Their event pricing would require you to log something on the order of tens of millions of events to get at a 10k monthly bill. If you were turning it on in a high-volume environment, how did you unexpectedly run up a 10k bill after doing the math in half a day? Something about this doesn't add up.
This a while ago, so I don't remember the source of the miscalculation.
We have a popular website, and this was the first analytics product outside of Google Analytics that we reached for. We used several other metered tools that didn't charge extreme amounts for our volume, so this was our very first real surprise.
They're selling to websites with low visitor counts, expecting people to sample small subsets traffic, or catering to large orgs with huge budgets for this.
The out of the box behavior for new customers should be to downsample until a flag is turned on or to set billing thresholds.
Or you can go with a self-hosted solution, but depending on how you host it, you might run into the same issues. This is why I prefer to have fixed-costs VPS/dedicated machines instead of "auto-scalable" infra. For analytics, with one single server of $5/mo you can track millions of pageviews per month.
I’ve never heard of PostHog and from “Side Project Insurance” I thought this was some form of super cheap E&O insurance which sounded like a neat idea.
11 comments
[ 4.1 ms ] story [ 39.7 ms ] threadI did some napkin math, turned on their product, and ran up a $10,000+ bill in my first half day of trying the service.
You have to talk to their sales funnel to ask for forgiveness. It was an awful experience.
Not only do they not have reasonable billing limits or triggers in place, but these are quite hidden.
We're using free tools now: Google Analytics (obviously), Microsoft Clarity (session recording), and hand-rolled feature flags.
2. You shouldn't have to read docs to find it.
3. It should have been covered in their onboarding.
I'm convinced this is a dark pattern for their sales funnel.
From the PR this links to.
Their event pricing would require you to log something on the order of tens of millions of events to get at a 10k monthly bill. If you were turning it on in a high-volume environment, how did you unexpectedly run up a 10k bill after doing the math in half a day? Something about this doesn't add up.
We have a popular website, and this was the first analytics product outside of Google Analytics that we reached for. We used several other metered tools that didn't charge extreme amounts for our volume, so this was our very first real surprise.
They're selling to websites with low visitor counts, expecting people to sample small subsets traffic, or catering to large orgs with huge budgets for this.
The out of the box behavior for new customers should be to downsample until a flag is turned on or to set billing thresholds.