Spreadsheet History

Pre-1960s

Ancient Accounting

Ledgers, clay tablets, and the story we tell ourselves about the grid

You might picture ancient accountants hunched over enormous sheets of paper: rows and columns, pencil moving, numbers changing, erasing, starting again. Grinding work. Excel in parchment, or so the story goes.

That picture is wrong.

Serious accounting did not work that way. It ran on ledgers: bound books of transactions recorded in order. A sale goes in. A payment goes in. A correction does not overwrite the original line. You post an amendment, clearly labeled, that explains what changed and why. The history stays visible. Debits and credits balance. The past is not edited away.

Before computers, before the spreadsheet

Humans still organized numbers in rows and columns for thousands of years. Mesopotamian clay tablets. Roman abacuses. Medieval merchants' daybooks. Venetian double-entry bookkeeping from the 1400s. Wide ruled paper in the 1800s. The grid is old. The habit of rewriting the grid is not.

Ledgers answered a different question: "What happened, in order, and can we prove it?" Forecasting and scenario planning came later, on different paper, for different jobs.

The software rewrote the past

VisiCalc made the editable grid feel inevitable. Dan Bricklin said the killer use case was forecasting: enter a number, watch dependents update, change your mind, try again. That is scenario planning, not bookkeeping. We read today's spreadsheet back into yesterday's accountants because the software won. Immutability is the opposite of what spreadsheets became.^[2][4]

Earliest mentions of the word spreadsheet

For most of the 20th century, people wrote spread sheet as two words: a physical sheet of paper, often wide, laid out in rows and columns. Printers used spread for facing pages in a magazine. A spread sheet was paper wide enough to hold a full layout at once. Accountants borrowed the phrase for analysis tables: costs spread across columns, totals visible in one glance.

Later the words merged. Spreadsheet (one word) shows up in professional dictionaries by the 1950s as settled vocabulary. By the 1960s it picked up a second meaning: software that mimicked the paper form. Same lineage, different era. The hyphen and the space came first. The compound came after. VisiCalc did not invent the word. It inherited it.^[6][15]

Linear math, written down

Before calculators, clerks used comptometers and adding machines. Before those, columns of figures on paper with a stubby pencil. The grid let you lay out numbers and follow the math with your finger instead of holding it all in your head. That was layout and arithmetic, not live recalculation across linked cells.

What changed in the 20th century was not the concept of tabular data. It was automatic recalculation: change one cell, and every dependent cell updates. That required computers. In the 1960s, those cost more than a house.

1960s

Early Signs of Life

Large machines, hundreds of thousands of dollars, and the first computerized grids

In the 1960s, computers filled rooms, and the people who programmed them rarely touched the machines. Fortran analysts fed punch cards to mainframes at universities, defense contractors, and phone companies. Most had never seen a mouse; Douglas Engelbart's demo was still a few years away. When Bell Canada needed a budget form changed, those analysts were who managers called first.

Professor Richard Mattessich at the University of British Columbia pioneered computerized spreadsheets for business accounting between 1961 and 1964, culminating in Simulation of the Firm Through a Budget Computer Program. He was not building a toy. He was simulating an entire firm's budget on a mainframe.^[5][12]

Richard Mattessich

UBC · 1961-1964

Pioneered computerized budget simulation on mainframes. His work predates VisiCalc by nearly two decades and proves the idea was academically alive long before the personal computer existed.

René Pardo

LANPAR Technologies · 1969

Met Landau at McGill in the mid-1960s. Co-founded LANPAR Technologies and co-invented LANguage for Programming Arrays at Random. Pitched eighteen U.S. telephone-company vice presidents in 1971. Later pursued a Master's of Education at Harvard and helped build early educational multimedia.

Remy Landau

LANPAR Technologies · 1969

Co-invented LANPAR with Pardo after a dog-walk conversation that became a programming language. Recognized the forward-referencing breakthrough on the spot. Won U.S. Patent 4,398,249 with Pardo in 1983 after twelve years of litigation.

