Where are my users coming from?
Real attribution, not vibes. Sources ranked by quality, not volume.
You'd think GA4 source data is the answer — Astra cross-references UTM, first-touch, and a post-signup question to give you real attribution.
The short answer
Astra maps every signup to its real source weekly using 3 signals stitched together. (1) UTM + GA4 first-touch attribution — captures direct paid + organic search referrals. (2) PostHog session recording metadata — catches indirect referrals (Reddit/Twitter clicks that lose UTM). (3) Post-signup 'how did you hear about us?' question — captures word-of-mouth, podcast, conference, friend recommendation. She cross-references all 3, resolves conflicts (post-signup self-report > UTM > first-touch), and outputs a weekly source breakdown ranked by both volume AND quality (paid conversion rate × LTV per source). The Lark report shows: top 5 sources by paid customers (not signups), each source's CAC, blended CAC, and the cheapest channel to scale next. Most founders find their #1 paying-customer source isn't where they're spending money — and the gap is the next move.
How Astra actually does it
- 1Map UTM + first-touch from GA4
GA4 acquisition reports for every signup. Capture utm_source, utm_medium, utm_campaign + first-touch organic source. Backfill last 90 days.
GA4 - 2Layer PostHog session metadata
PostHog initial_referring_domain. Catches Twitter/Reddit/HN signups that drop UTM in shared links. Reconcile with GA4 attribution.
PostHog - 3Capture post-signup self-report
Add 'how did you hear about us?' as required field on signup or first onboarding step. 7 options + free text. This is the highest-confidence signal.
PostHogIntercom - 4Resolve conflicts and rank by quality
Source priority: self-report > UTM > first-touch. Compute paid conversion rate per source, LTV per source, true CAC per source. Rank by quality not volume.
- 5Deliver weekly report
Lark: top 5 sources by paid customers, CAC per source, blended CAC, biggest opportunity (cheapest scale-up source).
Lark
What it looks like in chat
Weekly Lark report: top 5 sources by paid customers, CAC per source, blended CAC, top opportunity to scale.
Weekly Friday 4pm. Quarterly deep-dive on source quality trends.
Ask Astra this right now
We'll spin up your workspace, hand the prompt to Astra, and you see the answer in 60 seconds. Free.
Try this with AstraFrequently asked questions
Why is GA4 attribution alone not enough?
GA4 misses ~30-40% of true source — anything from a non-UTM tagged link (most social shares), email forward, copied URL, or app-to-app click loses attribution. Layering self-report + PostHog session data captures these. Confidence on source attribution jumps from ~60% (GA4 only) to ~88% (3-source stitch).
Won't asking 'how did you hear about us?' hurt signup conversion?
Yes — by ~2-4% if it's required at signup. Better placement: first onboarding step after signup ('one quick question'). Conversion impact <1% there, response rate ~85%. Astra defaults to onboarding-step placement and tracks both signup conversion and response rate weekly.
What if my customers are mostly word-of-mouth?
Common for product-led growth. Astra captures word-of-mouth via self-report and segments it by referrer (friend/coworker, podcast guest, Twitter post, etc.). The actionable insight is usually: invest in the highest-converting word-of-mouth driver (often podcast appearances or specific Twitter influencers) to make organic word-of-mouth more frequent.
How do I distinguish 'direct traffic' from real direct?
Direct traffic in GA4 is mostly mis-attributed referral (lost UTM). Astra cross-references with self-report: real direct = customer typed your URL knowing your brand (usually previous trial users or word-of-mouth). Mis-attributed direct = self-report says Reddit/podcast/Twitter. She separates these cleanly so you don't double-count word-of-mouth as 'direct.'
Run your one-person company.
Hire your AI team in 30 seconds. Start for free.
Free to start · No credit card required · Set up in 30 seconds