Back to Explore
GuideVibeFix editorialUpdated Nov 19, 2026

They Added to Cart and Left: Abandoned Cart Recovery for Your Vibe Project, in Three Messages

7,200 add-to-carts, only 622 orders paid: people did not stop wanting the bag, they turned back at the doorway. Sort out who really left and who just forgot, then bring them back in order — email at 1 hour, SMS at 24 hours, a discount reminder at 72 hours, discount last and only for some. The walk-through project, people, and numbers in this guide are fictional.

A real photo: a fingertip about to tap the red Checkout button on a tablet showing 1 item added to cart with a $99.99 subtotal.

They added to cart and left: the cart is not the finish line, it is the doorway

A-Yuan runs a small shop selling handmade canvas bags. He calls it Slow Bag. He cuts the patterns and works the sewing machine himself, and each bag sells for 268 yuan. Three months in, word of mouth is decent: buyers come back for a second bag and send him photos of the first one in use. But the numbers in his dashboard keep bothering him. Plenty of people add a bag to the cart every day, and very few of them actually pay. The carts are full of bags; the checkout page is quiet. At first he assumed the price was too high, so he cut 20 yuan off to test it. Add-to-carts stayed the same, payments still did not move. Then he asked a few people who had added a bag and never checked out, and their answers surprised him. Nobody said it was too expensive. Nobody said the bag was bad. One said, "I was on the subway, I meant to buy it when I got home, and then I forgot." Another said, "I wanted to compare with other shops first." A third said, "Twelve yuan for shipping just put me off, so I left it there."

The numbers make it clearer. Last month Slow Bag had 7,200 add-to-carts and only 622 completed orders. In between, 5,184 people still had not paid an hour after adding to cart, and they went into the recovery pool. Of those 5,184, 3,110 opened a recovery message, 1,555 clicked back to the cart, and 622 paid. In other words, out of every 100 people who added to cart, fewer than 9 ended up buying. The other 91 did not dislike the bag; they walked to the doorway and turned back. Adding to cart is not the end of buying intent, it is the start of hesitation: putting something in the cart usually means "I want this, let me hold it for now," not "I will buy this right now." Treat an add-to-cart as a sale, and you are hearing "let me think about it" as "take my money now," which throws off every step after it.

One thing up front: the author, the shop, and every number in this guide are fictional examples; the methods themselves are real and usable. Treat the numbers as worked examples for doing the math, not as industry data to quote.

Sort them first: who really left, and who is just comparing prices

Not everyone who fails to pay is worth winning back, and not everyone should be won back the same way. Sort the people who added to cart and did not pay into four groups before deciding who to talk to and what to say. The first group is guests: not logged in, no email, no phone number. You do not even know who they are, so no email or text can reach them. All you can do is remind them inside the shop when they come back ("there is still something in your cart") or bring them back with retargeting ads. For these people, do not count on a recovery sequence; first find a way to get contact details at the add-to-cart step — for example, "leave your email and we will tell you if the price drops," trading a real benefit for a way to reach them. The second group is the comparers: they have three shops open at once, and the cart is just a bookmark. These people hate being chased; the harder you push, the more they feel pressured. What helps them is not a discount but information that settles the decision: how many are left, what other buyers said, how returns work. The third group was scared off by a detail in the price: shipping, tax, a free-shipping threshold — one extra line on the checkout page is enough to talk them out of it. For them, doing the math in the message beats a discount: "The bag is 268 yuan, shipping is free this time, 268 yuan in total, one tap to pay." The fourth group simply forgot: they added the bag on the subway and forgot about it at home. These are the easiest to win back; one reminder is enough, no discount needed.