In the summer of 1969, AT&T changed roughly 2,000 cells in the budgeting forms Bell Canada depended on. When the plant manager asked how long reprogramming would take, internal MIS groups quoted two years. Pardo and Landau had another idea: let the manager type the form the way he would write it on paper, separating cells with semicolons, and let the computer resolve the dependencies.^[14][21]

Landau looked at him and said, "You know what you just said? You just invented a computer language." They formed LANPAR Technologies that year. The name was a mash-up of their surnames and an acronym: LANguage for Programming Arrays at Random. They rented a hotel room and wrote the first compiler over six weeks, often working from eight in the morning until two at night.^[21]

Their breakthrough was forward referencing: formulas could point at cells not yet evaluated, so users did not have to think about execution order. Drop your punch cards, shuffle them, rerun the job. The grid still recalculated correctly. Today we call that reactive programming. In 1983 it became a patent.^[14][21]

LANPAR was the grid language; OUTCOM was the upper command layer for generating reports. Cells could branch on conditions and reach external databases. The team licensed the software rather than selling source outright: they copyrighted the code and kept ownership. In 1971, Pardo pitched eighteen vice presidents of every U.S. telephone company. AT&T's rule was unanimous consent. They got it. The deal licensed what the inventors' families later called the world's first spreadsheet, a decade before VisiCalc.^[14][21]

LANPAR ran on mainframes. It was not interactive in the VisiCalc sense: you submitted jobs, waited, read printouts. But it was a programming language for two-dimensional arrays with formulas, the conceptual ancestor of every cell reference today.^[14]

The 1960s also set an expectation that stuck: computers should deliver answers. Management wanted one tool for everything. Spreadsheets became that tool mostly because they were flexible, not because they were perfect for every job.

1960s

The Market Commercializes

Go to the moon, price the impossible, forecast what has never been sold

The business problem of the 1960s was not bookkeeping. Ledgers worked fine. The hard part was forecasting. How do you price a product that has never existed? How do you budget a lunar program? How do you model a market that has no precedent?

Defense contractors, aerospace firms, and telcos needed scenario tools. Not immutable accounting records, but living models where assumptions could be swapped and outcomes compared. LANPAR at Bell Canada and GM was exactly this: budgeting at scale, across divisions, with arrays that could be reprogrammed. When AT&T rewrote 2,000 budget cells, Bell Canada's MIS group quoted two years. With LANPAR, the plant manager could do it himself over a weekend, without waiting on Fortran programmers.^[14][21]

Dan Bricklin would later sit in a Harvard Business School classroom and hit the same wall: case studies required rebuilding entire financial models by hand every time a single assumption changed. Businesses already lived on forecasts. The tools were just slow.

1970s

Spreadsheets Get Personal

The Apple II, a boring finance class, and the birth of the modern grid

In the spring of 1978, Dan Bricklin, a Harvard MBA student and MIT electrical engineering graduate, had a daydream: what if his calculator had a ball on its back, like a mouse? What if he could see a virtual spreadsheet hanging in the air like a fighter pilot's heads-up display, circle numbers, and answer "10% will be fine"? Not bookkeeping. Forecasting, with the immediacy of a cockpit HUD.^[2]

He prototyped a five-column, twenty-row matrix in Apple Integer BASIC over a weekend on a borrowed Apple II. It could not scroll. It could not do much arithmetic. But you could see the grid form as you typed. The metaphor was "an electronic blackboard and electronic chalk in a classroom."

Before production code, Bricklin mapped the whole program as a state diagram: what each keystroke did, how replicate worked, where help appeared. In November 1978 he sketched it on the back of Harvard Business School spreadsheet paper. The blue grid lines show through in the scan. Paper became prototype. The software inherited its grid because the design was literally drawn on one.^[22]

Dan Bricklin

Co-creator · Software Arts

Conceived the visible calculator. Wrote the first prototype. Chose the name VisiCalc (visible calculator) over Calcu-ledger and Calcu-paper. Called Bob Frankston the "co-creator" who made it real.

Bob Frankston

Co-creator · Software Arts

Rewrote VisiCalc in 6502 assembler on a Multics timesharing system. Packed the entire program into roughly 20KB of memory. Added scrolling, faster arithmetic, and the production polish that made a demo into a product. In a 1984 BYTE interview, he recalled the naming hunt: "electronic ledger, electronic blackboard, visible calculator. That's what we finally based the name, VisiCalc, on."^[1]

Dan Fylstra

Publisher · Personal Software

Byte Magazine associate editor who insisted VisiCalc run on the Apple II. Founded Personal Software (later VisiCorp). Ran the teaser ad in Byte that launched an industry.

