CAPI Tracking for E-Commerce Brands

What is CAPI (Conversion API)?

CAPI (Conversion API) allows you to send conversion data directly from your server to Google Ads, bypassing browser-based tracking entirely. Instead of relying on JavaScript tags that can be blocked by ad blockers, restricted by browser privacy settings, or broken by iOS restrictions, you upload conversion events via the Google Ads API after the sale has completed on your server.

CAPI vs Client-Side Tracking

Conversion Types Supported

• Click conversions — Purchases, signups, or leads that happen after an ad click (requires GCLID)
• Call conversions — Phone orders tracked through call tracking systems
• Store sales — In-store purchases attributed to online ad exposure
• Enhanced conversions — Conversions enriched with first-party customer data (email, phone, address) for fallback matching

Why CAPI Matters for E-Commerce

For Google Shopping advertisers, conversion data directly impacts Smart Bidding performance. Missing conversions mean Google's algorithms make decisions on incomplete information — undervaluing your best products and misallocating budget.

The Impact on Your Campaigns

Without accurate conversion data, Smart Bidding makes systematically bad decisions. It sees a product convert at 1.2% when the true rate is 1.8%, and bids accordingly — starving profitable products of budget while overspending on visible-but-mediocre ones. Here is what this looks like in practice:

Why Accurate Data Enables Profit Optimisation

Tracking accuracy is not just a reporting problem — it directly impacts bid optimisation. When 30% of conversions are invisible to Google, Smart Bidding operates blind. Profit-based bidding tools like ProfitClarity include a tracking accuracy adjustment that compensates for known loss rates, but the foundation is always accurate server-side data. CAPI is that foundation.

Step 1: GCLID Capture & Storage

The GCLID (Google Click ID) is the critical link between an ad click and a conversion. When someone clicks your Google ad, a unique identifier is appended to the landing page URL:

https://yourdomain.com/product?gclid=EAIaIQobChMI...

Your job is to capture this parameter, persist it through the shopping journey, and store it alongside each order in your database.

Step 1a: Enable Auto-Tagging

1. In Google Ads, go to Settings → Account Settings
2. Enable Auto-tagging
3. This automatically appends gclid to all ad click destination URLs

Step 1b: Capture GCLID on Your Site

Add this script to every page to capture GCLIDs from the URL and store them for later use. It also captures wbraid and gbraid — Google's iOS web and app attribution parameters:

// GCLID Capture Script — place on every page
(function() {
    const CONFIG = {
        GCLID_KEY: 'grow_gclid',
        EXPIRY_DAYS: 90
    };

    function getGclidFromUrl() {
        const params = new URLSearchParams(window.location.search);
        return params.get('gclid') ||
               params.get('wbraid') ||  // iOS web fallback
               params.get('gbraid');    // iOS app fallback
    }

    function storeGclid(gclid) {
        if (!gclid) return;

        // localStorage (primary storage)
        localStorage.setItem(CONFIG.GCLID_KEY, gclid);
        localStorage.setItem(CONFIG.GCLID_KEY + '_ts', Date.now());

        // Cookie (backup — survives localStorage clear, works cross-subdomain)
        const expires = new Date(Date.now() + CONFIG.EXPIRY_DAYS * 86400000);
        document.cookie = CONFIG.GCLID_KEY + '=' + gclid +
            ';expires=' + expires.toUTCString() +
            ';path=/;SameSite=Lax;Secure';
    }

    const gclid = getGclidFromUrl();
    if (gclid) storeGclid(gclid);
})();

Step 1c: Pass GCLID Through Checkout

Add a hidden field to your checkout form so the GCLID is submitted with the order:

<!-- Add to checkout form -->
<input type="hidden" name="gclid" id="gclid_field">
<script>
    // Populate hidden field with stored GCLID
    document.getElementById('gclid_field').value =
        localStorage.getItem('grow_gclid') || '';
</script>

Step 1d: Store GCLID in Your Database

Add GCLID columns to your orders table. Index the column for efficient lookup during upload jobs:

-- Database schema additions
ALTER TABLE orders ADD COLUMN gclid VARCHAR(255);
ALTER TABLE orders ADD COLUMN gclid_type VARCHAR(20); -- 'gclid', 'wbraid', or 'gbraid'
ALTER TABLE orders ADD COLUMN conversion_uploaded_at TIMESTAMP;
ALTER TABLE orders ADD INDEX idx_gclid (gclid);

Step 2: Create Conversion Actions in Google Ads

Before uploading conversions via the API, you need a dedicated conversion action configured for offline/server-side imports. Do not use the same conversion action as your website tag — keep them separate so you can diagnose issues independently.

Create the Conversion Action

1. In Google Ads, go to Goals → Conversions → Summary
2. Click + New conversion action
3. Select Import → Other data sources or CRMs
4. Choose Track conversions from clicks
5. Configure the settings below

Get the Conversion Action Resource Name

After creating the action, note the Conversion Action ID from the URL or the API. You will need it in this format for upload requests:

customers/{customer_id}/conversionActions/{conversion_action_id}

Step 3: Upload Conversions via the Google Ads API

With GCLIDs captured and a conversion action created, you can now upload conversions. The API endpoint is ConversionUploadService.UploadClickConversions. For new implementations from early 2026 onwards, Google recommends the Data Manager API.

