Skip to content
Tolinku
Sign In Start Free
Tolinku
Sign In Start Free
App Growth · · 6 min read

A/B Testing Email Deep Links

By Tolinku Staff
|
Tolinku app growth strategies dashboard screenshot for growth blog posts

Email remains one of the highest-converting channels for driving app engagement. When you add deep links to emails, users can tap straight into specific app screens instead of landing on a generic homepage. A/B testing these deep links (the destinations, CTA text, link placement, and fallback behavior) lets you find the combination that maximizes opens, taps, and in-app conversions.

For Universal Links in email, see Universal Links in Email Campaigns: Setup and Pitfalls. For e-commerce email deep links, see E-Commerce Email Deep Links That Convert.

Tolinku A/B testing dashboard for smart banners The A/B tests list page showing test names, status, types, and variant counts.

What to Test

Test Variables by Impact

Variable Impact Difficulty Notes
Deep link destination High (15-30% lift) Medium Where in the app does the user land?
CTA text High (10-25% lift) Easy "Open in App" vs. "View Your Order"
Link placement Medium (5-15% lift) Easy Above fold vs. inline vs. footer
App vs. web destination High (10-40% lift) Medium Route to app or mobile web?
Number of deep links Medium (5-10% lift) Easy One primary CTA vs. multiple links
Fallback page Medium (10-20% lift) Medium Where non-app users go
Subject line (with app mention) High (10-20% open lift) Easy "Open in app" vs. no mention

App Screen Variants

The same email can route users to different in-app screens:

const emailDestinationTest = {
  id: 'email_destination_v2',
  emailType: 'order_shipped',
  variants: [
    {
      id: 'order_detail',
      weight: 50,
      deepLink: '/orders/:orderId',
      ctaText: 'Track Your Order',
      fallbackUrl: 'https://example.com/orders/:orderId',
    },
    {
      id: 'tracking_map',
      weight: 50,
      deepLink: '/orders/:orderId/tracking',
      ctaText: 'See Live Tracking',
      fallbackUrl: 'https://example.com/orders/:orderId/tracking',
    },
  ],
  primaryMetric: 'email_cta_tap',
  secondaryMetrics: ['app_open', 'order_detail_view', 'repeat_purchase_7d'],
};

App vs. Web Destination

Test whether routing to the app or mobile web produces better outcomes:

const appVsWebTest = {
  id: 'app_vs_web_v1',
  emailType: 'weekly_digest',
  variants: [
    {
      id: 'app_deep_link',
      weight: 50,
      link: 'https://example.com/content/:id', // Universal Link, opens app if installed
      ctaText: 'Read in App',
    },
    {
      id: 'web_link',
      weight: 50,
      link: 'https://example.com/content/:id?no_app_redirect=true',
      ctaText: 'Read Now',
    },
  ],
  primaryMetric: 'content_read_complete',
  secondaryMetrics: ['session_duration', 'next_action'],
};

Typical results:

Scenario App Link Wins Web Link Wins
User has app installed Always Never
User doesn't have app Never (friction) Always
Transaction email (order, receipt) Usually Rarely
Content/newsletter Sometimes Often
Re-engagement campaign Usually Rarely

The key insight: if most of your email audience has the app installed, deep links win. If many recipients don't have the app, web links avoid the friction of an app store redirect.

CTA Text Testing

Email CTA Variants

const emailCtaTest = {
  id: 'email_cta_v3',
  variants: [
    { id: 'generic', ctaText: 'View Details', weight: 20 },
    { id: 'app_specific', ctaText: 'Open in App', weight: 20 },
    { id: 'action_specific', ctaText: 'Track Your Package', weight: 20 },
    { id: 'benefit', ctaText: 'See Delivery Updates', weight: 20 },
    { id: 'urgency', ctaText: 'Check Status Now', weight: 20 },
  ],
  primaryMetric: 'email_cta_tap',
};

Results by email type:

Email Type Best CTA Style Example
Order confirmation Action-specific "View Your Order"
Shipping notification Benefit "Track Your Package"
Promotional Benefit or urgency "Claim Your 20% Off"
Re-engagement Direct "Come Back and See What's New"
Weekly digest Content preview "Read This Week's Top Stories"
Account alert Urgency "Review Your Account Now"

Subject Line Testing

Subject lines that mention the app can increase open rates:

const subjectLineTest = {
  id: 'subject_app_mention_v1',
  variants: [
    {
      id: 'no_mention',
      subject: 'Your order has shipped',
      weight: 50,
    },
    {
      id: 'app_mention',
      subject: 'Your order has shipped - track it live in the app',
      weight: 50,
    },
  ],
  primaryMetric: 'email_open_rate',
  secondaryMetrics: ['email_cta_tap', 'app_open'],
};

Mentioning the app in the subject line typically increases open rates by 5-10% for users who have the app installed, but has no effect (or slight negative effect) for users without the app.

const linkPlacementTest = {
  id: 'link_placement_v1',
  variants: [
    {
      id: 'single_cta',
      weight: 50,
      layout: {
        heroLink: true,    // One big CTA button
        inlineLinks: false,
        footerLink: false,
      },
    },
    {
      id: 'multiple_links',
      weight: 50,
      layout: {
        heroLink: true,     // Primary CTA
        inlineLinks: true,  // Links within content
        footerLink: true,   // Repeat CTA at bottom
      },
    },
  ],
  primaryMetric: 'any_link_tap',
  secondaryMetrics: ['hero_tap', 'inline_tap', 'footer_tap'],
};

Typical results:

  • Single CTA: Higher tap rate on the one button (focused attention)
  • Multiple links: Higher total taps (users who scroll past the hero CTA tap the footer one)
  • For short emails (< 300px height): single CTA wins
  • For long emails (> 500px height): multiple links win

CTA Button vs. Text Link

const linkStyleTest = {
  id: 'link_style_v1',
  variants: [
    {
      id: 'button',
      weight: 50,
      style: 'button', // Styled HTML button
    },
    {
      id: 'text_link',
      weight: 50,
      style: 'text', // Underlined text link
    },
  ],
  primaryMetric: 'email_cta_tap',
};

Buttons outperform text links by 20-30% in most cases. The exception: emails that are meant to feel personal (like a note from a founder) perform better with text links because buttons feel like marketing.