Bricklin and Frankston formed Software Arts Corporation on January 2, 1979. In May, Fylstra's Personal Software began marketing VisiCalc. It shipped in November 1979 for under $100. Within years, roughly one million copies sold. Dealers bundled Apple II machines with VisiCalc because businesses bought the computer to run the spreadsheet.^[14][20]

The command tree on paper became a print ad. Fylstra had to teach the grid to buyers who had never seen one.

Steve Jobs

Apple · 1970s-80s

Credited VisiCalc as the event that propelled the Apple II into offices. Later argued the same pattern held for Lotus 1-2-3 and the IBM PC: killer spreadsheets sold the hardware.

The Apple II was not the first personal computer. But Jobs credited VisiCalc as the reason it succeeded in offices. If VisiCalc had shipped on another machine first, he said later, "you'd be interviewing somebody else right now."^[3] By 1985, the Wall Street Journal was running editorials that assumed readers already knew what VisiCalc was. Accountants divided their careers into before and after the computerized spreadsheet. A twenty-hour weekly bookkeeping task could collapse to minutes.^[14]

Bricklin and Frankston squeezed VisiCalc into 20KB of 6502 assembler. I rebuilt it in Google Sheets: green-on-black cells, slash commands, live recalc. No Apple II required.

That letter-and-number split answers a question most users never think to ask. Column A, row 1 is A1. If both axes used numbers, 11 could mean row eleven or row one, column one. On a tiny screen, letters disambiguate at a glance. Microsoft's Multiplan later bet on R1C1 instead; A1 won because people read it faster.^[14]

1980s

Modern Spreadsheets Rise

Graphical interfaces arrive. Lotus leads. Microsoft follows with a mouse.

While VisiCalc's creators fought legal battles with VisiCorp (beginning September 1983), the IBM PC arrived and the spreadsheet market grew fast. Mitch Kapor, former VisiCalc product manager at Personal Software, had already sold VisiPlot and VisiTrend back to VisiCorp for $1 million. With Jonathan Sachs and venture capitalist Ben Rosen, he founded Lotus Development Corporation in 1982.^[10][18]

Kapor walked an early Lotus prototype into VisiCorp, the company he'd just left, and offered to sell it back. Executives looked at what Sachs had built and said the functionality was "too limited." Kapor kept it. When 1-2-3 launched in 1983, he and Sachs owned the new standard almost overnight.

Mitch Kapor

Lotus Development · Founder

Former VisiCalc product manager. Built 1-2-3 to feel familiar to VisiCalc users while exploiting the IBM PC's memory and speed. Lotus revenue hit $53M in year one, $156M in year two.

Jonathan Sachs

Lotus Development · Co-founder

Co-architect of Lotus 1-2-3. Built it with Kapor for the IBM PC.

Ben Rosen

Sevin Rosen Funds · Lotus investor

Venture capitalist who financed Lotus Development in 1982 and sat on its board. Also backed Compaq, Electronic Arts, and Silicon Graphics through Sevin Rosen Funds.

Bill Gates & Excel team

Microsoft · 1985

Excel 1.0 for Macintosh (1985) brought pull-down menus and mouse-driven GUIs to spreadsheets. Many bought Macs specifically for Excel. Excel 2.0 for Windows arrived October 31, 1987.

Doug Klunder

Microsoft · Excel 1.0

Lead developer on Multiplan and Excel 1.0. Invented "intelligent recalc" so only dependent cells updated, a speed edge over Lotus 1-2-3. The team's motto: "Recalc or Die."

Richard Frank

Sorcim · SuperCalc

Founded Sorcim in 1980 and shipped SuperCalc for the Osborne 1 portable. Bundled with WordStar, it gave CP/M users a VisiCalc-class grid before the IBM PC dominated.

Kapor and Sachs packed three tools into one product: spreadsheet, graphics, database. They named it 1-2-3. Sachs engineered named cells, cell ranges, and macros so VisiCalc veterans could level up without relearning the grid. Kapor's pitch was familiarity with a faster machine underneath.

Steve Jobs made the same argument about platform shifts: VisiCalc sold Apple IIs; Lotus 1-2-3 sold IBM PCs. Two spreadsheets, two industry explosions.^[13]

Lotus also made a costly choice: it largely ignored the Macintosh market. The revenue looked small, but building a graphical spreadsheet on the leading GUI taught Microsoft useful lessons with Excel for Mac while Lotus stayed focused on the IBM PC.