API Requirements

• Developer token — Request via your Google Ads account under Tools & Settings → API Center
• OAuth 2.0 credentials — Service account or web app credentials from Google Cloud Console
• Customer ID — Your Google Ads account ID without dashes (e.g. 1234567890)
• Conversion Action resource name — From the conversion action you created

Conversion Upload Payload

Each conversion requires these fields. The order_id field is critical — it acts as an idempotency key and prevents duplicate uploads if your retry logic fires more than once for the same order:

{
    "gclid": "EAIaIQobChMI...",                            // Required — exact as captured
    "conversion_action": "customers/123/conversionActions/456",
    "conversion_date_time": "2026-03-23 14:30:00+00:00",  // ISO format with timezone
    "conversion_value": 99.99,                             // Order total OR profit value
    "currency_code": "GBP",                                // ISO 4217 currency code
    "order_id": "ORD-12345"                                // Prevents duplicate uploads
}

Upload Frequency and Batching

Step 4: Enhanced Conversions for Better Match Rates

Enhanced conversions add hashed first-party customer data to your conversion uploads. When a GCLID is missing, expired, or never captured, Google can still match the conversion to an ad click by matching the customer's hashed email or phone number against signed-in Google account data.

User Identifiers to Include

• Email address — Most important, highest match rate. Normalise before hashing.
• Phone number — E.164 format (+441234567890)
• First name and last name — Lowercase, no special characters
• Street address, city, state, postal code, country — Match to Google account billing info

Data Normalisation and Hashing

All user data must be normalised and hashed with SHA-256 before leaving your server. Google only receives the hash — never the raw data:

// Node.js — normalise and hash before upload
const crypto = require('crypto');

function hashForEnhancedConversions(value, type = 'email') {
    let normalized = value.trim().toLowerCase();

    if (type === 'email' && normalized.includes('@gmail.com')) {
        // Gmail: remove dots from the username portion
        const [username, domain] = normalized.split('@');
        normalized = username.replace(/./g, '') + '@' + domain;
    }

    if (type === 'phone') {
        // Remove all non-digit characters, then prepend country code if missing
        normalized = normalized.replace(/D/g, '');
        if (!normalized.startsWith('+')) normalized = '+' + normalized;
    }

    return crypto.createHash('sha256').update(normalized).digest('hex');
}

// Example usage
const hashedEmail = hashForEnhancedConversions('John.Doe@Gmail.com', 'email');
const hashedPhone = hashForEnhancedConversions('+44 7700 900123', 'phone');
// Include these hashes in your conversion upload payload

Multi-Layered Matching

Enhanced conversions create a fallback hierarchy so Google can attribute conversions even when your primary signal is missing:

1. Primary match — GCLID directly links the conversion to a specific ad click
2. Fallback match — Hashed email or phone matches to a signed-in Google account
3. Cross-device recovery — Customer clicks on mobile, converts on desktop; email links the journey
4. Cookie deletion recovery — GCLID cookie expired, but email match bridges the gap

Best Practices

Data Quality Checklist

• Unique order IDs — Always include order_id to prevent duplicate conversions on retries
• Accurate values — Use actual order totals or actual profit — never placeholder or average values
• Correct timestamps — Always include timezone offset (e.g. 2026-03-23 14:30:00+00:00)
• GCLID preserved exactly — Store and upload the value character-for-character as received
• Multiple identifiers — Email + phone + address gives the highest match rate
• 90-day retention — Store GCLIDs long enough to cover your full attribution window
• Mark uploaded rows — Track conversion_uploaded_at to avoid re-uploading

Profit-Based Conversion Values

Most brands upload order revenue as the conversion value. Smart Bidding then optimises for revenue — treating a £10 profit sale the same as a £60 profit sale. This is a fundamental misalignment between what you tell Google to optimise for and what actually matters to your business.

MarginStack solves this by computing true per-order profit (after COGS, shipping, payment fees, return costs, and platform fees) and passing that value as the conversion value in the CAPI upload. Google's algorithms then optimise for actual profit, naturally bidding higher on high-margin products and pulling back on low-margin ones.

Layer CAPI with Tag Gateway

CAPI and Tag Gateway are complementary, not alternatives. Together they give you the most complete tracking picture possible:

Monitoring & Diagnostics

CAPI is not a set-and-forget system. Monitor these metrics weekly to catch issues before they affect campaign performance.

Key Metrics to Track

Google Ads Diagnostics

1. Go to Goals → Conversions → Summary in Google Ads
2. Click your CAPI conversion action
3. Open the Diagnostics tab
4. Review: upload status, error breakdown, match rate for enhanced conversions, and recommended improvements

Implementation Roadmap

A practical 8-week timeline for getting CAPI live and fully optimised:

Expected Results After Full Implementation

• 15–25% increase in tracked conversions
• 5–17% improvement in Smart Bidding performance
• 80%+ match rate for GCLID and user identifier combinations
• Recovery of iOS and ad-blocker blocked conversions that were previously invisible
• Accurate ROAS reporting that reflects real business outcomes

Next Steps

CAPI is the tracking foundation that every other optimisation layer depends on. Once your server-side conversion data is flowing accurately, you can layer on profit-based bidding, SKU-level ROAS targets, and automated campaign management.

Related Reading

• Tag Gateway: First-Party Tracking via Cloudflare — Layer server-side script proxying on top of CAPI
• How to Win with Google Shopping in 2026 — The full strategic picture for profitable Shopping campaigns
• COGS & CAC: Understanding Your True Margins — Build the cost data that powers profit-based bidding