If you start sending messages before sorting people into these groups, everyone gets the same line — "You forgot something! Buy it before it is gone" — and the person who genuinely forgot finds it baffling, the comparer feels pushed, and the person scared off by shipping still has no idea what shipping actually costs. The first step of recovery is not writing copy, it is sorting people. You do not need much to sort them: three fields are enough — whether they are logged in, whether you have a way to reach them, and how much is in the cart — plus one timing fact: how long it has been since they added to cart. Those four fields together are usually enough to tell which group someone is in and what to say to them. But do not overdo the sorting either: four groups are enough. Any finer, and you are making work for yourself; the more complicated the rules, the less anyone maintains them.

The recovery funnel: 7,200 add-to-carts, 622 orders back

Put the funnel on the table before deciding which step to fix first. Slow Bag last month: 7,200 add-to-carts; 5,184 people who still had not paid after an hour went into the recovery pool — that step lost 2,016 people, all of them fast payers who never needed recovery, so do not mourn them; 3,110 opened a recovery message, an open rate of about 60% (3,110 / 5,184 = 60.0%), losing 2,074 people at that step; 1,555 clicked the link back to the cart, a click-back rate of exactly 50% (1,555 / 3,110 = 50.0%), losing 1,555 people; and 622 finally paid, a payment rate of 40% (622 / 1,555 = 40.0%), losing 933 people. Add up the losses at the four steps: 2,016 + 2,074 + 1,555 + 933 = 6,578, which is exactly 7,200 − 622. The books balance. Across the whole funnel, fewer than 9 out of every 100 add-to-carts came back (622 / 7,200 = 8.6%); but if you only count the people who entered the recovery pool, about 12% came back (622 / 5,184 = 12.0%). Do not mix the two numbers up: 8.6% is the conversion rate of all add-to-carts in the shop, while 12% is the scorecard for the recovery work itself.

Recovery funnel: 7,200 add-to-carts through the recovery pool, message opens, and clicks back to cart, ending in 622 paid orders

The biggest hole in the funnel is not the last step, it is the second: 2,074 people never even opened the message. If the message is never opened, nothing written after it matters. So fix the message itself before the payment page: does the subject line say clearly that the item is still in the cart, is the sender name recognizable, is it sent at a time when people actually look at their phones? The last step, on the other hand, is not bad at 40%: four out of ten people who clicked back went on to pay, which means the people who come back want the bag badly, and the problem is not price or the checkout flow, it is that the earlier steps did not bring them back. Fix the funnel in this order: lift the open rate above 60% first, then the click-back rate, and only then touch the payment page. Plenty of people start by redesigning checkout and adding payment methods, which is pushing at the narrowest part of the funnel while the 2,074 people lost earlier stay lost.

Three channels: email at 1 hour, SMS at 24 hours, discount at 72 hours

A recovery message is not one message, it is a set, sent spaced out over time. Fix the sequence: an email 1 hour after the add-to-cart, a text message 24 hours later, and one more reminder with a discount 72 hours after that. Why this rhythm? At the 1-hour mark, the person has only just walked away and still remembers what they added, so one line in an email is enough — "The canvas bag you put in your cart is still there, tap here to finish checkout." No discount needed, and no need to sound anxious. At the 24-hour mark, whoever was going to forget has forgotten, so the text has to be short and direct: "Slow Bag here: the canvas bag in your cart is still there. Order today and it ships tomorrow." Keep it under 70 characters besides the link, and make the link land straight back in the cart. At the 72-hour mark, the people the first two steps did not bring back are mostly stuck on price or still hesitating, and that is when the discount comes out: "Three days now, we kept the bag for you — free shipping this time, valid for 48 hours." The discount is the last card to play; play it first and there is nothing left to play.

The three channels do not perform the same. Slow Bag's records over four straight weeks, counted per 100 sends: email opened 46% of the time, with an 8% repurchase rate within 7 days of the send; SMS opened 63%, with a 12% repurchase rate; the 72-hour discount reminder opened 39%, with a 10% repurchase rate. SMS opens most because a text on a phone is the hardest to ignore; the discount reminder opens least, but its repurchase rate is still solid — the people who open the third message were already hesitating, and the discount is exactly what tips them over. Note that the 60% open rate in the funnel is the share of people who opened at least one of the three messages combined, not the open rate of any single message, so do not try to reconcile 46% against 60%. And one hard rule: stop after three messages, do not send a fourth. From the fourth message on, unsubscribes and complaints climb, and the few extra orders do not make up for the trust you lose.

