Nick QuirkFor about fifteen years, B2B SaaS had one default answer to the pricing question: charge per seat, per month. It was easy to explain, easy to forecast, and it grew revenue automatically as customers grew. Then it started breaking. Today a meaningful share of B2B tools — project management, CRMs, helpdesks, analytics platforms — are testing or fully committed to flat-rate SaaS pricing, and the ripple effects reach everyone who buys software for a team.
If you run an agency, this matters more than it sounds. Pricing models shape how you use a tool. Per-seat pricing quietly teaches your team to share logins, exclude freelancers, and keep clients out of the workspace. Flat pricing removes that friction — and introduces different tradeoffs.
What flat-rate SaaS pricing actually means
Flat-rate pricing means one fixed fee for access, regardless of how many people log in. Basecamp's Pro Unlimited plan at $299/month is the most cited example in project management: unlimited users, unlimited projects, one invoice. Other vendors use softer versions — a flat tier with a generous user ceiling, or unlimited internal seats with paid external collaborators.
It's worth separating the three models you'll actually encounter when buying:
- Per-seat: $9–$25 per user per month. Asana, Monday, ClickUp, Notion and most of the category default here. Cost scales linearly with headcount.
- Flat-rate: one price, unlimited or near-unlimited users. Cost is fixed; cost per user drops as you grow.
- Usage or value-based: billed on contacts, records, API calls, AI credits, or active projects. Cost tracks consumption, not headcount.
Most 2025-era pricing pages are hybrids. A flat platform fee plus AI credits. A per-seat base plus unlimited free guests. The clean categories are collapsing, which is exactly why buyers get confused.
Why flat pricing gained ground
Three forces pushed the shift.
Seat counts stopped correlating with value. Per-seat pricing assumes more users equals more value extracted. That held when software replaced manual labour one desk at a time. It stopped holding when automation and AI meant a six-person agency could service the client load that used to need fifteen people. The vendor's revenue shrank while the customer's outcome improved — an unsustainable arrangement for the vendor.
Buyers started rationing access. Every per-seat tool creates an internal debate: does the intern need a licence? Do we add the freelance link builder for a six-week sprint? Does the client get a view-only seat or do we just email PDFs? These decisions are rational and they systematically reduce how deeply a tool gets embedded. Vendors noticed that seat-gated products had worse retention because fewer people touched them.
Procurement got tired of surprise bills. A 14-person agency on a $19/seat plan that hires four people mid-year sees the annual cost jump from roughly $3,200 to $4,100 with no new contract. Flat pricing makes budgeting boring, which finance teams love.
The math: where flat pricing wins and loses
Do this arithmetic before you fall for either model. Take a $299/month flat plan versus a $15/seat plan:
- 5 users: $75/mo per-seat vs $299 flat. Per-seat wins by $2,688/year.
- 15 users: $225 vs $299. Per-seat still wins, but the gap is under $900/year.
- 20 users: $300 vs $299. Break-even.
- 35 users: $525 vs $299. Flat wins by $2,712/year.
- 60 users: $900 vs $299. Flat wins by $7,212/year.
The break-even point is simply flat price ÷ per-seat price. Everything above that line favours flat. Below it, you're subsidising larger customers.
But raw headcount understates the agency case, because agencies have three user populations, not one:
- Full-time staff — predictable, always licensed.
- Contractors and freelancers — rotating, often 20–40% of the roster in a busy quarter.
- Clients and their stakeholders — potentially 3–8 people per account, across 10–30 accounts.
Run the third group through per-seat pricing and the numbers get absurd. Twenty clients × four stakeholders × $15 = $1,200/month just to let the people paying you see their own work. Most agencies respond by keeping clients out entirely, then rebuilding status updates by hand in email and slide decks. That's the real cost of per-seat pricing: not the invoice, but the parallel reporting workflow it forces you to maintain. We've written about the alternative in the client-centric project workspace guide.
What flat pricing does to vendors
From the supply side, flat pricing is a strategic bet with real costs.
It kills automatic expansion revenue. A per-seat SaaS company with 110% net revenue retention gets a third of that growth from customers simply hiring. Flat-rate vendors must earn expansion through tier upgrades or add-ons, which means the product has to keep getting better in visible ways.
It also inverts unit economics. Under per-seat, a customer with 80 users pays 16× more than one with 5 and costs maybe 3× more to serve. Under flat, the 80-user account pays the same and consumes far more storage, support, and compute. Flat-rate vendors tend to build simpler, more opinionated products partly because complexity is expensive when revenue is capped.
And it forces clearer positioning. If you can't grow revenue by counting heads, you grow it by winning a specific segment decisively. That's one reason flat pricing correlates so strongly with vertical SaaS — a theme we cover in vertical SaaS vs all-in-one tools.
The AI squeeze on flat pricing
Here's the tension nobody had in 2019: AI features carry real marginal cost. A per-seat plan at $12/user with unlimited AI summaries can go negative if one power user runs 500 generations a month. Flat plans are more exposed still.
The market's answer has been the credit hybrid: a flat or per-seat platform fee plus a monthly allowance of AI actions, with overage sold in blocks. It's honest but it reintroduces exactly the budget unpredictability that flat pricing was supposed to solve.
When you evaluate a tool in 2025, ask specifically: what happens when we exhaust the AI allowance? Does the feature stop, degrade, or auto-bill? An agency generating content
