As of November 2026, Sortly's free plan covers 100 items and one user. Custom starts at $575.
On this page
- Where the spreadsheet still beats everything you could buy
- The count goes wrong on the second person, not the 500th item
- Rent the shelf count before you commission one
- A barcode only knows a number, and the rest is your job
- Keep a list of movements, and let the count add itself up
- Count one shelf a week and skip the January heroics
Where the spreadsheet still beats everything you could buy
Nobody profits from telling you this, so here it is plainly. If you sell a modest list of items from one room, and you are the only one taking things off the shelf, a spreadsheet is a fine inventory system. It is free or nearly free. You already understand it. You can add a column on a whim. Google Workspace Business Starter is listed at $7 per user per month as of November 2026, so confirm the term at checkout, and that gets you a good deal more than a sheet.
A spreadsheet wins in these specific situations. Keep it, and spend the money on something the business needs more. See spreadsheet versus custom app for the general version of this argument.
- One person touches the stock. If nobody else changes the count, nothing can change behind your back.
- You count on a schedule anyway. A monthly count corrects small errors before they matter.
- A stockout is an annoyance, not a lost job. You order more, you apologize, and the customer waits two days.
- Your items do not move quickly. A shop selling a dozen handmade things a week has no need for a scanner. It needs a good pen.
- You could not tell us today what a wrong count cost you last year. If you do not know, you cannot yet justify a system to prevent it.
The count goes wrong on the second person, not the 500th item
Owners assume that inventory systems are about volume. Mostly they are about hands. Say an appliance repair shop in Boise with two vans and a garage, invented for this example. The owner keeps a good sheet. A helper grabs a control board for a Friday job and means to log it. By Tuesday the sheet says two, the shelf holds one and the van holds one nobody remembers. The owner promises a repair, drives across town and finds the part is not where it should be.

A spreadsheet can rarely reconstruct that. The cell says two. Version history, where your sheet has it, may show who typed a number. It cannot show that a part left for a van, because nobody typed that, and two people working from memory in two parking lots will cheerfully overwrite each other's figures. That loss of trust is the true cost. Once people stop believing the number, they walk to the shelf and look, and at that moment your system has become a decoration that you still have to keep updated.
So here are the signs you have outgrown it. Stock lives in more than one place, including vehicles. More than one person takes things without telling you. A missing part costs a return trip or a job. You sell the same thing in two channels, and online orders need accurate counts, which ecommerce automation deals with. Any two of those and it is time to look at the next section. Any one of them and you might simply fix the habit first.
The first bug in any inventory system is a person in a hurry.
Rent the shelf count before you commission one
| Price as of November 2026 | Limits | Fits | |
|---|---|---|---|
| A shared spreadsheet | Free to $7 per user a month with Google Workspace | Shows what was typed, not what moved; no real scanning | One person, one location |
| Sortly Free | $0 | 100 unique items, 1 user license | A small parts list to try the idea |
| Sortly Advanced | $49 a month; yearly plans are listed lower | 500 unique items, 2 user licenses | Owner and one helper |
| Sortly Ultra | $149 a month; yearly plans are listed lower | 2,000 unique items, 5 user licenses | A small crew with vans |
| A custom web app | From $575, depending on complexity | What you specify, and nothing else | A rule or workflow no packaged tool holds |
| A custom iPhone or Android app | From $3,750, plus Apple's $99 a year or Google's $25 once | Upkeep is yours | Heavy scanning all day, poor signal |
Two cautions on that table. Sortly ties its lower yearly figures to a new customer's first year, so we quote the monthly price here. Read the renewal terms when you sign up. And the page's comparison table lists in-app barcode and QR scanning, but the text we read did not confirm which plans include it, so check before counting on it. Your point-of-sale and your field software may also track stock already. Look there first. Paying for a second system that disagrees with the first is a particularly modern way to be wrong.
If you sell online, your cart usually counts for you. Our Shopify versus custom store page covers that decision, and if you hold very little stock at all, selling without much inventory may be the better plan, and the cheaper one.
A barcode only knows a number, and the rest is your job
Barcodes sound like the expensive part of inventory. They are the simple part. A barcode is a number you can read quickly and without typos. It does not know what the thing is, where it lives or how many you have. Your system has to know that. So the first decision is whose number you use: the code the manufacturer already printed, or labels you print yourself. Use the printed code when your suppliers provide one. Print your own for the bins, the shelves and anything that came without one. A bin label also lets a helper count a shelf without knowing a single part number.
The second decision is what does the reading. A native app can use the phone's camera directly. A web app has a catch. MDN's page on the Barcode Detection API, read November 6, 2026, marks it as experimental with limited availability, not working in some of the most widely used browsers. So a browser-based scanner usually needs a separate scanning library, and testing on your own phones. The usual cheap alternative is a handheld scanner paired with the phone or tablet, most of which type the number like a keyboard and do not care about browsers. For a counter or a garage, this is often the right answer, and it is far cheaper than a build.
Think about the task too. Scanning 40 things at a shelf is fast with a camera in good light. Scanning in a dim van with gloves on is another matter, and that is where the honest case for a native app begins. Try the phones and the light you actually have. A demo in a bright conference room with a brand-new label tells you nothing about the third shelf down.
Keep a list of movements, and let the count add itself up
This is the single design decision that separates an inventory system you can trust from a sheet with a nicer look. Do not store "quantity: 2" as a number people overwrite. Store every movement: received 10, moved 2 to the van, used 1 on a job, adjusted minus 1 after a count, and who and when for each. The quantity is whatever those lines add up to. If you ask a developer about their design and hear only about a quantity field, ask again.
Scan and act
Opens to the camera. Scan a part, then tap one of three actions: used, moved or received. That covers almost all daily work, and everything else can live on a laptop.
