How budgeting built the spreadsheet

Who invented the spreadsheet? The history of the spreadsheet offers several answers, and for anyone in FP&A they share a flattering detail: every time someone built one, the problem on the desk was a budget. Your budget workbook is doing the job the tool was made for.

This page is part of the research behind In defence of Excel, which argues that finance didn’t adopt the spreadsheet so much as specify it. Here is the historical evidence, from the first budget models to how Excel beat Lotus 1-2-3. Where an inventor is describing his own work, we say so.

Who invented the spreadsheet?

It depends on what you count. What the candidates share is the problem: each time, someone needed a budget to recalculate.

The accountant’s spread sheet

The name came first, and it came from accounting. Richard Mattessich traces it to the first edition of Eric L. Kohler’s Dictionary for Accountants (1952), where a spread sheet “refers to a worksheet providing a two way analysis of accounting data”. Dan Bricklin, who conceived VisiCalc, is plain about it: “First of all, we did not invent the word ‘spreadsheet’.” VisiCalc took the name “precisely because it was like the paper spreadsheets, such as the ones I did as a student at Harvard Business School.” The electronic version copied the ledger pad on purpose, which is why a new analyst can read a model without being taught how.

A budget simulation, 1961–64

The first computerised spreadsheet was a budget model. In July 1961 Mattessich, then an accounting professor at Berkeley, published “Budgeting Models and System Simulation” in The Accounting Review. His 1964 book, Simulation of the Firm through a Budget Computer Program, included the program itself, written in FORTRAN IV by two research assistants. An encyclopedia of accounting history, quoted on Mattessich’s own page, says the book “foreshadows the basic principles behind today’s computer spreadsheets: the use of matrices, (budget) simulation, and, most important, the calculation that support each matrix cell.” That is the annual budget workbook, designed in 1964. It had to wait for cheaper machines: as Mattessich notes, “computerized spreadsheets became only popular in the 1980s, after the introduction of Personal Computers.”

LANPAR and the telephone companies, 1969

The second invention was built for corporate budget departments. In 1969 Rene Pardo and Remy Landau created LANPAR which, in Pardo’s own account, was “licensed to and used by the budgeting department of AT&T and every telephone company in North America throughout the early 1970’s”. Their patent, filed in 1970 and granted in 1983, describes a compiler that keeps scanning the formulas and queues each one for calculation once its inputs are defined, “until all formulas have been defined”. So formulas can be entered in any order. Pardo calls this “Forward Referencing, the cornerstone functionality of modern spreadsheets”. It’s why a formula can point to a cell further down the sheet, and it began as a fix for budget forms.

Bricklin says of these earlier programs: “Most of these were not something that Bob or I had seen.” The need was simply the same each time: recalculate linked budget figures cheaply, and let people who aren’t programmers change them.

VisiCalc: an MBA student’s magic blackboard

The third invention is the one that reached everybody. Bricklin dates the idea to “the spring of 1978”, when he was an MBA student at Harvard Business School working cases by hand. In a 2016 talk he recalled the frustration: “you’ve done all your homework, you come in the next day only to find out that you made an error and all of the other numbers you did were wrong.”

“I imagined a magic blackboard that if you erased one number and wrote a new thing in, all of the other numbers would automatically change, like word processing with numbers.”

His journal dates the first Apple II prototype, built over a weekend, to 8 October 1978. He tested it on a case about the Pepsi Challenge: “I did five-year projections. I did all sorts of different scenarios. I aced the case.” Steven Levy’s 1984 account names a different assignment, a projection of “one company’s acquisition of another”, and neither source reconciles the two. Both are corporate-finance projections.

The publisher, Dan Fylstra, recalls Bricklin pitching “an ‘electronic blackboard,’ useful for building budgeting and financial planning models.” Bricklin’s finance professor thought the modelling tools of the day, which ran on time-sharing systems (shared mainframes rented by the month), already served that market. Fylstra disagreed:

“I could see that this tool for budgeting and planning would offer a far better human interface than the established, command line oriented financial modeling tools on time-sharing systems. I had the youthful chutzpah at the time to predict that we would make the time-sharing based financial planning products obsolete. As far-fetched as it seemed in 1978, we proceeded to do exactly that with VisiCalc.”

Earlier forecasting tools, Bricklin notes, “ran on timesharing systems and cost >$1000/month or used card input”, while VisiCalc “ran on an affordable, personal machine, so it was accessible to all.” Finance shaped the details. Shown the program, Professor Glauber told Bricklin: “Net present value should be in it… you should have IRR.” Its first buyers used it the way finance would. Writing in 1984, the computer scientist Alan Kay observed that most people who bought VisiCalc “exploited it to forecast the future rather than to account for the past”.