Three-channel comparison: open rates and repurchase rates for the 1-hour email, the 24-hour SMS, and the 72-hour discount reminder

How to write it: one reminder is enough — do not push, do not beg, do not lead with a discount

A recovery message has exactly one job: make the person remember what they added, then tap back. Three rules for writing it. First, talk about the item, not about yourself. "The black canvas bag you put in your cart is still there" is good; "We miss you, come back and take a look" is bad — nobody owes your shop their missing; they only remember the bag. Second, do not push and do not scare. Do not write "Buy now before it sells out" or "Only three left, grab it fast" unless there really are only three left; fake urgency works the first time, and by the second time people have learned to ignore every word you send. Third, do not lead with a discount. A "20 yuan off now" in the first email tells everyone that waiting gets them a cheaper price, so next time people will be in even less of a hurry to pay. Keep the discount for the third message, and only for the people who are genuinely hesitating.

The subject line deserves its own note, because it is the switch for the open rate. Write about the item, not the feeling: "Your canvas bag is still in your cart" beats "Are you really going to leave it behind?" — the first states a fact, the second sounds like a quarrel. Do not pile on emojis, do not write in all caps, and do not use words like "free" or "off now" that trip spam filters — once a message lands in the spam folder, no amount of work on the open rate will save it. Test two subject lines at a time: split one batch of recipients in half, show one half "Your canvas bag is still in your cart" and the other "Black canvas bag in your cart, 268 yuan," then compare open rates after a week. Only switch if the gap is more than 5 points; a gap of one or two points means the sample is still too small, and switching on that is just guessing.

The three messages can follow this shape. The first email, subject "Your canvas bag is still in your cart," three lines in the body: line one names the item, "Black canvas bag, 268 yuan, still in your cart"; line two does the math, "Pay now, ships today, shipping 12 yuan, 280 yuan in total"; line three is one button, "Continue to checkout." The second message is one text: "Slow Bag here: the black canvas bag in your cart is still there. Order today and it ships tomorrow — tap here to continue to checkout." The third, the discount reminder, subject "We kept it for three days — free shipping this time," with the terms spelled out: "Free shipping if you order within 48 hours, back to the regular price after that, valid only for this one cart." None of the three should carry anything unrelated to paying: no new-product recommendations, no brand story, no QR code for your social account. People open the message to finish something they started, not to browse your new arrivals.

How much discount: when to give it, how much, and who not to give it to

The discount is where recovery work most often goes wrong. Give it too early and you teach users to wait for a discount; give too much and the margin is gone; give it to the wrong person and someone who would have paid full price takes the discount anyway. So set three rules for discounts. First, timing: the first two messages only remind, they never mention price; the third message is the first to offer anything, and it offers only once. Second, a cap on size: do not go below 10% off, or just waive shipping. Slow Bag's bag sells for 268 yuan with a margin of roughly 40%; waiving 12 yuan of shipping is a 4.5% giveaway, which is fine; 20% off would give away more than 53 yuan, and if three out of ten of those orders would have paid full price anyway, the math goes under water. Third, only some people get it: high-value carts, returning customers, and people who still have not come back after 72 hours make the discount list; first-time visitors and carts holding a single cheap item do not. Write the test into the rules instead of deciding by feel.

Discount codes need care too. Give each cart its own one-time code that dies after one use; never send a generic code — a generic code ends up pasted on a coupon site, and then the whole shop is buying with it. Set the expiry to 48 hours, and when it expires, it expires; extend it once and people learn your deadline is fake. Another detail that is easy to miss: the discount only applies to the cart being recovered, not to anything else the person feels like buying. And keep a separate ledger for recovered orders: did this order use a discount, how much did the discount cost, and would this order have come back on its own without recovery? Slow Bag did this math and found that about half the people the third, discounted message brought back would have returned anyway — the discount only truly added half the orders it appeared to. Once you know that number, the next round knows who should get a discount and who should not.