In 1985, Lotus acquired Software Arts and discontinued VisiCalc. A Lotus spokesperson told reporters that "1-2-3 and Symphony are much better products so VisiCalc is no longer necessary" (per D.J. Power, A Brief History of Spreadsheets).^[14]

1990s

The Industry Emerges

Pivot tables, VBA, lawsuits, and the tipping point toward Excel

Lotus engineers shipped Improv in 1990 with pivot tables years before Excel had them, then asked mainstream offices to abandon the familiar grid and buy a NeXT instead of a Mac or PC. Bold bet. Wrong audience.

By 1993, journalists were writing Lotus obituaries. Most people who switched from Lotus to Excel did so between 1993 and 1995. Excel shipped a Lotus macro interpreter so existing 1-2-3 files could keep running during the migration.^[14][17]

Excel 5.0 (December 1993) replaced the Excel macro language with Visual Basic for Applications. Corporate America had been building Lotus macros for a decade. VBA turned spreadsheets into an application platform, and they have stayed that way since.

Jim Manzi

Lotus CEO · 1986

Took over as Mitch Kapor resigned. Led lawsuits against Paperback Software (VP Planner) and Mosaic (TWIN) over Lotus 1-2-3's "look and feel." Lotus won those cases but kept losing users to Excel.

Borland / Quattro Pro

Philippe Kahn · 1990s

Quattro Pro challenged Lotus and Excel with fast recalculation and a strong interface. Borland founder Philippe Kahn led one of several contenders that eventually lost to Excel.

James Stephenson

VP-Planner · Paperback Software

Developed VP-Planner for Adam Osborne's Paperback Software, a Lotus 1-2-3-compatible spreadsheet sold at paperback prices. Lotus sued over look-and-feel; Paperback folded after the case dragged on.

The legal battles mattered. In June 1990, Judge Keeton upheld Lotus 1-2-3's copyright on its menu structure. The decision shaped software UI law for years. Lotus won in court but kept losing users to Microsoft as Windows 3.0 (1990) and Windows 3.1 pushed Excel to the front.

2000s

Mass Market

Office ubiquity, OpenOffice, and spreadsheets everywhere

The 2000s did not invent new spreadsheet ideas so much as lock in the ones that already won. Microsoft Office ran most offices. Excel was the default. Every MBA learned it. IT departments tried to replace it and usually lost.

OpenOffice.org (and its descendants LibreOffice Calc and Apache OpenOffice) brought a free alternative. Governments and schools adopted it. The file format wars (.xls vs .ods) mattered to archivists more than to most users. Excel had already won the mindshare.

In 2006, Google acquired Writely from Sam Schillace's team at Upstartle, engineers who had already put a word processor in the browser while most vendors assumed spreadsheets would stay on the desktop. That acquisition became Google Docs and, eventually, Google Sheets. The browser was not yet a credible spreadsheet engine. Schillace and his colleagues spent the next decade proving it could be.

This era also produced famous spreadsheet errors. The European Spreadsheet Risks Interest Group (EuSpRIG) kept a running list. TransAlta lost $24 million to a cut-and-paste mistake. Fidelity miscalculated a $2.6 billion dividend.^[8]

2010s

The Cloud

Google Sheets, Airtable, and the return of the database

Google Sheets (rebranded from Google Spreadsheets by 2012) showed that a browser could run a serious grid. Real-time collaboration, shared cursors, comments, and version history gave it an edge Excel did not match at the time.

Airtable (founded 2012) combined spreadsheet familiarity with relational links, views, kanban boards, and API access. Rows and columns were no longer enough for many teams.

The decade also brought Rows, Coda, Notion databases, Smartsheet, and dozens of niche players. Each took a slice: automation, docs plus data, project management, no-code apps.

Turn the grid into an app—without code

By the late 2010s, builders stopped asking users to stare at rows. They imagined a spreadsheet as backend: data in cells, experience in an interface. Glide turned Google Sheets into mobile apps. Softr connected Airtable to websites. Google bought AppSheet and made it official. The grid was no longer the destination. It was the database behind the screen.

