A Bell Built for Honesty

The cash register was not invented to make shopping pleasant. It was invented, according to the story usually told, because a Dayton, Ohio saloon owner suspected his bartenders of skimming from the till. James Ritty, who ran the Pony House Tavern, is said to have gotten the idea from a dial on a ship that counted the turns of its propeller, and he set about building a machine that would count sales instead.

With his brother John, a mechanic, Ritty made a device called Ritty's Incorruptible Cashier. Sources differ slightly on the timing: one account, published by NCR's own history site, dates the invention to 1878, while both sources agree the patent came on November 4, 1879. Either way, the first version was simple. Pressing a key added a sale to a total shown on a dial, and there was not even a cash drawer.

Drawer, Bell and Paper Roll

The Ritty brothers could not keep up with demand and sold the patent. The buyer, Jacob Eckert of Cincinnati, formed a company to make the registers and improved them with a cash till and a bell that rang with each sale. Those additions turned a counting device into something a customer could hear and see. Every sale now announced itself.

By 1884 John H. Patterson had bought the business and renamed it the National Cash Register Company. The Computer History Museum dates the beginning of NCR to that year, when Patterson began making mechanical registers. One account credits the company with adding a paper receipt roll over time, while another describes a paper-roll machine among the Ritty brothers' own prototypes, so the origin of that feature is not settled here. Either way, it gave customers a copy of the record. The receipt and the drawer are the two features that the modern checkout still carries.

Levers and Counters at Work

Inside a mechanical register, pressing a key moved a lever that turned counting wheels, much like the mechanisms in mechanical calculators. The total rose, the drawer sprang open, and the bell rang. The clerk never needed to trust memory, because the machine held the running figure.

That design had a strong social effect. A sale could not be quietly taken, because the machine would have to be operated and the count would show. Owners could compare the day's total on the register with the money in the drawer, and any gap invited a question. What the register could not do was tell the owner what had been sold, since it knew only amounts.

The machine also changed the relationship between customer and clerk. The customer could watch the amount appear and hear the drawer, so a wrong total was easier to spot and dispute. In that sense the register was a shared instrument of trust, visible to both sides of the counter, and that public quality is something screens facing only the clerk do not always preserve.

Bars, Stripes and the First Scan

Knowing what was sold required a different kind of record. In 1970, grocery retailers and manufacturers called for a standard machine-readable symbol, and IBM's design, led by George Laurer, became the Universal Product Code, which IBM says was officially born on April 1, 1973. It was meant to improve inventory handling and checkout.

The first store use came in 1974. Supermarket News reports that Marsh Supermarkets in Troy, Ohio, put an NCR prototype scanning system into service on June 26, and that IBM production systems followed about a month later at stores in Montreal and South Plainfield, New Jersey. The register was no longer just a counter of sums. It was becoming a terminal that identified products and could send information to the rest of the store.

IBM's account stresses that the payoff went beyond shorter lines. Once each item had a code, retailers could track stock and see what sold and what did not, which made a lasting difference to inventory management. Adoption took work, since manufacturers had to print the symbol and stores needed scanners and price files, and the early scanning pilots wrestled with equipment breakdowns and huge numbers of labels. The checkout had turned from a place where money was counted into a place where information was collected, a change that later payment systems built upon.

The Register Becomes a Terminal

A modern point-of-sale system builds on that idea. The counter device is a computer: it reads codes or taps on a screen, looks up prices, applies tax rules, takes payment and updates stock. The same record feeds accounting, ordering and reports, so the checkout is now a data source for the whole business.

Payments changed as well. Card networks pushed retailers toward EMV chip readers, and on October 1, 2015, in the United States, liability for certain in-store card fraud shifted from card issuers toward merchants whose terminals did not support chip cards, according to a law firm's client alert. The old drawer of cash remained, but the terminal increasingly handled money that never touched a hand.

Power Cuts, Queues and Uptime

In a blackout, a mechanical register keeps working, because it is powered by the clerk's hand. A modern terminal depends on electricity and often on a network, though many can fall back to limited offline modes. That gap matters most for small businesses in places where power or internet is unreliable.

