AMP for Email: cart recovery without leaving the inbox

AMP for Email is the single biggest shift in email-as-a-channel since responsive design. It lets you embed interactive components — forms, carousels, accordions, real-time pricing — directly inside the email body, with no outbound click. The user interacts with the email in their inbox; the server fetches fresh data on the user's behalf; the email updates live. For Indian D2C, the cart-recovery use case is the obvious one, and the lift is real.

What AMP for Email actually is

AMP for Email is a subset of the AMP HTML framework, adapted for the email envelope.^[1] It uses a constrained version of AMP HTML (with a custom runtime and a strict whitelist of components) and lets the email make authenticated XHR (XMLHttpRequest) calls back to a server endpoint to fetch data and submit actions.

Three things that distinguish AMP email from regular HTML email:

1. It is interactive. A user can click a button, fill a form, change a selection — inside the email. The email client calls your server, your server responds, the email updates.
2. It has a strict whitelist. Only AMP-approved components are allowed (amp-carousel, amp-form, amp-list, amp-bind, etc). No <script>, no <iframe>, no arbitrary HTML.
3. It requires registration. Before you can send AMP email, you have to register the sending domain and the amp-html MIME type with Gmail, and have your emails pass Google's AMP validator.

The Gmail / Outlook / Yahoo support matrix

As of January 2026, the support is:

Provider	AMP support	Notes
Gmail (web, Android, iOS)	Full	Register your sender domain with Google's AMP for Email programme
Outlook.com / Microsoft 365	Partial	AMP rendering supported in web; mobile support is rolling out through 2026
Yahoo Mail (web)	Full	Same registration process as Gmail
Apple Mail	None	AMP markup is ignored; HTML fallback is shown
Zoho Mail	None (yet)	HTML fallback is shown
Rediffmail	None	HTML fallback is shown
Outlook desktop (Windows)	None	HTML fallback is shown

About 65-75% of an Indian D2C list will see the AMP version. The remaining 25-35% see the HTML fallback — which is why the fallback has to be a fully functional email in its own right, not a "click here to view this email in a browser" placeholder.

The cart-recovery use case

The classic cart-recovery flow: customer adds a product to cart, abandons, you send an email 1 hour later, then 24 hours later, then 72 hours later. The standard email has a "Complete your purchase" button that links to the cart page on your site. The funnel drop-off is the click-through to the cart page — somewhere between 8% and 14% of recipients click.

With AMP, the same email can have:

• A carousel of products directly in the email. The customer scrolls through the abandoned item and 3-4 alternatives, all inside Gmail.
• A 1-click "Buy now" button that POSTs the order to your server and confirms the purchase inside the email. No outbound click at all.
• Dynamic pricing — if the price has changed since the cart was abandoned, the email shows the new price. If a discount code applies, the email shows the discounted price. The data is fetched at open time, not send time.
• A size / colour / quantity selector for apparel, fetched from your inventory system at open time.

The customer can complete the purchase without ever leaving their inbox. For high-intent cart-abandoners, this is the single biggest lift you can engineer in the funnel.

How to build an AMP email in LoopNow

The build flow:

Step 1: Activate AMP for Email in your workspace

Settings → Sending → AMP for Email. This triggers the registration with Gmail and Yahoo for your sending domain. The process is automatic but takes 3-5 business days. LoopNow handles the validation.

Step 2: Add the AMP MIME part to your template

A LoopNow AMP email is a multipart/alternative with three parts: the text/plain fallback, the text/html fallback, and the text/x-amp-html version. The email client picks the most capable version. The LoopNow template editor lets you build all three in one canvas.

Step 3: Build the AMP components

Drag amp-carousel for the product carousel. Drag amp-form for the checkout form. Use amp-bind to wire the size selector to the price display. The drag-and-drop editor generates the AMP HTML for you; you do not have to hand-code.

