Adopt

5 surfaces

Surfaces

Integrations

Strategy & Discovery

2
VistalyProductboard

Research & Insights

1
Dovetail

Knowledge & Docs

1
Notion

Design

1
Figma

Analytics

1
PostHog
AnalyticsBeta

Coming from PostHog

22% entity overlap

Import

Open-source analytics with native hypothesis fields: PostHog experiments become hypothesis_claim + experiment node pairs.

The structural finding

PostHog is the only analytics tool in the series with a native `hypothesis` field on experiments. The adapter extracts it as a separate hypothesis_claim node, making PostHog the most UPG-aware analytics tool in the ecosystem.

What UPG adds

1

Hypothesis chains: PostHog experiment hypotheses become typed hypothesis_claim nodes connected to experiments and outcomes.

2

Feature flag traceability: flags become feature nodes, with rollout history traceable to the decisions that drove it.

3

Open-source alignment: PostHog and UPG are both open. No vendor lock-in on either side.

Schema mapping

Direct matches
Feature Flag→feature_flag
Experiment→experiment
Experiment `hypothesis` field→hypothesisDirect field match, unique in analytics tools
Variant→variant
Cohort→behavioral_segment
Survey→survey_response
Dashboard→dashboard
Insight (saved query)→report
Event Definition→event_schema
Maps with caveats
Session Recording→observationRecorded session = behavioral observation with media
Group (B2B)→accountWhen group_type = company/org
What UPG adds
outcomeopportunityobjectivekey_resultfeaturepersona
Skipped:Raw event stream · Data warehouse sources · Plugin system

The key edge

The single most important semantic relationship this integration enables.

key_result_tracked_by_metric

Defined in the UPG specification

Sample import

PostHog experiment with native hypothesis field → UPG hypothesis chain

Input (from PostHog)
Output (UPG nodes)
Experiment
Homepage CTA test
hypothesis: "Moving the CTA above the fold will increase signups by 15%"
→
experiment
Homepage CTA test
Experiment.hypothesis field
Moving the CTA above the fold will increase signups by 15%
→
hypothesis
Moving the CTA above the fold will increase signups by 15%
↳ experiment_run_validates_hypothesis
Feature Flag
enterprise-beta
active: true, rollout: 20%
→
feature_flag
enterprise-beta
↳ service_toggles_feature_flag
Cohort
Activated in first session
→
behavioral_segment
Activated in first session

Roundtrip

API ROUNDTRIPPlanned
PostHog ──import──▶ UPG ──enrich──▶ UPG ──push via REST──▶ PostHog

Import experiments and feature flags → enrich with UPG opportunity and outcome context → push strategic rationale back to PostHog via API.

Writes back
  • ·Outcome and opportunity context written back to PostHog experiment descriptions
  • ·UPG learning node conclusions written back to experiment result summaries
  • ·Feature flag rollout rationale linked from UPG decision nodes back to flag metadata

How to import your PostHog data

Via the CLI (coming soon)

The PostHog adapter is built and tested. A one-command import via the UPG CLI is in development:

npx unified-product-graph import --from posthog

This will fetch your PostHog data, map entity types automatically, and write everything to your graph. Today you can use the Markdown import or the adapter directly in code.

▸ For developers: use the adapter directly
import { PostHogAdapter } from '@unified-product-graph/adapters'
// Pass pre-fetched PostHog data as SourceItem[]
const adapter = new PostHogAdapter()
const result = await adapter.convert(items)
// result.nodes → typed UPG entities
// result.edges → canonical relationships
// result.warnings → items needing review

Get started

The PostHog adapter ships in @unified-product-graph/adapters. Install it, then run the import command.

# one-time install
$ npm install -g unified-product-graph

# then import
$ upg import --from notion

Other tools in Analytics

← All integrations