Feature flag pricing models: why we don't charge by MAU

Most developer-tool pricing charges for scale: per monthly active user, per seat, per million requests. Whatever the metric, the shape is the same: the bill climbs exactly when the product is working, because the thing being metered is your growth. Sign up small, succeed, and watch the line item become a conversation with finance.

We priced smplkit against that shape, and this post explains the model. Fair warning up front: pricing developer tools is hard, the unit of value doesn’t map cleanly onto any billing metric, and we spent more time on this than on some features.

The models already out there

LaunchDarkly charges by monthly active users across all environments; a startup with 50,000 users and meaningful flag usage still pays for the 50,000 users, and the bill tracks user count from there. ConfigCat prices by SDK keys plus MAU, with rates that improve at enterprise volume — a reasonable structure, but the MAU dependency is still in it. Flagsmith’s open-source version is free, and its hosted service charges per SDK request, which ties cost to usage more directly but gets expensive for high-traffic services and quietly discourages evaluation-heavy patterns. DevCycle splits the difference with seats for the console and API-call pricing for evaluation, which is two models to reason about instead of one.

These are all defensible designs. But all of them charge for scale in a way that makes the product more expensive exactly when it’s working well — when lots of users are evaluating lots of flags frequently. We wanted heavy usage to feel like the point, not a penalty.

Our model: charge for what you manage, not the traffic

Smpl Flags has four flat tiers: $0, $99, $249, and $599 a month. What moves between them is how much you’re managing — 10 flags on the free plan, then 50, 250, and unlimited — plus how elaborate each flag’s targeting can get: five rules per flag on free, unlimited on every paid plan.

What never moves: traffic. Servers, SDK connections, API requests, and evaluations are unlimited on every tier, including free. The free plan caps the console at one write user; read-only seats are unlimited everywhere.

The reasoning: the number of flags you manage tracks the value you’re actually extracting. Ten flags is a team trying the product. A few hundred flags with real targeting rules is a team that ships with it. Evaluations track nothing — a flag is worth the same whether it’s checked a hundred times a day or a hundred million — so we don’t meter them.

The same idea, applied to Config and Logging

Smpl Config runs $0 to $249 a month on the same shape: tiers differ by how many configurations you manage, how many keys each one holds, how deep inheritance nests, and how large a value can get. Smpl Logging runs $0 to $249, gated by log groups and how many loggers you can control at runtime — you can see every logger on every plan; the tier decides how many you can actually flip.

The pattern is the same everywhere: the free tier is genuinely useful for small projects, and paid tiers price the scale of what you’re managing, never the traffic you serve.

Discounts, not bundles

Products are sold individually, and subscribing to more of them discounts all of them: 15% off with two products, 33% with three, 35% with four, 40% with all five. It lands on the invoice automatically — no bundle SKU, nothing to manage or forget to cancel. Teams that adopt one smplkit product tend to add the others within months, so the schedule prices what actually happens.

The part we can’t know yet

The competitive analysis behind all this used each vendor’s public pricing as of early 2026, and public pricing rots quickly; check their pages.

The bigger caveat: this pricing hasn’t been validated by enough real customers making real purchase decisions. It reflects our best first-principles read of the market, and the hardest thing about pricing is that you can’t know if it’s right until customers tell you — either by paying or by not paying. We’ve optimized for being in the ballpark on principles we can defend.