We read 25 project management homepages.They nearly all say the same thing.

We scored 25 work management homepages across 9 elements of message market fit, against a single shared axis, on the same day.

We found something we didn't expect. This category doesn't have a middle anymore.

Key findings

1. The category has split in two. 6 companies now describe themselves as AI-agent platforms. 6 describe themselves exactly the way they did in 2019. Only 1 sits between them.

2. Saying the category's words doesn't mean your page works. Jira uses the new agent language and scores 19 out of 45 overall — second-worst in the set. Basecamp refuses the language entirely and scores 32.

3. 20 of 22 homepages score 2 or lower on what happens after the sale. Not one page scores above 3. In a category built on multi-year contracts.

4. 9 companies are named by nobody. They all position against Asana, Monday and ClickUp. None of them gets named back.

5. "Your team" appears word-for-word on 18 of 23 homepages. That's 78% of the category opening with the same two words.

6. Three companies absorb all the attention. Asana, ClickUp and Monday are named by 17, 16 and 15 of the pages. Everyone else is a rounding error.

The category split in two

Every homepage in this category is being pulled along the same axis. At one end is the shared workspace: a place where humans plan and track work together, described in boards, views, schedules and collaboration. At the other is the human-agent OS, a layer where AI agents are teammates that plan, execute and report on work alongside people.

We scored all 22 in-category homepages on that axis, 0 to 1.

There's almost nothing in the middle.

And the one company that is in the middle — Teamwork — got there by accident. It sells professional services automation, so its agent language sits on top of a completely different job.

You can see the split in the headlines without any scoring at all.

The agent pole:

• Asana: The OS for human-agent teams
• Monday: People and agents working as one team
• Linear: The product development system for teams and agents
• Jira: Turn plans into agent-ready tasks
• Wrike: Deliver work, powered by humans and agents
• Nifty: Built for humans + AI agents

The workspace pole:

• Basecamp: The refreshingly straightforward project management system that's rock-solid and easy to use
• TeamGantt: Online Gantt chart software for schedule-driven teams
• Trello: Capture, organize, and tackle your to-dos from anywhere

Same job. Same buyer. Two completely different industries, if you go by the copy.

Saying it isn't doing it

This is the finding that should worry everyone at the top of that table.

Position measures whether your page codes for where the category is going. It says nothing about whether your page actually works.

So we scored all 9 elements: what a page communicates before a buyer has words for it, who it signals to, and what it does when the buyer is ready to act.

The two came apart badly.

Look at the top and bottom rows.

Jira says the category's sentence — orchestrate work across your team and agents — and scores 19 out of 45. Second-worst page in the set. Campfires: 1. Conversion: 1. Connection: 1.

The biggest player in this market adopted the new vocabulary and built almost nothing behind it.

Basecamp is the mirror image. It sits further from the agent pole than any company here. It scores 1 out of 5 on Codes. And its page still reads 32 out of 45 — fourth best overall, with a 5 on Community.

Basecamp isn't failing to keep up. It's refusing, deliberately, and executing the refusal better than most of the companies that moved.

Across the whole set, agent-pole companies average 29 out of 45 and workspace-pole companies average 24. But the ranges overlap almost completely: 19–36 versus 18–32.

Translation: adopting the category's language buys you very little. What you do underneath it is what separates these pages.

More on this: A new headline is not a new position

Nobody sells the second year

We score a page's post-sale promise as Connection: what a buyer can see about life as a customer, and whether anything on the page gives them something to repeat to a colleague.

Twenty of 22 homepages score 2 or lower. Not a single page scores above 3.

Which is strange, because of what this product is.

A work management platform is a workspace a team lives inside for years. Switching costs are the entire defence against churn. Renewal is the business model.

And not one homepage in this category makes an argument about year two. Every page here is built to win the evaluation, then stops at the signature.

The rest of the category, by element:

These companies know how to argue their way onto a shortlist. Consideration is their strongest element by a clear margin.

They just have nothing to say once they're on it.

More on this: Nobody sells the second year

Nine companies are named by nobody

Every read produces the set of competitors a page implicitly positions itself against. Do that across 24 pages and you get a map of who is paying attention to whom.

Three companies absorb the category's entire attention. After that it falls off a cliff.

9 companies in this category are named by nobody at all.

Basecamp · Hive · Nifty · Plaky · Podio · Shortcut · Taiga · TeamGantt · Teamwork

Every one of them positions against Asana, Monday or ClickUp on their own homepage. Not one gets named back. That's one-way attention, and it's probably the most useful number in this report.

Two of those nine have among the best pages in the set. Basecamp scores 32 out of 45 and TeamGantt scores 31. Both beat Asana, both beat ClickUp, and both beat Jira by a wide margin.

