Audience Demographics → Persona Creation

Pull GA4 demographic intersections (age, gender, country, device) with conversion metrics and create data-backed Mavera Custom Personas for your actual converting audience

Scenario

Your GA4 property has thousands of converting visitors segmented by age, gender, country, and device — but your personas are based on gut feel. You pull a RunReport crossing userAgeBracket, userGender, country, and deviceCategory with conversion metrics, identify the top-performing demographic intersections, and create a Mavera Custom Persona for each. The result is a persona library calibrated to your actual converting audience.

Architecture

Code

Python

import os, requests, time
from google.analytics.data_v1beta import BetaAnalyticsDataClient
from google.analytics.data_v1beta.types import (
    RunReportRequest, Dimension, Metric, DateRange, OrderBy,
)

PROPERTY_ID = os.environ["GA4_PROPERTY_ID"]
MV = os.environ["MAVERA_API_KEY"]
MB = "https://app.mavera.io/api/v1"
MH = {"Authorization": f"Bearer {MV}", "Content-Type": "application/json"}

client = BetaAnalyticsDataClient()

report = client.run_report(RunReportRequest(
    property=f"properties/{PROPERTY_ID}",
    dimensions=[
        Dimension(name="userAgeBracket"),
        Dimension(name="userGender"),
        Dimension(name="country"),
        Dimension(name="deviceCategory"),
    ],
    metrics=[
        Metric(name="totalUsers"),
        Metric(name="conversions"),
        Metric(name="sessions"),
        Metric(name="engagementRate"),
        Metric(name="averageSessionDuration"),
    ],
    date_ranges=[DateRange(start_date="30daysAgo", end_date="today")],
    order_bys=[OrderBy(metric=OrderBy.MetricOrderBy(metric_name="conversions"), desc=True)],
    limit=200,
))

intersections = []
for row in report.rows:
    age = row.dimension_values[0].value
    gender = row.dimension_values[1].value
    country = row.dimension_values[2].value
    device = row.dimension_values[3].value
    users = int(row.metric_values[0].value)
    conversions = int(row.metric_values[1].value)
    sessions = int(row.metric_values[2].value)
    engagement = float(row.metric_values[3].value)
    avg_duration = float(row.metric_values[4].value)

    if conversions < 5 or age == "(not set)" or gender == "(not set)":
        continue

    intersections.append({
        "age": age, "gender": gender, "country": country, "device": device,
        "users": users, "conversions": conversions, "sessions": sessions,
        "engagement": engagement, "avg_duration": avg_duration,
        "conv_rate": conversions / max(sessions, 1),
    })

intersections.sort(key=lambda x: x["conversions"], reverse=True)
top = intersections[:10]

total_conv = sum(i["conversions"] for i in intersections) or 1
print(f"Total intersections: {len(intersections)} | Creating personas for top {len(top)}")

created = []
for seg in top:
    conv_share = seg["conversions"] / total_conv
    name = f"GA4: {seg['gender'].title()} {seg['age']} — {seg['country']} ({seg['device']})"
    desc = (
        f"Data-backed persona from GA4 (last 30 days). "
        f"Age: {seg['age']}, Gender: {seg['gender']}, Country: {seg['country']}, Device: {seg['device']}. "
        f"Conversions: {seg['conversions']} ({conv_share:.0%} of total). "
        f"Engagement rate: {seg['engagement']:.1%}. Avg session: {seg['avg_duration']:.0f}s. "
        f"Conv rate: {seg['conv_rate']:.2%}."
    )

    r = requests.post(f"{MB}/personas", headers=MH, json={
        "name": name,
        "description": desc,
        "demographic": {
            "age_range": seg["age"],
            "gender": seg["gender"],
            "countries": [seg["country"]],
        },
        "psychographic": {
            "device_preference": seg["device"],
            "engagement_level": "high" if seg["engagement"] > 0.6 else "medium",
            "conversion_share": f"{conv_share:.1%}",
        },
    })
    r.raise_for_status()
    persona = r.json()
    created.append({"name": name, "id": persona["id"], "conversions": seg["conversions"]})
    print(f"  {name} → {persona['id']} ({seg['conversions']} conv, {conv_share:.0%})")
    time.sleep(0.3)

print(f"\nCreated {len(created)} demographic personas from GA4 data")

Example Output

{
  "total_intersections": 142,
  "created": 10,
  "personas": [
    { "name": "GA4: male 25-34 — United States (desktop)", "id": "per_ga4_01", "conversions": 312, "share": "18%" },
    { "name": "GA4: female 25-34 — United States (mobile)", "id": "per_ga4_02", "conversions": 245, "share": "14%" },
    { "name": "GA4: male 35-44 — United Kingdom (desktop)", "id": "per_ga4_03", "conversions": 189, "share": "11%" },
    { "name": "GA4: female 35-44 — United States (desktop)", "id": "per_ga4_04", "conversions": 156, "share": "9%" },
    { "name": "GA4: male 25-34 — Canada (mobile)", "id": "per_ga4_05", "conversions": 98, "share": "6%" }
  ]
}

Error Handling

(not set) dimensions

GA4 can’t classify all users by age/gender — especially those without Google sign-in. The code filters these out. Expect 20–40% of traffic to be unclassified.

Thresholding / data redaction

GA4 applies thresholding to protect user privacy. Small segments may be hidden. Widen your date range or reduce dimension cardinality to get more data.

Quota exhaustion (429)

50,000 requests/day is generous, but bursts of 10+ concurrent requests hit the per-property limit. Add 300ms delays between calls and retry on 429 with exponential backoff.

Service account permissions

If you get 403 PERMISSION_DENIED, verify the service account email is added as a Viewer (or higher) in GA4 → Admin → Property Access Management.

---

What’s Next

GA4 Integration

Back to GA4 integration overview

Interest Category → Content Strategy

Build a content plan from audience interests

Personas API

Full reference for POST /api/v1/personas

Mave Agent

Full reference for POST /api/v1/mave/chat