You have five abandonment problems. A cart flow solves one.
A standard abandoned-cart flow triggers on checkout started, waits for a sync, and emails the products left behind. By then the shopper has survived four earlier exits that nothing tracked: a search that returned junk, products viewed and left, a category browsed and abandoned, a wishlist never bought.
- Baymard: 70.22% of carts are abandoned, across 50 studies
- SearchSpring: 68% of shoppers leave a site after a poor search
- Multichannel recovery brings back 20% to 30% of baskets, email alone 10% to 15% (InternetRetailing)
Cart abandonment gets the attention because it is easy to measure. Baymard's meta-analysis of 50 studies puts the rate at 70.22%, stable for a decade. But a cart is the last place a shopper leaves, not the first, and the earlier exits are bigger and mostly invisible.
Klaviyo's flow lives at the cart. Its own help center says some integrations sync once an hour and recommends a delay of at least 1 hour and 15 minutes before the first email. It reaches the shoppers who gave an email, by email and SMS, with the products they already chose. Everything before the cart, and everyone it cannot see, is left to chance.
There are five failure points: search, browse, category, wishlist and cart, and most stacks recover only from the last one. Two channels is not multichannel: InternetRetailing finds push, email and SMS together recover roughly double what email alone does. And the recovery message should add what the shopper will buy next, not just mirror the cart.
The five failure points
1. Search abandonment
A shopper searches, the results are wrong or empty, they leave. SearchSpring estimates that up to 68% of shoppers abandon a site after a poor search experience. It is the biggest exit by revenue and the one with the least recovery. An email tool does not know what the shopper searched for, so it cannot follow up. In one kids' fashion store, the 4.9% of visitors who searched produced 27.5% of revenue once search suggested the right products from the first keystroke (case study).
2. Browse abandonment
Products viewed, nothing added. Intent is there, below the cart threshold. Klaviyo offers a browse abandonment flow, built separately, that shows the viewed products. It cannot say why the shopper did not add them: wrong size, wrong price, wrong color. Recommendations on the page that answer those questions prevent the exit; an email four hours later cleans up after it.
3. Category abandonment
A category page browsed, products compared, no click into a product. The aisle failed, not the product. Only on-site personalization can act here, in the moment, with products ranked for the person looking.
4. Wishlist abandonment
Saved, never bought. The intent is explicit and the timing is unknown. The right recovery waits for the predicted purchase window and nudges on the channel the shopper answers.
5. Cart abandonment
Added, checkout started, gone. This is where the standard flow lives, and even here it covers only the shoppers it can track, on two channels, with a mirror of the cart.
Four gaps in the standard flow
The tracking gap
A flow can only trigger on events it receives, and browser scripts lose events to ad blockers, Safari's cookie limits, consent banners and cross-device journeys. The shopper who adds to cart on their phone and returns on a laptop looks like two strangers. Server-side tracking records the event where it happens (guide).
The channel gap
Email and SMS are two channels. InternetRetailing reports that a sequence of push, then email, then SMS recovers 20% to 30% of abandoned baskets, roughly double email alone, and that push produced 21% of push-attributed orders from 3% of sends. On-site personalization prevents the exit before any message is needed. The standard flow has neither.
The product gap
The recovery email shows what the shopper already chose. The agent adds what they are likely to buy next: a complementary product, an alternative at a different price if the model says they left on price. One womenswear brand moved its abandoned-browse flow to the agent and added recommendations: 3 times the shoppers reached and 9.4 times the revenue on the same flow (case study).
The timing gap
InternetRetailing identifies a window of 5 to 15 minutes after abandonment that produces more than 60% of recovered revenue without any discount. Klaviyo's help center notes that some integrations sync once per hour and recommends a delay of at least 1 hour and 15 minutes for the first email. By then the no-discount window has closed, and the email needs a coupon to do what a push at five minutes would have done for free.
What recovery looks like with one agent
Search that understands intent and corrects spelling. Recommendations on every page that adapt as the shopper browses, so most exits never happen. A browse-abandonment journey triggered by server-side events, with products picked by prediction. Wishlist nudges timed to the predicted purchase window. And for the cart: push at five minutes, an email at one hour with the cart plus what goes with it, SMS later if nothing moved, all from one memory of the shopper.
The economics change with the channel and the timing. In one online grocery, the cart email leads to an order 5.7% of the time against 0.19% for a newsletter, opens at 44% against 18%, and under 1% of emails produce 21% of email revenue (use case). For Shopify brands on Klaviyo, the agent runs next to the existing flows: start with one flow, reach every abandoner, and let each email pick its own products (use case).
Five questions for your setup
What share of add-to-cart events reaches your email tool? How many of the five exits do you recover from? How many channels does recovery use? Does the recovery email contain anything the shopper did not add? Does a small cart from your best customer get the same flow as a small cart from a first-time bargain hunter? Three or more wrong answers, and the flow is not the problem. The architecture is.
Every exit, every channel, one memory
Reach every abandoner on Shopify and Klaviyo
Start with one flow next to what you already run. Three times the shoppers reached on the same trigger, and each email picks its own products.
Read the use case →CompareReleva and Klaviyo: keep your CRM, add a sales agent
What each one does, from Klaviyo's own pages and ours, and how they run side by side.
Read the comparison →Questions about abandonment recovery
What is the difference between cart and browse abandonment?
Cart abandonment is a shopper adding products and leaving before paying; Baymard puts the average at 70.22%. Browse abandonment happens earlier, when products are viewed and nothing is added. It has lower intent and far higher volume, and most stacks do nothing about it.
What is search abandonment?
A shopper searches your store and leaves without clicking a product. SearchSpring estimates that up to 68% of shoppers leave a site after a poor search. An email tool cannot follow up because it does not know what was searched.
Which channels recover the most baskets?
InternetRetailing reports that a sequence of push, then email, then SMS recovers 20% to 30% of abandoned baskets, roughly double email alone, and that the 5 to 15 minutes after abandonment produce more than 60% of recovered revenue without a discount.
Can the agent run next to Klaviyo?
Yes. Klaviyo stays your CRM. The agent runs one flow next to it, reaches every abandoner including visitors who never logged in, and lets each email pick its own products. Start with one flow and compare in a matched month.
See which of last week's visitors you missed
Thirty minutes, with your store and ad accounts open. Then 30 days free. A holdout group decides: if the agent doesn't add orders in 30 days, you don't pay.
Book 30 minutes
You'll talk to one of us.