How the spreadsheet sold the PC to business

A killer app, as Thomas Haigh defined it when interviewing Fylstra for the Computer History Museum, is “an application that was sufficiently compelling that people would go out and acquire the entire platform just to get access to that one application program.” The museum says VisiCalc “arguably became the first killer app, selling tens of thousands of Apple II computers”, though Fylstra adds that “the Apple II was essential for VisiCalc as well.” In Bricklin’s words, VisiCalc “was a catalyst to the personal computer industry, by introducing personal computers to the financial and business communities and others.”

Sales climbed fast. The first copy shipped on 17 October 1979 and 1,293 went out that month, according to Peter Jennings of Personal Software, VisiCalc’s publisher. At the peak, Fylstra says, royalties to the program’s authors ran at “about a million dollars per month”. Lotus 1-2-3 followed in January 1983 at $495. In the account of the economists Stan Liebowitz and Stephen Margolis, it “was immediately acknowledged to be a better product than VisiCalc”, and by October 1983 PC World reported that 1-2-3 was outselling VisiCalc. Switching was painless: 1-2-3 “could open VisiCalc files and it emulated most VisiCalc keystroke commands”, Fylstra notes.

By 1984 Levy could report that VisiCalc alone had sold over 700,000 copies, that “More than a million computer spreadsheet programs worth more than $250 million will be purchased in the United States this year”, and that “every student in the Harvard Business School is now required to be proficient in using electronic spreadsheets.” The historian Martin Campbell-Kelly sums up the decade that followed: the spreadsheet became “an indispensable tool of modern business, employed by tens of millions of users who had little or no direct computer experience.”

What Levy saw it do to business thinking

Levy’s Harper’s essay, “A Spreadsheet Way of Knowledge”, was written while the change was under way. He likened its effect to that of double-entry bookkeeping in the Renaissance: “The electronic spreadsheet is to double entry what an oil painting is to a sketch.” What he recorded reads like the operating model of modern FP&A. “The spreadsheet is a tool, and it is also a world view”, he wrote, a view of “reality by the numbers”. Users were charmed by “the ‘what if’ factor: the ability to create scenarios, explore hypothetical developments, try out different options”. Bob Frankston, VisiCalc’s co-author, told him that people who once “would have taken a guess” now “feel obligated to run the numbers.” And a CPA, Allen Sneider, described what it did to planning:

“Want to see what happens with a different return on investment? Sheltering? Interest rates changing by half of a percent? It’s done in a minute. Before you’d be tempted to say, ‘Let’s leave it the way it was.’ The whole mental attitude toward preparing projections has changed.”

Levy also saw the spreadsheet “diminishing the power of data processing”, moving models into the hands of the people who use them. And he added a fair warning: the benefits “will be meaningless if the spreadsheet metaphor is taken too much to heart.” A model is only as good as its assumptions.

What economists found in the spreadsheet market

By the late 1980s several spreadsheets were competing, and economists used the market to test for network effects: a product has them when each extra user makes it more valuable to everyone else. Two studies used hedonic pricing, a statistical method that splits a price into what buyers pay for each feature, and both found the effect.

Neil Gandal (1994) found that buyers paid “a significant premium for spreadsheets that are compatible with the Lotus platform and for spreadsheets that offer links to external databases”. Erik Brynjolfsson and Chris Kemerer (1996), using prices from 1987 to 1992, found that “a one percent increase in a product’s installed base was associated with a 0.75% increase in its price”, the installed base being the number of people already using it. And “products which adhered to the dominant standard, the Lotus menu tree interface, commanded prices which were higher by an average of 46%.” Their explanation is the part worth keeping:

“A switch by an organization to another user interface would require an investment by every user in re-training, whereas incompatibility in reading and/or writing files could be solved through one-time construction or purchase of a converter.”

Their data bear it out: 98% of the product observations could already read Lotus files, but only 49% copied the Lotus menus. The scarce asset was the trained user. The authors note that their estimates moved with the sample and method, so the percentages are best read as direction and rough size. Liebowitz and Margolis add the automation we build into our files: “many users, particularly business users, are loath to switch to a spreadsheet that disables their old macros.” Macros then, and VBA and queries now, are the stickiest part of a finance stack.