2020s

AI in the Grid

APIs, BigQuery, machine learning, and built-in suggestions

The 2020s added intelligence. Google connected Sheets to BigQuery. Explore surfaces charts without a prompt. =AI() and Copilot turn natural language into formulas. Bricklin's heads-up-display idea is closer than it was in 1978. A wrong AI suggestion in cell F47 is the same kind of problem as a cut-and-paste error.^[2][4]

Better Sheets lives in this era, teaching Apps Script, automation, and workflows that turn a shared grid into something your team can run. Browse courses →

2030s

40% of the World

Two billion users is not saturation. What comes after the grid?

Some estimates put spreadsheet users at roughly 40% of all internet users, more than two billion people. Excel is a generic noun in many offices ("send me an excel"). Google Sheets is the default for startups and schools.

Will the next spreadsheet still look like a grid? Bricklin imagined a cockpit with heads-up displays and simulation. AR and VR might change how data is arranged. Or the grid disappears entirely and lives only as a backend for other tools.

Guinness World Records recognized Ariel Fischman for the largest collection of physical spreadsheet software: 506 boxed titles verified in Mexico City in 2018, spanning VisiCalc-era floppies to rare beta releases.

Ariel Fischman 414 Capital · Collector

The word catches up

The Google Books Ngram for spreadsheet barely moves for a century, then climbs fast after 1980. The two-word form spread sheets fades as the compound takes over.

Sheetscape

Who makes spreadsheet software today, and who used to

A map of the industry: the big three, the challengers, and the dead products along the bottom.

Hundreds of creators on YouTube and blogs teach spreadsheet skills: Better Sheets among them, plus channels on Excel formulas, Google Apps Script, financial modeling, and no-code automation.

Spreadsheet Opportunities

Where AI, interface design, and the grid still fall short

Spreadsheet software is about fifty years old and still growing. Microsoft and Google are bolting AI onto chat sidebars and =AI() cells. The grid itself has barely moved.

What AI still misses

Copilots draft VLOOKUPs and explain errors. They stumble on the real work: exploring a model, testing assumptions, trusting the result.

• Explainability in the cell. Copilot suggests a formula. You accept it. Why that structure? What breaks if you change one input?
• Deterministic AI. =AI() can return different text on repeat runs. Forecasting and compliance need answers you can reproduce.
• Agents on ranges, not workbooks. Real work happens in a block of cells: clean this column, find Q3 outliers, model an 8% rent hike.
• Structured extraction. Pulling tables from PDFs and invoices into typed columns is still manual or brittle.
• Natural language that compiles. "Revenue by region, stacked bar, exclude returns" should produce a live view, not a screenshot.

Interface and experience gaps

VisiCalc put a grid on screen in 1979. Excel added ribbons in 2007. Google Sheets added real-time cursors. The core interaction model is largely the same: infinite rows, A1 references, a formula bar above the sheet.

• Dependency maps you can click. Precedents and dependents exist as menu items. A visual graph of what feeds cell F47, with one-click navigation, would help anyone debugging a model they did not build.
• Mobile that is not a compromise. Sheets on a phone is read-mostly. Editing formulas, selecting ranges, and reviewing wide tables on a small screen is still painful.
• Onboarding without a course. Millions use spreadsheets daily and never learn INDEX/MATCH or pivot tables. Inline help at the point of need would close the gap faster than another tutorial.
• Collaboration with intent. Shared cursors show where someone is, not what they mean to change. Lock regions. Propose edits on sensitive cells.
• Views beyond the rectangle. Kanban, calendar, map, and timeline views exist in Airtable and Notion. The classic grid is one lens. Switching views without losing the underlying model is still rare in mainstream spreadsheets.
• Error states people understand. #REF!, #N/A, and circular reference warnings assume you already speak spreadsheet. Plain-language explanations at the cell level would save hours of forum searching.

Problems neither AI nor UI has solved

Some gaps are older than the current AI wave:

• Auditability. Who changed cell F47, when, and can you prove it in court?
• Interoperability. =IMPORTXML still breaks when a source site changes its HTML.
• Scale without a data warehouse. Million-row analysis for people who will never write SQL.

Bricklin imagined a heads-up display: circle a number, ask a question, see the answer. Copilot and Explore are early steps. The win is making the sheet itself smart, not just the chat box beside it.