Key Takeaways: Intelligent Usage Metrics

Rob Litterst

6d • Pricing & Packaging Strategy

Thanks so everyone who attended Office Hours yesterday! It was a lively session that was jam-packed with insights from

@Ulrik Lehrskov-Schmidt

and

@Emil Eriksson

Breaking down my 5 top takeaways below for those who couldn't make it:

1️⃣ Measure internally at maximum granularity, but expose a metric customers intuitively understand.

Emil made the point clearly: internally you want usage data as granular as possible to understand costs, but what you surface to customers must map to an outcome that makes sense in their world. His example of Lovable charging fractional credits — where one prompt costs 15 credits and another costs half a credit — illustrates what happens when that translation layer is missing. The customer experience becomes a black box.

2️⃣ Build a usage baseline before you price anything.

Emil's strongest practical advice was to run shadow pricing for a few months against real usage data, then sit down with friendly customers and say "here's what you would have paid." This prototype-testing approach — borrowed from product design — de-risks the model before launch and creates internal buy-in across finance, product, and sales.

@Steve Blanck

from M-Files distilled it well: the first ask of any internal stakeholder should simply be "get the data."

3️⃣ Credit systems beat pure metering for handling seasonality and revenue predictability — but only if they self-balance.

Ulrik laid out a specific architecture: monthly credit subscriptions that roll over (so customers don't over-calculate), combined with an auto-rebalancing mechanism where the next contract period adjusts to match actual prior usage. This "transposed usage-based pricing" flattens revenue volatility, reduces renewal friction, and — in his experience — keeps balance-sheet carry-over under 5% of ARR. The key design details: monthly credits rot after 12 months, annual credits roll once then expire, and you recognize revenue at the average price-per-credit in the customer's balance.

4️⃣ Packaging differentiates customers that pricing alone can't.

Ulrik was direct: when you have wildly different customer segments (Paramount vs. a YouTuber, in

@Majed Alhajry

's case), you won't solve the value-capture problem with a single pricing metric. You need different packages — different bundles of functionality, service levels, and volume structures — so that enterprise customers get unlimited seats but pay on a value-aligned volume metric, while smaller customers stay on simpler seat-based plans. The pricing model shift itself becomes a vehicle for repositioning the product conversation with enterprise buyers.

5️⃣ Treat pricing as a living system, not a one-time architecture decision.

Emil flagged what he sees as a hidden trap: teams build "glue and duct tape" to wire up their first pricing model, and that technical debt then prevents them from iterating. In an AI-accelerated environment where cost structures and product capabilities shift weekly, your billing and usage infrastructure needs to support experimentation — not just the model you launched with. If you can't change your pricing without an engineering project, you've already lost agility.

Thanks again to all who joined, and for asking such thoughtful questions. Looking forward to the next one!

Have a great weekend,

Rob

2 comments