The History of Project Management Software

By Victor Hoang · Updated Oct 24, 2026

Picture a DuPont engineering department in late 1956. Design, procurement and construction each kept their own schedule. A survey started that autumn. It asked whether a computer could prepare a master schedule and revise it as conditions changed. Until then, a plan lived on paper. A late job meant redrawing it by hand.

Project management software began in 1957, when DuPont and Remington Rand tested critical path scheduling on a UNIVAC I computer in Newark, Delaware. For about 25 years it stayed with specialist schedulers on mainframes and minicomputers. Then personal computers (PCs) put it on the desktop in the 1980s. The web turned it into a shared team tool in the 2000s. In 2024 vendors began shipping AI agents inside it. The through-line is who gets to see the plan: first a few experts, then every scheduler, then everyone on the team.

Timeline at a Glance

The Paper Planners (1896-1956)

Plans lived on charts that people redrew by hand

Picture a drawing office around 1900, with a long sheet pinned to a board and a draftsman adjusting strips of paper when a delivery slipped. The Polish engineer Karol Adamiecki worked this way. Patrick Weaver's biography of Henry Gantt dates Adamiecki's harmonogram to 1896 and describes a date scale down the left side with the activities listed across the top. Our piece on Adamiecki and the harmonogram covers why it stayed little known outside Poland.

Henry Gantt's charts were a different tool. Weaver notes they tended to be retrospective and diagnostic, filled in after a day's work to show where production fell short, and that Gantt's first book, Work, Wages and Profits, appeared in 1916. The charts that carry his name in software today are closer to the forward-looking bar chart. The history of the Gantt chart tells that split in full.

These are ancestors, not software. They share one limit: the chart records a plan but can't recalculate it. Change one date and a person redraws everything downstream. That limit set the problem that the first programs went after.

What this means for teams today: a plan you can't recompute quickly is a drawing, not a model. Software earned its place by making change cheap, and that's still the test.

The Mainframe Age (1957-1970s)

DuPont's UNIVAC trial turns a schedule into a program

A UNIVAC I in Newark, Delaware, ran duPont's early scheduling jobs. In their 1959 paper, Kelley and Walker record that Remington Rand and duPont began a cooperative effort in May 1957, with Remington Rand supplying the programs. The critical path method was programmed for the UNIVAC I, 1103A and 1105, and the paper lists the size limits: 739 jobs on the UNIVAC I, 1,023 on the 1103A and 3,000 on the 1105. The full story is in the history of the critical path method, and the Navy's parallel effort is in the history of PERT.

Software became a product when a vendor took it over. Weaver reports that the contractor H.B. Zachry Company and IBM signed a joint venture in March 1963, and that in early 1964 IBM announced its Project Control System for the 1440 computer. The system used the name "precedence diagramming". Weaver adds that IBM's hold on the mainframe market made that notation the dominant one, and that by 2004 it was virtually the only commercially available system in general use.

In the 1970s, minicomputer systems such as MAPPS on Wang hardware and Artemis on HP and DEC hardware cost less to run. The same source says they brought the rapid demise of mainframe scheduling, and that by the mid 1980s the mainframe era was over. What stayed constant was the user. These systems still needed skilled schedulers, so only a few people in a company ever touched them.

What this means for teams today: the first scheduling tools assumed a specialist operator. If your current tool only works when one expert maintains it, you've rebuilt the 1970s.

The Desktop Years (1980-2001)

Scheduling escapes the specialist and lands on every PC

In 1980, a small British firm shipped a scheduler for the Apple II. Weaver says Micro Planning Services released Micro Planner v1.0 that year after 14 months of development, based on the ICL PERT mainframe system, and calls it the first commercial scheduling software for a micro computer. "The Planner" followed for the IBM PC in 1982. Primavera was founded in May 1983 by Dick Faris, Joel Koppelman and Les Seskin, and according to Weaver it shot to prominence with a system able to handle 10,000 activities in the late 1980s.

Microsoft's entry was quieter. The Microsoft Project Users Group dates version 1 to 1984. It says Microsoft bought all rights in 1985 and released later versions for DOS, and it describes MSP 3.0 in 1990 as a significant upgrade that ran on Windows 3.0. The group's list continues with Windows 95 and server products later in the 1990s and 2000s.

Cheap tools changed who made schedules. Weaver writes that by the early 1990s nobody was scheduling manually and that the number of people creating schedules part-time and untrained exploded. He also notes that the professional schedulers who remained were mostly consultants, particularly claims consultants. Spreading the tool had not spread the skill, which became the era's new problem. Many of those schedules drew a bar chart on a PC without checking the logic behind it, a gap our Gantt chart and critical path method guides try to close.

What this means for teams today: a PC tool doesn't make anyone a scheduler. Check that someone on the team can read the dependency logic, not only the bars.

The Web Turn (2002-2011)

Browsers move the plan from one desk to the whole team

Two small companies show the change. Atlassian's 2015 registration filing says the company was founded in 2002 and that "We launched JIRA in 2002". It describes a customer, Cochlear, that adopted Jira in 2003 as a bug tracking system. Jira began as a tool for software teams to track issues, not to draw schedules, and it grew into a general project tool as other teams adopted it.