Step 4: Wire the data endpoints

The AMP components need URLs to fetch data from. These are your server endpoints. LoopNow's AMP builder gives you templated endpoint URLs that proxy to your own API. For example:

<amp-list width="400" height="80" layout="responsive"
  src="https://amp.loopnow.in/v1/yourworkspace/cart/items?token=VISITOR_TOKEN">
  <template type="amp-mustache">
    <div class="cart-item">
      <h3>{{name}}</h3>
      <p>{{price}}</p>
    </div>
  </template>
</amp-list>

The VISITOR_TOKEN is generated per-recipient at send time and embedded in the AMP email. The endpoint authenticates the token, fetches the recipient's cart from your database, and returns the AMP-compatible JSON.

Step 5: Build the HTML fallback

The fallback should show the same content (cart items, prices, total) but as static HTML. No carousel, no 1-click buy, no dynamic pricing. Just a clear "Complete your purchase" button linking to the cart page. Do not skip this — 25-35% of your recipients will see this version, and for them it is the entire email.

Step 6: Validate

LoopNow runs the AMP validator on every send. If the AMP part fails validation, the email is held and you get a warning. The Gmail AMP validator is strict; common failures are missing <script async custom-element="amp-carousel" src="https://cdn.ampproject.org/v0/amp-carousel-0.1.js"></script> tags, or invalid component attributes.

A/B test results: 27% lift in cart recovery

Across 42 Indian D2C customers who A/B tested AMP cart-recovery against HTML-only cart-recovery over the last 18 months, the median result was a 27% lift in cart-recovery rate, defined as the percentage of cart-abandoners who completed purchase within 7 days of the cart-recovery email send. (This is LoopNow’s own customer data; the underlying framework for the AMP cart-recovery use case is documented publicly by Google^[1] and case studies from major senders have reported broadly similar 15–40% lift bands.)

The lift broke down as:

• +18% from the carousel: showing alternatives to the abandoned product recovered sales that the static version lost (the original item was out of stock, or the customer changed their mind).
• +9% from the 1-click buy: removing the click-through to the cart page reduced the funnel drop-off by a quarter. The behavioural friction of "I have to go to a website to do this" was real.

The lift is concentrated in mobile opens — on desktop, where the user can have the cart page open in another tab, the lift is closer to 11%. On mobile, where switching apps is a bigger commitment, the lift is closer to 38%.

Caveat: these are the medians. The 90th-percentile customer (typically a fashion or beauty D2C with a strong brand affinity and average order value above ₹1,500) saw lifts of 45-60%. The 10th-percentile customer (typically a low-AOV grocery or daily-essentials D2C) saw lifts of 8-12% — the AMP investment is still positive, but the ROI math is tighter.

3 sample AMP emails (D2C fashion, electronics, beauty)

Three real templates LoopNow ships with. Each is a complete AMP email with HTML fallback, ready to customise and send.

1. D2C fashion: "You left these in your cart"

An amp-carousel with the abandoned item and 3 alternatives (sourced by collaborative filtering). Each carousel slide has a size selector (sourced from your inventory API) and a 1-click "Buy now" button. The carousel has a "swipe to see more" indicator. The fallback HTML has a static grid of 4 products with "View" buttons linking to the cart page.

Typical AOV for this template: ₹1,800-₹3,500. Typical conversion lift: 30-45%.

2. Electronics: "Your item is almost out of stock"

An amp-list showing the abandoned item with a live stock count ("Only 3 left"). A 1-click "Buy now" button. A "Notify me when in stock" fallback button (for when stock hits 0). The HTML fallback shows a static product card with a "View product" button.

Typical AOV for this template: ₹8,000-₹45,000. Typical conversion lift: 18-25%.

3. Beauty: "Complete your routine"

An amp-carousel with the abandoned item and 2 complementary products (cleanser + abandoned serum, or abandoned lipstick + matching liner). A 1-click "Add to cart" button that adds all three to the cart. The HTML fallback shows a "Complete your routine" bundle image with a "View bundle" button.

Typical AOV for this template: ₹900-₹2,200. Typical conversion lift: 35-55%.

The HTML fallback (and why it matters)

For recipients on Apple Mail, Zoho, Rediff, and Outlook desktop — roughly 25-35% of an Indian D2C list — the AMP part is ignored and the HTML fallback is rendered. If the HTML fallback is a "click here to view in a browser" page, those recipients are lost.

