pex
Growth Intelligence

Your team’s memory. Written in code.

Beliefs, scoring models, experiment ledger, and the Growth Journey: all versioned, auditable, and compounding. When your VP Growth leaves, their knowledge stays.

Belief graph · live

12 beliefs
Enterprise demos85%LinkedIn > Meta B2B72%Pricing page matters91%Pricing anchor works89%Webinar → pipeline48%Social proof → trust55%Sales-assisted ICP68%Onboarding email lift82%Demo duration → close77%Re-engagement timing64%Feature flag drop34%Free trial activates76%
Belief updated

Beliefs

A graph. Not a wiki.

Every growth assumption gets written down as a belief. “Enterprise demos convert at 3x self-serve.” “LinkedIn ads outperform Meta for B2B SaaS.” Linked in a graph. Updated by experiments. Auditable forever.
  • Beliefs are a first-class primitive in the data model, not a tagged doc and not a Notion page that will go stale in a quarter.
  • Confidence adjusts automatically as experiments complete, so a belief you wrote a year ago tells you how much to trust it today.
  • Linked in a graph, not buried in folders. One experiment’s outcome can update every belief it touches, including the ones you’d forgotten about.

Enterprise demos convert at 3x self-serve.

85%
12 experiments2 linked

Confidence history

6 snapshots

50%+35 pts85%

LinkedIn ads outperform Meta for B2B SaaS.

72%
8 experiments2 linked

High-anchor pricing tier lifts mid-tier conversion.

91%
15 experiments1 linked
Linked to 5 other beliefs in the graph

Scoring

Scoring as hypothesis, not black box.

Most lead scoring tools give you a number. Apex gives you the model. Version every weight change. Detect drift when the model stops predicting well. Let AI suggest corrections, or auto-apply them within the guardrails you set.
  • Versioned weights with one-click rollback. No more “this worked last month, what changed?” post-mortems.
  • Drift detection scores the whole model weekly and flags when the predictions stop matching reality.
  • AI proposes weight corrections grounded in your actual outcomes, not generic ML priors; approve or auto-apply with guardrails.

Scoring model · v3

healed 2d ago healthy
optimaldrift detectedauto-corrected

Weights

v3 → v3.1

3 variables retuned

Drift score

42 → 8

Back in healthy band

Lift

+4.2%

Rescored cohort accuracy

The Ledger

Every experiment. Forever.

525+ tests shipped. Zero lost to memory. Every experiment gets a permanent entry. Promoted winners leave a trail of beliefs updated. Losers leave context for the next hypothesis.

Decision ledger

Every experiment, forever

525 tests shipped0 lost to memory
exp-523

Landing page hero · v3 benefits-first

3 beliefs updated·Benefits-first copy lifts signup 22%

2 days ago
exp-518

Email subject · urgency vs. curiosity

1 belief updated·Curiosity lost to urgency by 14% (but lower open quality)

5 days ago
exp-511

Pricing card ordering (4 variants)

2 beliefs updated·No significance after 21 days; effect too small to detect

1 week ago
exp-502

Demo CTA placement on pricing page

1 belief updated·Above-fold CTA +18% demo requests

2 weeks ago
exp-494

Sign-up friction · remove company field

4 beliefs updated·Fewer fields = +34% completion, quality unchanged

3 weeks ago
exp-481

Trial length · 14 vs 30 days

1 belief updated·Activation timing matters more than trial length

1 month ago
exp-467

Social proof badge above fold

2 beliefs updated·Customer logos slowed perceived trust for enterprise

1 month ago
exp-450

Onboarding checklist v2 · 3-step

2 beliefs updated·3-step checklist + email reminders → +41% activation

6 weeks ago
exp-428

Referral incentive · cash vs credit

1 belief updated·Account credit outperforms cash by 2.3x for referrer

2 months ago
exp-410

Product tour · auto-start on first login

No beliefs changed·Users skipped tour at 67%, no activation lift

2 months ago

Losers leave context for the next hypothesis. Losers are learning too.

all 525

Growth Journey

Your team is also a growth experiment.

The Growth Journey tracks engagement score, level (Observer Explorer Architect Apex), and badges earned. Every step surfaces the next best action inside the product.
  • Named levels (Observer, Explorer, Architect, Apex) celebrate real milestones, not vanity points earned for logging in.
  • Next-best-action surfaces contextually inside the product, not as a static onboarding checklist users abandon on day two.
  • Churn risk is visible inside the dashboard before the user goes quiet, so your CSM team is told who to call, not who just left.

