spirby.Spirby home
all posts
by Dima

why per-tracked-user pricing breaks the feedback loop

metering on tracked users doesn't just raise the bill. it quietly teaches your team to collect less feedback. how the metric attacks each step of the loop.

a founder we know runs a public board on a metered tool. last quarter she stopped linking it from the product nav. not because the board wasn't working. because every new person who clicked it became a tracked user, and tracked users are what the invoice counts. the board still exists. almost nobody finds it any more.

we've written before about why we don't meter on tracked users. that post is about the invoice and the business model. this one is about something quieter and worse: what the metric does to the way a team actually runs its feedback loop. the bill is the part you notice. the behavior change is the part that does the damage.

the loop, in four steps

a feedback loop is four moves repeated forever:

  1. collect — people post and vote.
  2. read — you read what they posted.
  3. decide — you choose what to build, and what not to.
  4. close — you tell the people who asked.

the loop is a flywheel. wider collection means better signal, better signal means better decisions, better decisions and visible closes mean more trust, more trust means more people show up to collect from next time. it compounds, or it doesn't spin at all.

every one of those four steps is something you want to do more of. per-tracked-user pricing taxes three of them.

step 1: it taxes collection at the source

a tracked user is a unique identified end-user who touched your boards in a rolling window. the meter runs on reach. so the rational move, the one every team eventually discovers, is to reduce reach.

don't link the board from the nav. don't email the whole list about it. don't put it behind the "give feedback" button in-app. don't drop it in the launch announcement. each of those is exactly the thing a feedback board is for, and each one shows up on the invoice a month later.

nobody decides this in a meeting. it happens one small omission at a time. the board gets quietly demoted from "front door" to "link in the footer," and the people who would have posted the most useful things — the ones at the edge of becoming customers — never arrive.

step 2: it biases the sample toward who you already have

here's the part that's easy to miss. the people who are already tracked users cost you nothing extra to hear from again. they're paid for. the people who aren't yet tracked are the ones who tip you into the next tier.

so the cheapest feedback is from your existing power users, and the most expensive feedback is from newcomers. the meter puts a price on novelty.

that's backwards. the 50-vote rule post is about finding the requests with real mass behind them — but mass from whom matters. a loop that's cheapest to run on your existing base will, over time, optimize the product for the people you've already won and go deaf to the people you're trying to win next. you end up with a very well-served core and a product that stops reaching outward. the metric paid for that drift.

step 3: it taxes closing the loop

closing the loop means going back to the people who asked. notify every voter when a status changes. link the public roadmap from your changelog. let a "shipped" note pull someone back to the board to see what else is moving.

all of that is re-engagement, and re-engagement keeps people inside the tracked-user window. the discipline we think matters most — closing the loop even when the answer is no — is, under a reach meter, a line item. you are billed for following up.

a tool should make the highest-trust action the cheapest one. per-tracked-user pricing makes it the most expensive one. teams respond the way they always do to a tax: they do less of it. the loop stops closing, and a loop that doesn't close isn't a loop. it's an inbox.

step 4: the tax you pay in calendar time

step three of an actual feedback practice is read every post. it takes an hour a week and it's the part you cannot delegate.

now watch what the meter does to that hour. instead of reading the inbox, someone on the team is in the billing dashboard working out whether this month's reach tips you into the $311 tier, and which integration to throttle to stay under. that's the choreography metered tools train into you: pruning identified users, gating the board behind a login to suppress the count, debating whether the launch email is "worth" the tracked users it'll add.

every minute of that is a minute not spent reading what people actually wrote. the meter doesn't just tax the loop's inputs and outputs. it reallocates the team's attention from the one step — reading — that no tool can do for you.

what's fair to meter, and what isn't

this isn't an argument that all usage-based pricing is wrong. meter on the things that genuinely cost the vendor more as you use more, and that you'd want to use more of anyway: storage, seats on your side, api throughput. those scale with cost and they don't ask you to hide from your own customers.

reach is different. the marginal cost of one more person viewing a board is approximately zero. metering it doesn't recover a cost. it puts a price on the single behavior — being seen by more people — that a feedback board exists to produce. that's the tell. a meter that taxes the product's core purpose isn't aligned with the cost. it's aligned against the loop.

the cleaner default

we built spirby flat for exactly this reason, and we'll keep that argument where it belongs rather than relitigate it here. the short version: share the board everywhere, link it from the nav, email the whole list, notify every voter on every status change. none of it moves the number on the invoice. the loop is free to spin as fast as your discipline allows.

that's the whole point of getting the meter off reach. the constraint on your feedback loop should be how much you're willing to read and respond to. it should never be a billing tier you're quietly steering around.

if you'd rather run the loop wide open, start a trial. collect from everyone. close with everyone. the bill won't notice.