Where is it
Search by name or scan. Shows each location, including each van, with its count. This is the screen that prevents the drive across town.
Low stock
A short list of what is under its minimum, and what you usually pay. Tap to add it to a purchase list. A human still orders, because you know the supplier is closed Friday.
Count mode
Walk a shelf or a van and enter what is there. The app logs the difference as an adjustment, with your name on it, rather than quietly overwriting the truth.
History for one item
Every movement in plain order. When the count looks wrong, this is how you find out where it went, instead of blaming whoever is closest.
The records underneath are few: items, locations, barcodes, movements, suppliers and minimum levels. Connect them to a job where you can, so a part used is tied to a customer. That gives you something a sheet never will: what the job really cost. This is how we would scope the work. If it grows into the whole back office, that is an internal business application and deserves to be scoped as one.
Count one shelf a week and skip the January heroics
- 1
Count one area, not the whole building
A short count each week beats a heroic count in January. Pick one van or one shelf, and let the app show what it expected.
- 2
Look at the differences
A pattern matters more than the number. The same part always short means a habit needs fixing, which can be as dull as a basket on the van door for used-part tags.
- 3
Order from the low stock list
Takes a few minutes. Set the minimums from what actually ran out, and raise or lower them as you learn. Do not trust anyone's formula on day one, including ours.
- 4
Pass the totals to the books
Purchases and stock value belong in the accounting system, not here. Connecting QuickBooks settles which system holds which truth, and your accountant decides how inventory is valued for taxes.
If the habit holds and the tools still do not fit, we can quote the work at our published prices: web apps and API integrations, which go live the same day, from $575, depending on complexity; apps from $3,750; and financing for web applications at 0% with 25% down, up to 12 months. Run your own hours through the savings calculator first, including return trips for missing parts. There is no special rate for small companies, and we would rather lose this job than sell you a costly system that your shelves do not require. The microbusiness hub lists what else we build for businesses this size, and (210) 346-0848 reaches a person in San Antonio.
Questions owner-operators ask
There is no fixed number, and we would distrust anyone who gave you one. People and places break a sheet before item counts do. One careful person can run a few hundred items in a spreadsheet. Two people and two vans can make a mess of 40. Move on when the count stops being believed, rather than when the rows get long.
A phone camera works in native apps and many packaged tools. Inside a plain web app, it is less certain, because MDN marks the Barcode Detection API as experimental with limited availability. A small handheld scanner paired to a phone or tablet is often the cheap, dependable choice for a shelf or counter. Test on your own phones and in your own lighting.
Packaged options start low. As of November 2026, Sortly lists a free plan with 100 unique items and one user, and Advanced at $49 a month for 500 items and two users. A custom web app starts at $575, depending on complexity, and a custom iOS or Android app at $3,750, depending on scope. Add upkeep to any custom build.
Treat each van as a location. Moving a part to the van is a recorded movement, and using it on a job is another. Then the app can say which van holds it, which is the question that matters most in a small shop. Count one van a week, rather than everything at once, and your numbers stay close enough to trust.
Usually, and it is worth deciding first which system owns which fact. Often the store or inventory tool owns the count, and the accounting system owns the money. QuickBooks Online has a documented public API, and integrations we build go live the same day, from $575, depending on complexity. Have your accountant confirm how inventory value should be recorded.
Because a single number cannot explain itself. If it says two and the shelf holds one, you have no way of finding out why. Storing each movement, with who and when, makes the quantity a sum you can audit, and makes mistakes reversible. It costs little extra to build, and it is the difference between a record and a rumor.
Keep reading
- Two stores, one budgetMobile App Development for Small Business: iPhone, Android or a Web App FirstiPhone and Android apps for a small business: web app from $575, native from $3,750, Google Play and Apple fees as of 2026, and what to build first.
- Keep it or replace itSpreadsheet vs Custom Application: The Specific Moment the Sheet Stops WorkingA spreadsheet is good software, until one of seven things happens. How to spot the moment, repair the sheet cheaply and price a small app only if you must.
- One owner, many ordersEcommerce Automation for a One-Person Store: Orders, Labels, Stock and EmailEcommerce automation for a one-person business: which order, label, stock and email jobs to hand off first, what they cost, and how each one fails.
- One owner, one phoneWeb Design for One-Person Businesses: What You Need and What to RefuseWebsite design for one-person businesses: the short page list, what to refuse to pay for, and fixed prices from $1,500, or when a builder plan is better.
- The build sideSmall Business Website Development: Who Holds the Keys When the Builder LeavesWebsite development for independent contractors: who owns the code, domain and accounts, what custom means at $1,500, and what you must maintain yourself.
- Probably not. Sometimes yes.iOS App Development for Small Businesses: When an iPhone App Earns Its KeepWhen an iPhone app makes sense for a very small company, what from $3,750 buys, Apple's fees and rules as of 2026, and what to try before you build.
- Everything in MicrobusinessSee the section
From the Insights desk
- Own the Building: Why a Custom Next.js Site Beats WordPress and WebflowWordPress and Webflow both rent you a website — you may decorate, you may not renovate. A custom Next.js build hands you the keys: complete customizability, a CMS shaped to your business instead of a builder's data model, and a site you actually own.
- Rank First, Get Half: How to Create an Effective SEO Strategy for 2027With an AI Overview on the page, the top Google result draws about half the clicks. Here is how to build an SEO strategy for 2027 that still pays, in five plain steps and a 90-day start.
Tell us how stock moves through your shop
Describe your items, your locations and who touches them. A person at DataCram replies and says whether to keep the sheet, rent a tool or scope something custom.
Which of these sounds like you? (pick any)
Rather talk now? Call (210) 346-0848.