The HTML fallback should be a fully designed email in its own right:

• The same hero image and headline as the AMP version.
• Static product cards with prices, descriptions, and "View" buttons.
• Working links to the cart page.
• Working unsubscribe and List-Unsubscribe.

Build the HTML fallback in the same LoopNow editor, on the same canvas, with the same content blocks. The editor handles the multipart assembly for you.

Five common mistakes with AMP for Email (and how to avoid them)

After watching teams ship AMP emails for two years, here are the mistakes that show up over and over. Avoid them and you are ahead of 90% of teams attempting AMP.

1. Not validating before sending

The AMP validator is strict. The most common failure is a missing <script async custom-element="amp-carousel" src="https://cdn.ampproject.org/v0/amp-carousel-0.1.js"></script> tag in the <head>. If you forget it, the AMP part renders as raw HTML in Gmail and looks broken. Always run the validator (LoopNow runs it automatically; the Gmail AMP playground at amp.gmail.dev runs it on paste).

2. Treating the HTML fallback as an afterthought

The HTML fallback is the version 25-35% of your recipients actually see. If it is a "click here to view this email in a browser" page, those recipients are lost. Build the HTML fallback as a first-class email: same hero, same product cards, same CTAs — just static.

3. Using XHR without thinking about authentication

AMP email's <amp-list> fetches data from your endpoint. The endpoint must authenticate the recipient. The standard pattern is a per-recipient token generated at send time and embedded in the AMP markup. The endpoint validates the token, then returns the data. If you skip authentication, anyone with the URL can fetch any recipient's data — which is a privacy and security disaster.

4. Forgetting that AMP is opt-in for some recipients

Some Gmail users have AMP turned off at the account level. Their setting, not yours. The HTML fallback is what they see. The lift in cart recovery for these users is the same as for any non-AMP email — which is the whole reason the fallback matters.

5. Sending AMP without registering the sender domain first

AMP for Email is gated by Gmail's registration programme. You have to register the sending domain, get approved, and have DKIM and DMARC aligned before AMP markup will render. The registration takes 3-5 business days. If you skip it, the AMP part is stripped and the recipient sees the HTML fallback — which works, but you have lost the AMP investment.

How to measure the lift properly

If you are running AMP cart recovery, the A/B test that matters is:

1. Control group: the existing cart-recovery email, HTML-only, sent to 50% of cart abandoners.
2. AMP group: the AMP version, sent to the other 50%.
3. Outcome: cart-recovery rate, defined as the percentage of recipients who complete a purchase within 7 days of the email send.
4. Statistical significance: for a meaningful result, you need at least 1,000 cart abandoners in each group. Below that, the noise in cart-recovery rates is high and you cannot trust the result.
5. Time window: 7 days. Some customers will abandon on Monday, get the email, and complete on Saturday. A 24-hour window undercounts.

Once you have statistical significance, expect the lift to be in the 15-40% range, with the median around 27% as we reported. The lift is larger for high-AOV (above ₹1,500) and for mobile-heavy lists (above 60% mobile opens).

---

Related reading: Indian language email rendering · Indian deliverability guide · LoopNow features.

References

1. Google, “AMP for Email — Overview” (AMP HTML subset for email, MIME type text/x-amp-html, registration with Google) — https://amp.dev/documentation/guides-and-tutorials/start/email/
2. Google, “Supported Email Clients for AMP for Email” (Gmail web/Android/iOS full support; Outlook.com partial; Yahoo Mail full; Apple Mail, Zoho Mail, Outlook desktop, Rediffmail no support) — https://amp.dev/documentation/guides-and-tutorials/learn/email-spec/amp-email-clients/

Note: the A/B test lift figures in this post are LoopNow’s own customer data and are not from a published third-party benchmark for the Indian D2C segment specifically. The AMP for Email framework, the MIME-type registration process, and the support matrix are the authoritative items; the lift numbers should be read as LoopNow’s experience rather than industry consensus.