Latka logo
By Nathan LatkaBusiness Software4 min read

SaaS Pricing: Picking the Metric, Setting the Number, and Raising It Later

Every pricing page is three decisions wearing one design: what you charge for, how you package it, and what the number is. Here's how founders on the record made each call — including the ones who priced against salaries instead of software.

On this page
  1. Decision one: the value metric
  2. Decision two: packaging
  3. Decision three: the number
  4. The two most expensive pricing mistakes

SaaS pricing looks like one decision and is actually three:

  • The value metric — what unit you charge for.
  • The packaging — how features and units bundle into plans.
  • The number — what each plan costs.

Companies agonize over the third and inherit the first two by accident — usually by copying whoever they compare themselves to. The founders who’ve talked pricing on the record with GetLatka mostly got famous for decision one.

Decision one: the value metric

The value metric is the unit your invoice scales with — seats, usage, transactions, revenue share, or something invented. The test is whether the unit tracks the value the customer receives:

Per seat

Remains the default because it’s legible, but it taxes adoption — every new user is a procurement event.

Activity-based

Expensify’s twist: David Barrett charges only for seats that are actually active in a month — “we only get paid when you get active” — which turns the pricing model itself into a sales pitch: idle licenses cost nothing, so companies deploy wide without fear.

Take rate

Prices a marketplace’s flow: Magic Eden’s 2% of NFT volume scaled revenue perfectly with customer success — in both directions, as its history shows.

Per role

The AI era’s boldest metric yet: 1Mind prices its AI sales agents per role, benchmarked not against software budgets but salaries — “it’s the cost of a human, so it’s six figures,” per founder Amanda Kahlow, who is deliberately walking customers off per-conversation metering because predictability closes deals that meters scare away.

Kahlow’s reasoning generalizes: when your product replaces labor, price against the labor line, not the software line. The software budget is thin; the payroll budget is the company.

Decision two: packaging

Packaging segments willingness to pay without quoting anyone a custom price. Webflow’s 2018 structure is a compact masterclass — designer plans priced separately from hosting plans, two lines with opposite churn profiles:

Designer plans — the tool

Freelancers paid for their workflow. Churn ran 4–5% monthly.

Hosting plans — the outcome

Each client site generated its own recurring line. Churn was “essentially zero.”

Vlad Magdalin then added the layer most companies never think of: letting freelancers resell $20 hosting at $200 and taking a cut of the spread — packaging as a channel incentive. His honest summary of how they navigate it all: “Pricing is so hard… sensitivity surveys, figuring out which pricing levers are the most efficient — and then a lot of data, some gut feel, and a lot of experimentation.” The full Webflow breakdown is here.

The enterprise version of packaging is the services attach: Workboard’s paid coaching onboarding — kept out of the subscription line entirely — repaid acquisition cost on day one while making the software stickier. Packaging isn’t only plans on a grid; it’s every distinct thing a customer can pay you for.

Decision three: the number

Three disclosed heuristics beat any framework diagram:

  • Anchor to the alternative — 1Mind’s six figures reads cheap next to the “89 SDRs and 19 sales engineers” its HubSpot deployment displaced.
  • Raise with a reason — SafetyWing moved $35 to $45 over six years and framed it as “inflation adjustments” — a raise customers accept because the story is fair, not clever.
  • Check the payback, not the price — Outreach’s Manny Medina ran every segment against one gate — gross-margin-adjusted CAC payback under 20 months — letting price, deal length and acquisition cost trade off freely as long as the unit economics cleared.

A number is right when the machine around it clears its hurdle, not when it matches a competitor’s page.

The two most expensive pricing mistakes

Underpricing as strategy

Cheap feels like a growth hack and compounds into a ceiling: low price attracts low-intent customers, which drags churn up and rules out the sales motion that could reach better ones. (The ACV-to-go-to-market fit table shows the trap mechanically.) Expensify is the disciplined exception that proves the rule — Barrett keeps prices low on purpose, because his entire model is bottom-up scale: “much more focused on massive scale than trying to squeeze harder.” Cheap works when cheap is the strategy, not the accident.

Never revisiting

Pricing set at launch quietly ages while the product triples in value. The companies above all re-priced deliberately — new metrics, new packages, new numbers — usually annually. A pricing page that hasn’t changed in three years is a discount that compounds.

See what thousands of SaaS companies actually charge — and earn — at getlatka.com/saas-companies.

SourcesFounders’ on-the-record pricing conversations with GetLatka — Expensify’s David Barrett, 1Mind’s Amanda Kahlow, Webflow’s Vlad Magdalin, Outreach’s Manny Medina — and the GetLatka company breakdowns linked throughout.

Get the real numbers behind SaaS

CEO-confirmed revenue, growth, and valuation data for thousands of private SaaS companies.

Create Your Free Account →