Growth Journey

Level 4 of 5
0/ 100

Current level

Architect

9 points to Apex

Recent unlocks

Multi-Source

3+ integrations connected

5d ago

First Experiment

Your first test went live

2w ago

Scoring Architect

Custom model configured

3w ago

Full Spectrum

4 / 5

Connect an integration in every category (payments missing).

Proactive Insights

Insights that do the work.

Apex continuously scans your data and surfaces what actually needs your attention: drift, experiments ready to ship, data gaps, cascading belief updates. Each insight comes with a one-click action.

Insights feed

Ranked by severity and recency

live4 of 4
Drift#1

Score drift detected · 3 variables below threshold

Your ICP scoring model stopped tracking outcomes well over the past two weeks. Company size, pipeline stage, and demo attendance are the main drivers.

Experiment#2

Experiment ready to graduate · 2x uplift, 97% confidence

Homepage hero v3 has outperformed control for 12 consecutive days. Sequential test confirms significance. Promote to ship permanently.

Data gap#3

No conversion events from /pricing in the last 7 days

The pricing page is receiving traffic but no goals are firing. Either the tracking changed or a page update broke the form selector.

Correlation#4

ICP weighting update affects CAC · belief chain triggered

A recent scoring model change cascaded through 3 downstream beliefs. Review the lineage to decide whether the cascade is intentional.

Metric shifts

Week-over-week anomalies

Goal pacing

Projected vs. target

Correlation hints

Connected belief updates

Compounding Moat

Your competitor forgets. Apex teams compound.

Every quarterly planning cycle at most companies starts from zero. The lessons from last quarter live in a deck that nobody reads. Apex teams don’t reset; every experiment makes the next one smarter. That’s the moat.
  • Beliefs persist across leadership changes; a new VP Growth inherits the last VP’s thinking with confidence scores attached.
  • Scoring models keep improving between quarters; the weights a previous team tuned keep drifting toward reality without anyone’s attention.
  • New hires read the ledger on day one and catch up on the “why” in hours. The answers to their 30 “why did we decide that?” questions are already written down.

Institutional intelligence over time

24 months

Apex (compounding)Resets each quarterQ1Q2Q3Q4

After 24 months

Apex teams know 340+ beliefs

Per compounded experiment

Their competitors

Start over every quarter

Knowledge dies with the team

A new category

What does “growth intelligence” actually mean?

The category didn’t exist before. Teams improvised with docs, wikis, and dashboards. Here’s what each approach actually preserves, and what Apex adds.

Beliefs written down

Spreadsheets + Notion

In someone's personal doc

BI dashboards

Not part of the model

Apex

First-class primitive

Tied to experiments

Spreadsheets + Notion

Manually, if anyone remembers

BI dashboards

No experiment primitive

Apex

Every test updates beliefs

Updated with new data

Spreadsheets + Notion

Only when someone looks

BI dashboards

Charts refresh hourly

Apex

Confidence adjusts automatically

Audited over time

Spreadsheets + Notion

Lives in a Notion graveyard

BI dashboards

Query history, not reasoning

Apex

Full ledger, forever

Survives team turnover

Spreadsheets + Notion

Dies with the person

BI dashboards

Dashboards stay, context leaves

Apex

Your institutional memory, in code

The verdict

The system that thinks, not just the system that reports.

Get started

Three steps to a brain of your own.

Intelligence compounds with every experiment you run. The sooner the graph starts, the sooner your next decision is informed by the last one.

1

Step 1

Install

Drop the snippet

One tag in your <head>, or npm for SPAs. Events start flowing immediately, no configuration required.

2

Step 2

Watch the graph build

Beliefs form as data arrives

The first few weeks of data auto-generate starter beliefs. You review, approve, and extend them. Each experiment adds more.

3

Step 3

Start compounding

Every cycle makes the next one smarter

Scoring models self-correct. Insights queue up. The Growth Journey levels up with your team. Institutional memory, in code.

Growth intelligence shouldn’t have an expiration date.

Beliefs you wrote in year one still inform decisions in year five. Scoring models from two sprints ago keep self-correcting today. The ledger never forgets an experiment, or the hypothesis behind it.

No credit card required · Install in minutes · Beliefs start forming day 1