Contextual Signals

How to classify IP type, detect timezone mismatches, and categorize referrer sources for age assurance.

Overview

Contextual signals carry a 10% weight in the Signal Fusion engine. Unlike behavioral and input signals, some contextual signals require server-side computation — particularly IP classification, which needs an IP intelligence database.

Fields Reference

All three fields are optional — send whichever ones you can determine. Even a single contextual signal contributes to the assessment.

IP Type

What it measures: The classification of the user's IP address. Education network IPs are strong minor indicators (schools have concentrations of children). Datacenter IPs suggest VPN usage, which can indicate age circumvention. Corporate IPs are mild adult indicators.

Signal impact:

How to Collect

IP classification is a server-side concern. Your backend determines the user's IP address from the incoming request and looks it up against an IP intelligence provider.

// IP classification is server-side — your React app calls
// your backend, which classifies the IP before forwarding to A3.

async function fetchIpType(): Promise<string> {
const res = await fetch('/api/classify-ip');
const { ip_type } = await res.json();
return ip_type; // "residential", "education", "datacenter", "corporate"
}

// Usage in your component:
// const ipType = await fetchIpType();

Recommended IP Providers

Timezone Offset Delta

What it measures: The mismatch between the user's system clock timezone and the timezone expected from their IP-based geolocation. A mismatch greater than 60 minutes can indicate VPN usage or timezone manipulation — a classic bypass technique.

Signal impact: Mismatches > 60 minutes reduce the contextual score by 0.20 and tag the response with inconsistent_timezone_offset.

How to Collect

The client reports its system timezone. Your server compares it against the expected timezone from IP geolocation.

// Client-side: collect the timezone offset and send to your backend
function getTimezoneOffsetMinutes(): number {
return new Date().getTimezoneOffset();
}

// Send to your backend which compares against IP geolocation:
// fetch('/api/timezone-delta', {
//   method: 'POST',
//   body: JSON.stringify({ tz_offset: getTimezoneOffsetMinutes() }),
// });

Edge Cases

• Legitimate travelers may have a timezone mismatch (e.g., using a US phone in Europe). The -0.20 penalty is intentionally moderate — it flags the mismatch but doesn't dominate the score.
• Users who manually set their timezone will show a mismatch even without a VPN. This is uncommon but possible.

Referrer Category

What it measures: The classification of the traffic source that brought the user to your app. Referrals from parental control platforms or minor-oriented social networks are strong minor indicators.

Signal impact:

How to Collect

Read document.referrer on the client and classify the domain against known categories.

import { useMemo } from 'react';

type ReferrerCategory =
| 'direct' | 'social_minor' | 'social_general'
| 'parental_control' | 'search_engine' | 'unknown';

const PATTERNS: [ReferrerCategory, RegExp][] = [
['parental_control', /bark\.us|qustodio|netnanny|familylink|kaspersky.*safe|norton.*family|screentime/i],
['social_minor', /youtubekids|messengerkids|popjam|grom\.social|kinzoo/i],
['social_general', /instagram|tiktok|twitter|x\.com|facebook|snapchat|reddit|linkedin|threads\.net/i],
['search_engine', /google\.|bing\.|duckduckgo|yahoo\.|baidu|yandex/i],
];

function classifyReferrer(referrer: string): ReferrerCategory {
if (!referrer) return 'direct';
try {
  const hostname = new URL(referrer).hostname;
  for (const [category, pattern] of PATTERNS) {
    if (pattern.test(hostname)) return category;
  }
} catch {}
return 'unknown';
}

export function useReferrerCategory(): ReferrerCategory {
return useMemo(() => classifyReferrer(document.referrer), []);
}

// Usage: const referrer_category = useReferrerCategory();

Putting It All Together

# Complete contextual_signals example payload:
curl -X POST https://api.a3api.io/v1/assurance/assess-age \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
  "os_signal": "18-plus",
  "user_country_code": "US",
  "contextual_signals": {
    "ip_type": "residential",
    "timezone_offset_delta_minutes": 0,
    "referrer_category": "search_engine"
  }
}'

Next Steps

• Account Longevity — the simplest signal to implement
• Behavioral Metrics — the highest-impact category (43%)
• Signal Collection Overview — integration paths and coverage guidance