Miro, OpenAI, Loom, and Splunk on Pricing as a Growth Lever

Pricing Is a Feature, Not a One-Time Decision

Most teams treat pricing like a tax form: set it once, file it away, never touch it again. That's a mistake.

Akshay Sharma from Miro thinks about it differently. Pricing is another product feature. It changes customer behavior. You should evolve it as your goals shift, just like you iterate on onboarding or activation flows.

Early on, Miro had a 5-user minimum for paid plans. Seems counterintuitive when you're trying to grow revenue. But the goal wasn't revenue. It was user love. The team knew collaboration happened when multiple people showed up on a board. They set the threshold high to ensure buyers actually got value.

It was a conscious choice because the outcome was users need to love the product. That would help get to product market fit.

— Akshay Sharma, Miro

As Miro matured, they introduced a business plan with a 20-user minimum. Again, not to squeeze more dollars out of small teams. It was to prevent customers from paying for features like SSO before they were ready to use them. Later, when the product hit scale, they dropped the minimum to one user. Same lever, different outcomes, all intentional.

When Lowering Price Drives More Growth Than Raising It

At OpenAI, the pricing playbook flips conventional wisdom on its head. As models get better, prices go down. They've changed pricing far more than the standard "once or twice a year" advice.

Why? The goal isn't short-term revenue extraction. It's adoption. The AI market is new. OpenAI wants as many companies as possible experimenting with the platform. Lower prices mean more usage, more integration, more lock-in.

You think as we increase value, you should increase price. But really it's while we're trying to ignite the market and just get as many companies exposed to AI as possible.

— Alison Harmon, OpenAI

This only works if you're in a land-grab phase. If you're optimizing for revenue per customer or margin, this approach will destroy your P&L. But if you're OpenAI and the prize is defining the category, it's the right call.

Most Pricing Experiments Fail—and That's Fine

Janie Lee from Loom has a blunt take: don't experiment unless you're willing to roll it back. Pricing tests sound good in theory, but they require volume, time, and a strong stomach for bad news.

At Miro, the team tested letting free users access GIFs. Conversion spiked 15%. Win. Then they tested giving starter plan users a taste of top business-tier features, assuming it would drive upgrades. Conversion dropped. Users thought, "If that's all business offers, I'm good."

Assume the mindset that failures will happen and all of the de-risking and insurance you're doing that helps you continue doing experiments and learning, that's more important than the success of any given experiment.

— Akshay Sharma, Miro

Before you run a test, ask: What's the goal? Is it validating product-market fit for a new SKU, or de-risking a change to your core package? The answer changes whether you experiment at all. If you're enterprise or sales-led, you likely don't have the volume. Use surveys, betas, or geographic pilots instead.

Carsten Holm from Splunk recommends finding your "South Africa"—a market where you can test quietly without blowing up your brand if things go sideways. For some teams, that's Australia. For others, it's a specific customer segment or a limited beta.

Packaging Is About Segmenting Willingness to Pay

Splunk's platform sells separately: the core product, then add-ons like Enterprise Security. Customers buying IT Service Intelligence had to purchase two line items. Friction everywhere.

The team bundled it. They packaged the add-on with the platform and monetized on platform usage. Result? Higher adoption of the add-on because customers felt they could experiment. And higher ASP on the platform because the perceived value went up.

User adoption was much, much higher because people felt like they can experiment without any real risk and commitment. And the ASP for the platform actually went up.

— Carsten Holm, Splunk

Packaging isn't about good-better-best tier labels. It's about capturing different willingness to pay across customer segments. The mechanics—bundles, seat minimums, usage caps—are just tools to segment smartly.

The One Thing You Can't Change Later

You can tweak price points. You can adjust limits. But your value metric—the thing you charge for—is nearly impossible to change once it's in market.

Carsten and Akshay both flagged this. Seat-based, usage-based, event-based—pick the metric that reflects how customers derive value and how they think about fairness. If you get it wrong early, you're stuck.

Allison from OpenAI suggests you can change pricing more often than conventional wisdom says—if you're only lowering it or adding value. ChatGPT Plus has stayed the same SKU for over a year. Even in a fast-moving market, stability in packaging builds trust.

Janie's rule: you can tweak constantly, but don't overhaul packaging more than once every 1–2 years. That's when your goals, customers, or product actually shift enough to justify the disruption.

If you're early and don't have a pricing expert, go find one. The decisions you make now—minimum seats, value metric, tier structure—will haunt you or help you for years.