The other company came from design. According to 37signals' own account, it was a web design firm whose client work got hard to keep organized, so it built its own project system and clients asked to use it. Its anniversary post states, "We officially launched Basecamp on February 4, 2004." Five years later the same post says over 3,000,000 people held accounts. Basecamp's job was keeping client work organized, not calculating critical paths.

Trello added a third idea. On September 13, 2011, Joel Spolsky announced it on his blog as Fog Creek's newest product, calling it "an incredibly simple, easy-to-understand way for teams to collaborate online". The card board descends from the signal cards in the history of Kanban, though Trello's own pitch was simplicity.

Between those launches, enterprise vendors consolidated. Oracle announced on October 8, 2008 that it had agreed to acquire Primavera, described in the release as a provider of project portfolio management solutions. Portfolio tools were turning scheduling into a question of which projects to fund, the subject of project portfolio management.

What this means for teams today: the web changed the user from scheduler to participant. Pick tools your whole team will open without training, then add scheduling depth only where the work needs it. The working methods many of those teams ran inside these tools have their own histories, told in the history of Scrum and the history of lean project management.

The Work Management Boom (2012-2023)

Boards, timelines and automation merge into one category

By 2012 the pieces were on the table, and new vendors packaged them as work management. Asana's S-1 says the company was incorporated in Delaware on December 16, 2008 and counts its milestones from the launch of the Premium level in 2012. It lists tasks, lists, calendars, boards, timelines and workload as the ways users can view work, and notes that its Rules automation launched in October 2019. monday.com's F-1 filing says the company was incorporated as DaPulse Labs Ltd. in Israel in 2012 and changed its name in December 2017.

Consolidation continued. On January 9, 2017, Atlassian announced an agreement to acquire Trello, which it said had over 19 million users. The deal joined a card board with an issue tracker inside one vendor, and it's a good marker for the era: the Gantt view, the board and the backlog stopped being separate products and became tabs of the same one. The scheduling methods in the history of project management now sit behind a single interface.

The new problem was sprawl. A team could keep tasks in one tool, issues in another and status in a third, each with its own board that looked tidy. Visibility had arrived, and control still depended on someone acting on what the boards showed.

What this means for teams today: a unified tool can still hide a muddled process. Write down who owns what and what "done" means before you configure statuses.

The Agent Age (2024-present)

Vendors add assistants that draft, search and act

Today's planner opens a project and finds an assistant offering to summarize it. Atlassian first showed Rovo in May 2024, and its own Team '24 Europe post describes it as generally available for all customers, with search, chat and more than 20 built-in agents. On October 22, 2024, Asana's announcement of AI Studio described a no-code builder in which teams add AI agents to workflows, with early access starting that day for its top two tiers.

These features push the old pattern one step further. CPM recomputed a schedule when a plan changed in 1957, and today's agents read the plan and draft the update. The longer arc of automation in this field, from UNIVAC runs to generative tools, is in the history of AI in project management. Because the agents read whatever is in the workspace, their output is only as good as the tasks, owners and dates someone entered.

What this means for teams today: treat clean data as the entry fee for agents. A vague task list produces a confident summary that's wrong.

Key People

Key Facts

• Kelley and Walker's paper records that Remington Rand and duPont began working together in May 1957, with programs for a UNIVAC I.
• Primavera was founded in May 1983 by Dick Faris, Joel Koppelman and Les Seskin, according to Patrick Weaver's history of scheduling.
• Atlassian's filing says it launched Jira in 2002 and counted more than 37,000 Jira customers by September 30, 2015.
• 37signals launched Basecamp on February 4, 2004, and reported over 3,000,000 accounts by February 2009.
• On January 9, 2017, Atlassian said Trello had over 19 million users when it announced the acquisition.

The Pattern Behind the History

This is our synthesis, not published research. Call it the Widening Circle. Five points repeat:

• Each era widened who could see the plan. Specialists on mainframes, then desktop schedulers, then whole teams in a browser.
• A cheaper machine moved the tool down the org chart. Minicomputers ended mainframe scheduling, and PCs ended manual scheduling.
• Access outran skill. Weaver's account of untrained part-time schedulers has a modern echo in boards nobody maintains.
• New entrants solved a different problem. Jira tracked issues and Basecamp organized clients and conversations, while scheduling vendors tracked time.
• Consolidation follows each boom. IBM's 1964 product fixed a standard notation, and Oracle's 2008 agreement for Primavera and Atlassian's 2017 agreement for Trello pulled separate tools under one vendor.

What It Teaches Teams Today

• Choose the problem before the tool. Schedules, issues, conversations and cards were built for different questions. Name yours first.
• Keep a reader for the logic. Someone should understand dependencies well enough to check a tool's output, whether the tool is a 1964 mainframe or an agent.
• Expect consolidation. Features migrate into whichever tool already holds your data, so avoid scattering your plan across many.
• Clean the plan before automating it. Agents act on what's written down.