Sharpening Usage-Based Pricing Instincts

Usage-based pricing has become the growth narrative of the decade in SaaS. Align price with value. Let customers start small and expand. Remove friction from the initial purchase. Every company that's referenced Snowflake or Twilio in a board presentation is operating on this narrative. A lot of them are about to make an expensive mistake.

The instinct that usage-based pricing is universally better than subscription pricing is wrong. It's not wrong because usage-based is bad. It's wrong because the conditions required for usage-based pricing to improve financial outcomes are specific, and most companies don't check whether those conditions apply before they switch.

Where Money Leaves

The financial downside of a poorly executed usage-based pricing transition shows up in three places.

First, revenue unpredictability. Subscription annual recurring revenue (ARR) is predictable. Usage ARR is not. A SaaS company that converts a $15M subscription base to usage-based pricing often sees 20% to 35% revenue variance quarter to quarter as customers adjust usage in response to budget cycles, seasonality, and procurement reviews. That variance makes forecasting harder, capital allocation less efficient, and board confidence lower. At a 7x revenue multiple, higher revenue variance alone reduces valuation by 0.5x to 1.5x.

Second, customer behavior inversion. In a subscription model, the company wants customers to use the product as much as possible because usage drives retention and expansion. In a usage-based model, heavy users face rising bills and start optimizing usage down to reduce spend. The customers who were most engaged under subscription become the most cost-conscious under usage. Net revenue retention (NRR) in the first 12 months after a poorly designed usage-based transition often drops 6 to 10 points as customers "right-size" their usage.

Third, sales motion complexity. Usage-based pricing removes the clean annual contract value (ACV) number that makes enterprise sales motion work. Procurement teams don't like open-ended commitments. Enterprise deals slow down. Deal velocity falls. In a $30M ARR business transitioning to usage-based pricing, if average enterprise deal cycle extends by 30 days, the quarterly impact on recognized ARR is significant.

Building the System

Evaluating whether usage-based pricing is right for your business requires three tests.

Step 1: Test whether your usage metric is a reliable proxy for value delivered. The classic test: do your best customers by NRR use more than your worst customers by NRR? If high-usage customers churn and low-usage customers expand, usage isn't tracking value. If high-usage customers are your most satisfied, most likely to renew, and most likely to expand, usage is a valid value metric and usage-based pricing will likely work.

Step 2: Test whether your cost structure scales with the usage metric. If you're building on infrastructure where your costs scale with customer usage, usage-based pricing aligns revenue and cost and can generate healthy contribution margins. If your costs are largely fixed regardless of usage level, usage-based pricing adds revenue volatility without matching cost relief.

Step 3: Model the transition impact on NRR before you execute. Take your top 30 accounts and model what they would have paid last year under usage-based pricing versus what they actually paid. How many would have paid more? How many less? The answer tells you whether usage-based pricing is an expansion play or a pricing reduction dressed up as model innovation.

What Falls Apart

A data enrichment SaaS company at $26M ARR switched to usage-based pricing after seeing a competitor do it successfully. Their hypothesis: customers with smaller initial needs could start cheaper and expand as they grew. The transition was announced with a clear migration plan.

What they hadn't modeled: their largest 15 accounts, representing 38% of ARR, were already at the ceiling of their use case. Under usage-based pricing, those accounts would optimize down, not up. The company discovered this six months post-transition when those accounts began renegotiating at renewal based on actual usage data that showed they were over-contracted in the subscription model.

Before: $26M ARR subscription, 38% concentrated in high-usage accounts, NRR 104%.

After: Usage-based transition, top 15 accounts contracted at renewal, NRR dropped to 88% in the first year. They reintroduced a hybrid model with a subscription base plus usage overage, which stabilized NRR at 97% over 18 months, but with 18 months of damage first.

Do This in the Next Seven Days

Take your top 20 customers and estimate what they would have paid last year on a pure usage-based model. Compare that to what they actually paid. If the median number is lower, usage-based pricing as designed is a pricing reduction for your best accounts, not an expansion play.

That's worth knowing before you announce the transition.

Assess Your Commercial Health to get a structured view of whether your current pricing model architecture is capturing available value.

For a deeper look at willingness to pay as an input to usage-based design, see Why Your Instincts Are Wrong About Willingness to Pay Research and Stop Guessing: Monetization Strategy Driven by Data.