When everything works, the balance tips the other way. A scanner typically handles a full basket faster than a keyed register, and a stored price list avoids many typing errors. Speed at scale is the strongest argument for the newer system, in the same way that the calculator improved on the bead frame, a story told in abacuses and electronic calculators.

Fixing Gears, Replacing Modules

A mechanical register could be repaired by someone who understood its gears, and many old machines still work. A point-of-sale terminal usually has few moving parts, but its electronics and software are replaced as units, and its useful life often depends on a vendor's support and on changing security rules.

The data question is new. A mechanical register recorded little beyond amounts. A modern system can store what each customer bought and how they paid, and may pass information to payment processors and other services. That gives owners valuable insight and creates a responsibility to protect what is collected.

There is also a matter of scale. A single shop with a small range of goods may find little to gain from a database, while a store with thousands of items would struggle to operate without one. The choice is less about which technology is more advanced than about how many products, clerks and payment types the business has to handle, and how much it values being able to run without electricity.

What the Brass Machine Still Shows

The old cash register is a reminder that good design can be simple: a total, a drawer, a bell and a receipt. Those elements are still present in the checkout of today, translated into software and screens. Its logic of making each sale visible also echoes in other access and payment systems, such as those in paper tickets and mobile passes.

The modern terminal earns its place through capability: scanning, card payment and records that reveal how a business runs. Rather than choosing a winner, it is more accurate to see a continuous line from Ritty's dial to today's screen, with each stage adding something and asking something in return.

A contextual conclusion

Mechanical registers and point-of-sale systems belong to different eras of retail, and each has a clear place. The mechanical machine is durable, self-contained and easy to grasp, while the electronic system scans, accepts cards and turns each sale into usable business information. Neither is simply better. A small stall with a few prices might still suit a simple register, while a shop with many products and card payments benefits from a connected terminal.

  • Best for simplicity and independence from power Mechanical Cash Registers — A mechanical register needs no electricity, software or network to record a sale.
  • Best for speed with large product ranges Point-of-Sale Systems — Barcode scanning and stored prices remove the need to type every item.
  • Best for tracking stock and sales Point-of-Sale Systems — Each sale updates records that can be turned into reports without manual counting.

Historical impact

The cash register changed retail by giving owners an accurate, checkable record of every sale, which made it easier to trust employees and to run larger shops. Its bell and receipt also made transactions visible to customers. Point-of-sale systems extended that record into inventory, pricing and payment, turning the checkout counter into a source of business information as well as a place to pay.

How the two are related

Point-of-sale systems are the direct successors of the cash register, and the company most tied to the register, NCR, was also involved in early barcode scanning. The core idea, recording each sale at the moment it happens, carried over unchanged, while adding a computer let that record be connected to products, payments and stock. In effect, the register became a terminal.

Sources consulted

  1. Ritty's Incorruptible Cashier (1879), The American Table. Ritty brothers' dial machine with no drawer, November 4, 1879 patent, Eckert's till and bell, Patterson and NCR by 1884.
  2. The Inventor of the Cash Register: James Ritty, Ritty (NCR history site). Dates Ritty's invention to 1878 and the patent to 1879; describes dial, tablet and paper-roll prototypes and Patterson's purchases.
  3. National Cash Register Company (NCR), Computer History Museum. NCR began in 1884 making mechanical cash registers and later moved into computing.
  4. The UPC, IBM Heritage. Universal Product Code officially born April 1, 1973, the 1970 industry request, George Laurer, and the first scan.
  5. The Birth of Scanning, Supermarket News. June 26, 1974 Marsh Supermarkets NCR prototype; IBM systems in Montreal and New Jersey a month later; early pilot challenges.
  6. Retail Industry EMV Compliance Deadline Arrives Today, McGuireWoods. October 1, 2015 shift of in-store card-fraud liability toward merchants without EMV chip readers.

Dates and figures in this article are limited to those supported by the sources above. Something look wrong? Report a correction.