Meanwhile the spreadsheet got far cheaper. Adjusted for quality, prices fell by about 15% a year from 1986 to 1991 in Gandal’s data and 16% a year in Brynjolfsson and Kemerer’s. By 1997, Liebowitz and Margolis report, the typical price a vendor received was about $50, “a fall of over 80 percent from the typical price in 1988”. That is how a spreadsheet came to sit on every finance desk by default.

How Excel beat Lotus 1-2-3

If the network was that strong, why did Excel become the standard? The economists’ answer is quality, helped by timing.

Graphical from the start, and first on Windows

Excel debuted on the Macintosh on 30 September 1985, the date Microsoft gives. It was, as Liebowitz and Margolis put it, “created to run under a graphical operating system”, and at first “offered only for the Macintosh”. In late 1987 Microsoft ported it to the PC under Windows, a gamble because “Windows’ success was far from assured until version 3.0, which was not available until 1990.” Lotus and Borland, meanwhile, “did not produce Windows versions of their programs until after Windows 3.0 had proven itself as a successful operating system.” When finance desks moved to Windows, Excel was already there.

Brynjolfsson and Kemerer’s data show why that mattered. The premium for the Lotus menus was half as large across their whole sample as among DOS products alone, which, they write, “supports the argument that the initial absence of a Lotus-based product allowed entry and significant network creation by Microsoft Excel.” A new platform “‘levels the playing field’ and erases the advantage of the incumbent network.”

Reviewers preferred it for years

“Microsoft Corp. has just unleashed a spreadsheet that makes 1-2-3 look like a rough draft”, PC Magazine wrote in November 1987. Across ten years of head-to-head reviews, Liebowitz and Margolis found, “Lotus 1-2-3 just barely avoids a shutout, managing but a single win.” Excel won the reviewers years before it won the market.

It came in the first office suite

“Not only was Excel the best Windows spreadsheet, but it was also the first Windows spreadsheet. It was part of the first office suite, and it was also part of the best office suite.”

Liebowitz and Margolis call these “strategic gambles by Microsoft” that “paid off handsomely”. For a finance department, a spreadsheet that came with the word processor and presentation software its reports already used was the path of least resistance.

Lotus stumbled

Lotus, for its part, showed a “general disregard of the Macintosh market”, was “slow to create a version of 1-2-3 for Windows”, placed a “wrong bet on OS/2”, and launched Improv, which “introduced what has come to be known as the pivot table”, on the NeXT computer. It also sued rivals over the look and feel of its software: the incumbent knew its lock-in lay in users’ habits.

Even so, the shift was gradual. Network-effects theory predicts “tipping”, where one product pulls ahead and the whole market suddenly flips to it. Liebowitz and Margolis found instead “a fairly steady increase in Excel’s market share and a fairly steady decrease in 1-2-3’s.” Their verdict: “Lotus 1-2-3 failed because other competing products were judged to be superior in quality.” Fylstra adds that “it took years and took huge investments for Excel to overtake 1-2-3.” Finance standards change at platform shifts, and only for a tool that stays better across several releases.

Excel at 40: the timeline

Excel turned 40 on 30 September 2025. Microsoft marked it with “40 Days of Excel”, a countdown from 6 August, and a news feature in which a community manager spoke of uses “far beyond the typical finance or accounting scenarios”. That is Microsoft describing Microsoft, but forty years on, finance is still the typical case. The table puts this page’s dates and figures in one place.

What the evidence doesn’t show

Much of the early history rests on the inventors’ own accounts, several recorded decades later, and they don’t always agree. Sources date VisiCalc’s launch anywhere from May 1979 to the first shipment that October. The case for Mattessich rests on his own later summary and on historians quoted on his web page, not on the 1961 article, which isn’t openly available. A popular detail about LANPAR, that the telephone companies needed it because the 2,000 cells in their budgeting forms were so hard to change, can’t be traced to any primary document.

Some well-known numbers are weaker than they look. The market shares often quoted for the switch (Lotus from 70% in 1988 to 26% in 1997, Excel from 10% to 68%) are credited to Liebowitz and Margolis but don’t appear in the text of their chapter. A 2025 anniversary article in Public Accountant, an Australian accounting body’s magazine, cites a 1990 survey in which only 2% of AICPA members used Excel, and a user base of “more than 1 billion people”. Neither can be traced to its source. And no historian or economist we know of treats Excel’s file format as business’s exchange standard. The research puts the lock-in in people’s skills and macros.

None of this touches the main line, on which the sources agree: the spreadsheet was specified by people doing budgets and forecasts, and business, finance included, bought personal computers to run it. That is the historical half of the case in In defence of Excel. For why the grid suits the way we think, see Why the grid fits how we think. The next post in the series looks at where Excel breaks.