History · 5 min read

The spreadsheet that quietly rewrote the office

VisiCalc shipped in 1979 and made a category of work obsolete almost overnight. Nobody planned it.

Before 1979, a financial model was a physical object. Rows and columns drawn on wide green paper, filled in by hand, every figure the product of a desk calculator and someone's patience. Changing one assumption meant recomputing every cell downstream of it, by hand, for hours.

The idea in a lecture hall

Dan Bricklin was a business student watching a professor work through a model on a blackboard. When one number changed, the professor had to erase and redo the chain of consequences, and the room waited. Bricklin later described imagining a blackboard where correcting one figure updated everything that depended on it instantly.

With Bob Frankston he built it for the Apple II, a machine then sold largely to hobbyists. It was called VisiCalc, it cost around one hundred dollars, and it did essentially one thing: a grid of cells where a cell could contain a formula referring to other cells, recalculating whenever anything changed.

The first killer app

What happened next is the origin of the phrase killer application. Businesses bought Apple IIs not because they wanted a computer but because they wanted VisiCalc, and it ran nowhere else. Accountants and analysts brought machines in on their own budget, without asking the data processing department, because the machine paid for itself in the first week.

That detail mattered more than the software. It established that computing could enter an organisation from the edge rather than the centre — bought by the person doing the work, justified by their own productivity. Every wave since, from the personal computer to the smartphone to cloud tools bought on a credit card, has followed the same path.

What it destroyed

The spreadsheet did not eliminate accounting jobs. It eliminated a specific one: the person whose role was to perform the arithmetic. Whole floors of clerks existed to do the recomputation, accurately and slowly. That work vanished within a decade.

The jobs that replaced it were about asking questions. When a scenario takes seconds instead of a day, you run twenty of them, and the valuable skill shifts from computing the answer to knowing which question is worth computing. The people who moved up were the ones who had been thinking about the model rather than only filling it in.

The ending nobody expects

VisiCalc lost. Lotus 1-2-3 arrived faster and better suited to the IBM PC, and later Excel took the category outright. Software Arts was sold off in 1985, six years after shipping the thing that created the industry it lost.

The idea survived intact, though. Open a spreadsheet today and the model is unchanged: a grid, formulas referring to other cells, everything downstream updating at once. Forty-five years on, it remains the most widely used programming environment ever built, mostly by people who would never call it that.