Why Is My Cost Per Add-To-Cart Stable but Cost Per Purchase Doubling: Mid-Funnel Leak Detection

Cost per add-to-cart is stable at $2.20. Cost per purchase used to be $8.70. This month it is $17.50. The top of the funnel is fine. Something between ATC and Purchase is broken.

Mid-funnel leaks are the most expensive problem in D2C and the hardest to spot — because the ad-level metrics look normal. The leak hides between cart and checkout, between checkout and payment, between payment attempt and payment success.

First: Confirm It’s Mid-Funnel, Not Top or Bottom

Rule out the obvious before chasing the subtle.

• Check ATC volume. If ATCs dropped too, the leak is upstream — creative or audience.
• Verify Purchase event is firing correctly. A broken Pixel under-reports purchases — see our CAPI guide.
• Compare ATC to Initiate Checkout to Payment Info to Purchase. The drop tells you the exact stage.

The Root Cause: Cart Abandonment Has Many Names

Between ATC and Purchase there are at least six common leak points:

• Shipping cost shock — free shipping promise broken at checkout.
• Payment failure — declines on the locally dominant method, a payment option silently removed, the gateway timing out.
• Coupon expectation mismatch — ad says 30% off, code doesn’t apply on selected SKU.
• Slow checkout page on mobile data, especially on slower or congested networks.
• Trust gap — no tax invoice offered, no return policy visible, no order tracking explained.
• Cross-device drop-off — ATC on phone, intended purchase on desktop, lost in the gap.

The 4-Step Mid-Funnel Diagnostic

Step 1: Build the Funnel Ratio

Pull ATC, Initiate Checkout, Add Payment Info and Purchase over 30 days and calculate the step-to-step ratios. There is no universal healthy figure to compare against — the signal is the shape. Each step should lose a modest share of the previous one; the step that loses far more than the steps around it is your leak, and comparing this month against your own prior months finds it faster than any published benchmark.

Step 2: Check Shopify/WooCommerce Cart Recovery Reports

Your store’s analytics often tell you more than Meta. Shopify’s Abandoned Checkout report shows the exact stage where users dropped. WooCommerce has similar via plugins.

Step 3: Test Checkout on Slow 3G

Chrome DevTools → Throttling → Slow 3G. Try to complete a purchase. If the page is unusable, you’ve found the leak for the substantial share of buyers on slower connections.

Step 4: Audit Payment Methods

Run a test purchase through every method you offer — card, wallet, bank transfer, buy-now-pay-later, cash on delivery where it applies. Each one that fails or shows extra fees is a confirmed leak.

The Fix: One Leak at a Time

1. Show shipping cost on the PDP, not at checkout. Surprise charges kill conversion.
2. Always offer the locally dominant method plus one low-commitment alternative unless your order value makes the economics impossible.
3. Make the tax invoice and return policy visible on the PDP and the cart page.
4. Strip checkout steps. One-page checkout outperforms multi-step.
5. Set up abandoned cart recovery via email + WhatsApp within 30 minutes.
6. Reference our retargeting funnel playbook for full sequences.

How Bach.ai (by Wittelsbach AI) Catches Mid-Funnel Leaks

Bach.ai cross-references your Meta funnel data with site signals to find exactly where users drop off — and quantifies the $ recoverable from each fix. Mid-funnel leaks are usually fixable within a week once you know which step is losing people. Try Bach.ai on your account at app.wittelsbach.ai.

Frequently Asked Questions

What’s a healthy ATC-to-Purchase rate for D2C?

There is no universal figure, and the spread by category is wide enough that a published average would mislead you — high-AOV and high-consideration products convert a far smaller share of carts than everyday consumables, and both can be healthy. Read it against your own prior months instead, and treat a sudden drop as the signal rather than the level itself.

Could a Meta Pixel issue be causing this gap?

Yes — and it’s the most common false alarm. If your Pixel fires Purchase events on the wrong page or with wrong values, your reported cost per purchase will look inflated even though real purchases are healthy. Always validate Pixel + CAPI events match your Shopify/WooCommerce order count before declaring a mid-funnel leak.

How fast should abandoned cart recovery fire?

The first recovery message should hit within 30 minutes, on whichever channel that buyer actually reads — messaging apps and SMS are opened far more reliably than email, which is the whole reason to use them first. Follow with email at two hours, the messaging channel again at 24 hours with a small incentive, and a final reminder at 72 hours. Measure recovery against your own abandonment baseline rather than an expected rate.

Does cash on delivery really still matter for D2C?

In the markets where it is established — India, much of Southeast Asia, the Gulf — yes, and it scales with order value. At low order values it still carries a large share of orders outside the biggest cities. As order value rises its share falls and return-to-origin rates climb, which is why many premium brands cap COD above a certain basket size. The right call is COD enabled with smart filters (city, AOV, repeat-customer rules) — not blanket on or off.

Is one-page checkout always better than multi-step?

For D2C — almost always yes. Every additional step is another place to lose the buyer, and multi-step checkouts reliably lose more of them than single-page. The exception is high-AOV products where buyers want to feel a deliberate purchase — but even then, the steps should be visible-but-not-blocking, not navigation-required. Test one-page checkout on your top SKUs and measure ATC→Purchase ratio change.

Method and sources

“The leak is between cart and confirmation, not in the ads.”

Source: Where this guide describes platform behaviour, it follows Meta’s published advertising and Marketing API documentation, which changes without notice — verify anything load-bearing against the current version before you act on it. Every threshold the guide asks you to supply is first-party, drawn from your own account exports and commerce ledger, because no external benchmark can stand in for your own margin structure.