Implementation

Building Testable Email Templates

async function buildEmail(userId, emailType, data) {
  const experiment = getActiveExperiment('email', emailType);

  if (!experiment) {
    return buildDefaultEmail(emailType, data);
  }

  const variant = assignVariant(userId, experiment.id);

  // Generate trackable deep link
  const deepLink = await generateDeepLink(variant.deepLink || data.defaultDeepLink, {
    experimentId: experiment.id,
    variantId: variant.id,
    userId,
    source: 'email',
    campaign: emailType,
  });

  // Generate fallback URL for users without the app
  const fallbackUrl = variant.fallbackUrl || data.defaultFallbackUrl;

  return {
    subject: variant.subject || data.defaultSubject,
    html: renderEmailTemplate(emailType, {
      ...data,
      ctaText: variant.ctaText || data.defaultCtaText,
      ctaUrl: deepLink,
      fallbackUrl,
      layout: variant.layout || 'default',
    }),
    tracking: {
      experimentId: experiment.id,
      variantId: variant.id,
    },
  };
}

Universal Links have specific requirements in email clients. Some email clients (Gmail, Outlook) wrap links in their own redirect, which breaks Universal Link behavior.

function buildEmailDeepLink(path, params) {
  // Use a web URL that is configured as a Universal Link
  const baseUrl = 'https://links.example.com';
  const url = new URL(path, baseUrl);

  // Add tracking parameters
  Object.entries(params).forEach(([key, value]) => {
    url.searchParams.set(key, value);
  });

  // Add email-specific parameter for fallback handling
  url.searchParams.set('source', 'email');

  return url.toString();
}

For detailed Universal Link email setup, see the Apple developer documentation on Universal Links.

// Email sent
analytics.track('email_sent', {
  experimentId,
  variantId,
  emailType: 'order_shipped',
  userId,
});

// Email opened
analytics.track('email_opened', {
  experimentId,
  variantId,
  emailType: 'order_shipped',
  userId,
  openedAt: Date.now(),
});

// CTA tapped
analytics.track('email_cta_tapped', {
  experimentId,
  variantId,
  ctaPosition: 'hero', // or 'inline', 'footer'
  destination: 'app', // or 'web', 'app_store'
  userId,
});

// App opened from email
analytics.track('app_opened_from_email', {
  experimentId,
  variantId,
  emailType: 'order_shipped',
  deepLinkPath: '/orders/123',
  userId,
});

Fallback Testing

What Happens When the App Isn't Installed

const fallbackTest = {
  id: 'email_fallback_v1',
  variants: [
    {
      id: 'app_store',
      weight: 33,
      fallback: 'app_store', // Send to app store
    },
    {
      id: 'mobile_web',
      weight: 33,
      fallback: 'web_equivalent', // Same content on mobile web
    },
    {
      id: 'landing_page',
      weight: 34,
      fallback: 'install_landing', // Custom "install the app" page
    },
  ],
  primaryMetric: 'goal_completion', // Did the user complete the intended action?
};
Fallback Install Rate Goal Completion Best For
App store 15-25% Low (delayed) Growing app installs
Mobile web 0% High (immediate) Transactional emails
Landing page 20-35% Medium Campaigns with install incentive

Segmentation

Test Different Variants for Different Segments

async function getEmailVariant(userId, emailType) {
  const userProfile = await getUserProfile(userId);

  // Different experiments for different segments
  if (userProfile.hasAppInstalled) {
    return getVariant(userId, `${emailType}_app_users`);
  }

  if (userProfile.previousAppUser) {
    return getVariant(userId, `${emailType}_lapsed_users`);
  }

  return getVariant(userId, `${emailType}_web_only_users`);
}

Segment-Specific Results

The winning variant often differs by segment:

Segment Best CTA Best Destination Best Fallback
Active app users "Open in App" Direct deep link N/A
Lapsed app users "See What's New" Home screen Mobile web
Web-only users "Try the App" Landing page Mobile web
New subscribers "Get Started" Onboarding flow App store

Best Practices

  1. Segment by app install status: The optimal email is completely different for app users vs. web-only users.
  2. Test one variable at a time: Change the CTA text or the destination, not both.
  3. Run for at least 2 send cycles: If you send weekly, run the test for at least 2 weeks.
  4. Track through to in-app action: Email open and tap rates are vanity metrics. Track the downstream conversion.
  5. Account for email client behavior: Test on Gmail, Apple Mail, Outlook, and Yahoo. Each handles deep links differently.
  6. Use Universal Links, not custom schemes: Custom URL schemes (myapp://) don't work in most email clients. Use Universal Links / App Links (HTTPS URLs).

For A/B testing features, see Tolinku A/B testing. For deep link A/B testing setup, see the A/B testing docs.

A/B Testing conversions Deep Linking Email Marketing experimentation mobile-development optimization User Acquisition

Get deep linking tips in your inbox

One email per week. No spam.

In this article

Ready to add deep linking to your app?

Set up Universal Links, App Links, deferred deep linking, and analytics in minutes. Free to start.

Start Free View Documentation

Related articles