Being good is not the same as being in the comparison.

More on this: You name them. They don't name you.

The same two words, on 78% of pages

We also counted the actual words.

"Your team" appears verbatim on 18 of 23 homepages.

Not "team." Not a synonym. Those exact two words, on 78% of the pages a buyer would compare side by side.

Other phrases that show up across the category:

"Your team" isn't wrong. It's team software. Addressing the buyer as someone who has a team is close to forced.

That's the problem.

A phrase that everyone uses carries no information. A buyer who reads three of these pages reads "your team" three times and learns nothing that separates them.

"In one place" is the category's oldest claim, and it's still on 7 of 23 pages. Consolidation has been the pitch in this market since before half these companies existed.

More on this: Say something only you could say

The agent language arrived all at once

Six companies, on the same day, in the same 10-minute window, all reaching for the same idea:

• Asana: the OS for human-agent teams
• Linear: the product development system for teams and agents
• Monday: people and agents working as one team
• Wrike: powered by humans and agents
• Jira: turn plans into agent-ready tasks
• Smartsheet: AI that connects systems, runs projects, and helps teams act faster

Whatever "agents" was worth as a differentiator six months ago, it's now the category's shared vocabulary.

Which is how category language always works. Someone says the new thing. It works. Everyone copies it. Within two quarters it's furniture, and the company that said it first has to find something else.

The interesting question is what happens to the six companies at the other pole.

Right now they look like laggards. If the agent framing turns out to be a bubble, they'll look like the only ones who kept their nerve.

Nothing on any of these homepages tells you which it will be.

More on this: When the whole category moves at once

Two companies didn't belong here

We ruled two submitted companies outside the category rather than score them against an axis they're not on.

Coda now brands itself Superhuman Docs and compares itself to Google Docs, Airtable and Notion. That's a docs-and-database workspace, not project management.

Todoist sells the world's #1 to-do list app to 50+ million professionals who simplify work and life. Personal task manager, not team project management. It scores 4 on Context and 4 on Clarity — which is to say it's very clear about being something this category isn't.

And one company we changed our minds about. Teamwork calls itself AI-powered professional services automation. On an earlier run against a smaller set, that read as its own category and we ruled it out. Against the full 25 it reads as in: the core job on the page is running projects, resources and financials in one place, with the PSA framing sitting on top of it. That's also why it's the only company in the middle of the axis.

We're naming that because it matters. A category boundary drawn against 25 companies is a judgment about those 25, and it can move when the set does. All three of these calls are close, and we'd rather say so than pretend the line is obvious.

Method

The set. 25 work and project management homepages, chosen to cover the range from enterprise to open source. We didn't include communication tools, whiteboards, docs platforms or consumer task apps.

The capture. September 8, 2026. Every page rendered in a real browser, so client-rendered sites are read the way a visitor sees them rather than the way a crawler does. Navigation stripped by link density, not by class name.

One axis, derived once. This is the part that matters most.

If you read each homepage independently, every page nominates its own category and its own scale. We tried that first, on 18 of these companies, and got 18 different axes with zero overlap. One label showed up as one company's destination and another company's starting point.

So the axis here is derived from all pages at the same time, before any scoring, and then applied unchanged to every company. No company got to grade itself against a target it picked.

The scoring. 9 elements, 1 to 5 each, across three layers:

• Unconscious — Codes, Context, Clarity. What the page communicates before a buyer has words for it.
• Signal — Campfires, Community, Competition. Who it signals to and what it's against.
• Action — Consideration, Conversion, Connection. What it does when the buyer is ready.

Position on the axis is scored separately, against the fixed poles.

What this can't see. It reads a homepage. It doesn't read images, video, product, or pricing pages. It can't tell you whether a page converts — only what it says, who it's for, and where it sits against everyone selling the same thing.

Coverage. 25 submitted. 24 captured and scored. 22 in the category, 2 ruled out. Zero element failures across 216 reads.

Two disclosures on the captures themselves. Adobe Workfront failed to return a page on both attempts, so a significant enterprise player is missing from this volume and its absence is ours, not theirs. Notion's capture is marked partial: its homepage carries almost no readable text, so it's read from headings, hero copy, alt text and social proof rather than from body copy. Its scores are real, and they rest on a thinner capture than everyone else's.

Where does your homepage sit

Everything above is the category. The part that's about one company is different: your 9 element scores against your category's real axis, which competitor owns the language you're borrowing without realizing it, and the positions nobody in your market has claimed yet.

Start with the page itself.

Paste your URL and Yieldpage reads your homepage the way your buyer's mind does. No account, under 5 minutes.