Blog
product 3 min read

How to fix your $8k PostHog bill

Same product events and session replays for about 7x less. A quick guide to moving them from PostHog to Maple, with the code.

How to fix your $8k PostHog bill. About 7x less. Two bars compare one month of 5M events and 20k replays: $627.50 on PostHog, $89 on Maple.
On this page

My timeline this week is pretty much all PostHog invoices. If one of them is yours, this is the quick version of moving your analytics over to Maple.

Maple covers the stuff most teams use every day: product events, funnels, web analytics, a page per user, and session replay. An event costs the same whether the person is signed in or not. Every replay also links to the backend traces behind it.

Here’s what that comes to for a normal SaaS month, 5 million events from signed-in users and 20,000 recorded sessions:

One month, same usage
About 7× less on Maple
PostHog $627.50 Maple $89
PostHog Maple 5M events from signed-in users $560 $20 20k session replays $67.50 $30 Base plan $0 $39
The Maple plan also includes 100 GB each of traces, logs and metrics.

What maps to what

In PostHogIn Maple
EventsProduct events
Funnels and pathsProduct analytics
TrendsA dashboard chart on product events
Web analyticsWeb analytics
PersonsUsers, with each person’s events and replays
Session replayReplays, linked to the traces and errors
HogQLSQL

What it looks like

Funnels count people across the browser and your backend, with conversion, time to convert and where people drop off. These two are the real charts from the app on demo data, so hover them.

/product-analytics · funnels · last 7 daysDEMO
Entered1.2Kpeople reached step 1
Conversion5.1%61 reached plan_started
Time to convert2d 6hmedian
Biggest drop84%before plan_started
Signup to paid plan, counted by person
Conversion rate5.1%
signup_started100%1.2K
signup_completed62%742
62%38%3m
onboarding_connected32%388
52%48%1h 52m
plan_started5.1%61
16%84%2d 3h

Paths show what people did after (or before) any event or page.

/product-analytics · paths · last 7 daysDEMO
What people did after signup_completed742 people

Every replay sits next to the requests it made, so a failed checkout shows the 500 and the trace behind it.

A session replay of a checkout page, with the event list beside it showing a POST /api/checkout that returned 500 because payment-svc timed out.
A replay with the failing request right there in the event list.

And each user gets one page with their sessions, errors, traces and events.

One user's page in Maple: 14 sessions, 4 errors, 2.1K requests and 230 events, with a timeline mixing a checkout error, a failed API request, a checkout_started event and a session to replay.
One user, everything they did.

Switching

1. Add the SDK.

bun add @maple-dev/browser
import { MapleBrowser } from "@maple-dev/browser"

MapleBrowser.init({ ingestKey: "maple_pk_...", serviceName: "web" })

That’s enough to get page views, sessions, replays and errors coming in.

2. Swap your calls. They map almost one to one:

// posthog.identify(user.id, { email: user.email })
MapleBrowser.identify({ id: user.id, email: user.email })

// posthog.capture("plan_started", { plan: "pro" })
MapleBrowser.track("plan_started", { plan: "pro" })

// posthog.reset()
MapleBrowser.identify(null)

If your calls are all over the codebase, put both SDKs behind one small analytics.ts for now. Then you only have to touch one file when you remove PostHog later.

3. Move server-side events. You post them to one endpoint, from any language:

curl -X POST https://ingest.maple.dev/v1/events \
  -H "Authorization: Bearer YOUR_INGEST_KEY" \
  -H "Content-Type: application/x-ndjson" \
  --data-binary $'{"name":"plan_started","user_id":"user_123"}\n'

If you already trace that request with OpenTelemetry, just add maple.product_event.name to the span and skip the extra call.

4. Rebuild the funnels you actually look at. Usually that’s like five or ten. Skip the rest.

5. Run both for a week. Once the numbers agree, delete the PostHog lines.

What Maple doesn’t do yet

Tbh there’s a list:

  • Feature flags, experiments and surveys
  • Retention and cohort charts
  • Heatmaps
  • Mobile SDKs and mobile replay

If you use any of these, keep a small PostHog plan for just those and move the rest. Most of the bill is usually events and replays anyway.

Try Maple free for 14 days, or read the product events docs.

All posts

Read next

See your first trace today.

Add the SDK, point your OTLP exporter at Maple, and traces arrive.

maple.dev: observability on OpenTelemetry