There will be a million
B2B SaaS businesses.

The ones that win will have taste. Most landing pages don't.


from the editor's desk

The pages all say the same thing.

If you read forty B2B SaaS landing pages in a row, a strange thing happens. The pages start to merge. They use the same six adjectives — fast, easy, powerful, seamless, modern, intelligent. They name the same five integrations. They have the same three-column "Why us" section with the same three icons. After page ten, you cannot tell whose product is whose.

The pages aren't bad, exactly. They are unwilling to commit to anything. They describe the category instead of the product. They name "modern teams" instead of senior PMs at Series B startups. They promise to "transform workflows" instead of saving twenty minutes on the Tuesday playbook review. The writing is generic because being specific feels risky — and being specific is the thing that would have helped the visitor decide.

a typical hero, marked up The Modern Platform for Workflow Automation Run your customer-success playbooks without copy-pasting between Slack and Notion.

This was always a problem. It is about to be the only problem. Three observability launches on Show HN this week all said variations of "observability that fixes your bugs." None said whose bugs, which kind, for which engineer. Category copy on every one. Anyone can ship a SaaS product in a weekend now. The product itself is no longer the moat. The page selling it — the writing, the structure, the restraint, the willingness to mean one thing instead of six — is what will distinguish a business worth paying for from a business that fades into the category.

a typical bullet, marked up Powerful automation workflows 12-step playbooks that run while you're in a meeting.

That distinguishing thing has an old name. It is taste. You can read it on Linear's homepage, where the hero says "Linear is a purpose-built tool for planning and building products" — a sentence that names itself, names what it does, and refuses to promise an outcome the visitor would have to be skeptical of. Taste on a landing page means committing to one buyer instead of "modern teams." It means choosing four claims that are true instead of twelve that sound impressive. It means refusing to add a logo bar of Series-A startups nobody recognizes. It is, mostly, the discipline of removing things until the page says one thing well.

a typical CTA, marked up Get Started Free Send me your URL — I'll have a Snapshot back to you Wednesday.

Stet does that work. Founders send us their URL. We read the page closely, grade it on the same ten dimensions every time, and send back a written diagnosis — and, if you want it, the new copy already written. We don't take calls. We don't make decks. We send writing that reads like someone meant it. The framework is the same on every page; the eyes are careful; the turnaround is days, not months.

field notes

May 21 — read 12 Show HN launches today. Eight used "platform" in the hero. Two of those used it correctly.

— Stet · May 2026


in practice

Before, after.

Three heroes we read this month. The original on top, the version we’d write underneath. The fix is rarely a punchier sentence — it’s a sentence that pictures the next moment.

Modern billing for AI companies.

Bill your customers for tokens without hiring a CFO.

Modern did no work. AI companies was three buyers in one phrase. The rewrite names the founder who’s wiring Stripe at 11pm.

Observability that fixes your bugs.

Stop getting paged at 2am for things already in the logs.

Fixes your bugs was outcome theater. The rewrite names the engineer’s actual Tuesday and the specific pain they would pay to never feel again.

Simplest Messaging Platform.

Send messages between postboxes. Four lines of Python.

Simplest Messaging Platform was a claim. The rewrite is a demonstration. Buyers remember demonstrations.

All three originals are real. The afters are what we’d ship if these were our clients. The pattern across all three: the rewrite isn’t punchier — it pictures the next moment.


field notes

What we read this week.

A working journal of observations from public B2B SaaS launches. Long pieces Mondays, short notes through the week.

May 25 — a Show HN B2B API launch

A Show HN launch today did something most developer tools never manage: it kept its metaphor.

The product is CloudPostOffice. The unit in the SDK is a postbox. The verbs are send, listen, publish, subscribe — all the verbs you’d actually use at a post office. The code samples, product structure, and naming system are all speaking the same language. Most developer tools never get this far. They launch with one metaphor, describe themselves with another, then explain the features in generic infrastructure dialect. CloudPostOffice mostly resists that drift.

Then the hero loses its nerve: “Simplest Messaging Platform.” A platform. After all that careful vocabulary work, the page suddenly falls back into the most abstract noun in B2B SaaS — a word that explains nothing and remembers nothing. The post office metaphor had already done the harder, better job: it gave the product shape.

The mistake isn’t using a cliché. The mistake is overwriting a concrete mental model with category language the moment the company gets closest to saying what it actually is.

When a metaphor is doing real work, protect it aggressively. The moment you translate it back into generic SaaS terminology, you flatten the thing people were finally starting to remember.

— Stet

May 23 — read four B2B SaaS pricing pages today. Three buried the price below the fold. Two of those three used the phrase “flexible pricing for teams of all sizes” — the universal flag for we’ll tell you in a call. The fourth showed the number. That one will sell more.

May 21 — eight YC observability launches on HN last week. Each described what the tool does. None described what the engineer does differently after installing it. The buyer reads twelve features. They remember zero.


how to work with us

Three formats. Same framework.

Same diagnostic rubric on every page. The difference between the formats is how much of the rewriting you want us to do.

i.Snapshot

$147 · 48-hour delivery · one-time

A written diagnostic. The page graded across our ten dimensions, the top five fixes ranked, and the direction for each one named. You implement the changes. A small sample of what the studio sounds like — most founders go straight to the Rewrite.

see a sample → begin the Snapshot
the studio’s flagship

ii.Rewrite

$997 · 5-day delivery · one-time

The Snapshot, then the work. Every section of your page rewritten, ready to paste in. Three alternate headlines with reasoning. A sketch of the recommended section order. One revision pass. The version of your page that commits to one buyer instead of all of them.

The studio is new. The first five Rewrites go out at $997; after that, $1,500. The work is the same — the number only reflects the line forming behind it.

begin the Rewrite

iii.Standing

$1,497 per month · three-month minimum

For founders who want the page to keep improving as the product evolves. A relationship, not a deliverable. Monthly rewrites, conversion-data review, unlimited written Q&A. Cancel anytime after the third month.

ask about Standing

Send your URL. We'll read the page.

say hello