Four ways to price a seat, and the one nobody is defending

01The shape of the market

Three of the four competitors we looked at moved off per-seat pricing in the last eighteen months.¹ They did not all move to the same thing. Northvane went to consumption, billed monthly against a committed floor. Harbourly kept seats but made them nearly free and put the money on a per-workspace platform fee. Ordell went hybrid and, judging by how often their pricing page has changed, is still deciding.

The fourth, Cadence, still charges a flat per-seat rate and has not changed it since 2024. They are also the only one of the four with a published enterprise floor, which suggests the seat price is a list price that nobody actually pays.

Per-seat pricing survives where the buyer counts people. It dies where the buyer counts work — and in this category the buyer has started counting work. The pattern across all four pricing pages, stated plainly.

02What each one charges

Entry price is the cheapest paid tier. Confidence is mine, and reflects how much of the pricing is actually published rather than inferred from a sales conversation.

Prices are list, annual billing, as published on 22 August 2026.

03The gap, and why it is open

Nobody is selling to the team of three that has one heavy month a year. Northvane's free tier covers them right up until the month they need it, and then the bill is unbounded. Harbourly's $249 floor is more than that team spends on everything else combined. Ordell and Cadence want a seat count that does not describe how the team works.

This is not a segment anybody has decided to abandon. It is a segment that falls between two pricing philosophies, and the companies on either side each assume the other one has it.

Share of the segment served by at least one competitor at a price it would plausibly pay. The first row is the finding.

04What would have to be true

• That the spike is real and not seasonal noise. We have one year of our own usage data and it does show a spike, but one year is a pattern of one.
• That Northvane does not simply add a cap. A monthly ceiling on their free tier would close this gap in a week and cost them nothing. It is the obvious move, which is an argument for moving first rather than for not moving.
• That we can bill it without building metering. If serving this segment requires six weeks of billing infrastructure, the maths stops working. A flat cap with a hard stop needs far less than true metering.³

05What I would do next

Two weeks of work, in this order. Pull our own usage distribution and check whether spiky small teams exist in our data at the volume this argument assumes — if they do not, the rest is moot and we have saved a quarter. If they do, price a capped tier and put it in front of ten existing accounts before it is built.

I would not commission more competitor research. The pricing pages say what they say, and the open question is about our customers rather than theirs.