Run the totals once a month, and do not judge the work by the order count alone. The formula is simple: recovered revenue minus discount cost minus texting cost equals net gain. A text costs only a few cents, but a few thousand of them a month still goes on the ledger; email is cheap, but the hours spent writing it are a cost too. In Slow Bag's first month the 622 recovered orders included fewer than half that used a discount code, the net gain was positive but only barely better than sending no discount at all — so in the second month the offer changed from free shipping to "12 yuan off orders over 300 yuan," and once the threshold went up, the bargain hunters thinned out and the net gain went up instead. If the net ever comes out negative, cut the size of the third message's discount first, not the first message's reminder; the reminder is free money, the discount is the part that costs.

Making it work: triggers, deduping, unsubscribes, and privacy

However good the idea, four things have to be right when it goes live. First, the trigger: an add-to-cart counts as abandoned only if there is still no payment 30 minutes later, and only then does it enter the recovery sequence. Do not set that number too short — someone who adds a bag, pours a glass of water, and takes a phone call will be back in ten minutes, and a "you forgot something" landing then feels like being watched; do not set it too long either, because after two hours the memory of the bag has faded. Second, the cancel rule: the moment someone pays, the sequence stops at once. The worst experience is paying and then getting a "you forgot something" on your phone, which tells the customer your shop does not even know it made the sale. Third, deduping: one cart enters the sequence once; do not send two sets of messages because someone refreshed the page twice, and if one person has several carts, take the most recent one. Fourth, unsubscribes and consent: email and text only go to people who clearly agreed to marketing messages. People who never agreed get at most one order-related reminder — and even that one can only say what is in the cart, with no discount attached. Put the unsubscribe link at the bottom of every email where one tap ends it, instead of making people hunt for it; replying "T" to a text unsubscribes, and the moment it arrives, stop.

Privacy deserves its own line: do not write things in a recovery message that you should not know. Do not write "we saw you look at this bag three times at 11 last night," and do not write "the warehouse near your home still has it in stock." Being able to see someone's browsing history does not mean you should say so in a message. Mention only what is in the cart itself, never the trail of how the person behaved. And test it yourself end to end before it goes live: add a bag with your own phone number and email, wait 1 hour for the email, 24 hours for the text, 72 hours for the discount reminder, and check that the copy, the link, and the discount code all line up at every step. Plenty of people ship it straight to real users, and the link in the first email opens an empty cart — because the cart they tested with was already cleared. Walking through it once yourself catches that kind of basic mistake.

Three things to do this week

Do not try to build the whole system at once; three things this week are enough. First, wire up the trigger — no payment 30 minutes after add-to-cart — and send only the first email, with no discount, then run it for a week and watch the open rate and the click-back rate. Second, split the recovery pool into the four groups — guests, comparers, people scared off by price, and people who forgot — or at least split "guests with no contact details" from "people you can reach," so the first group gets a way to leave contact details and the second enters the sequence. Third, write down four numbers: the recovery rate (orders paid divided by people who entered the recovery pool — in Slow Bag's example 622 / 5,184 = 12.0%), the open rate (3,110 / 5,184 = 60.0%), the click-back rate (1,555 / 3,110 = 50.0%), and the discounted share of recovered orders (recovered orders that used a code, divided by all recovered orders). Look at the four numbers once a week; whichever one has dropped tells you which step to fix. Do not change copy, timing, or discounts by feel.

One last thing: recovery is not tricking people back, it is inviting them back. The people who added to cart already wanted the item; all you have to do is help them remember it, show them the math, and let them pay in one tap. The more a message reads like a reminder, the more people come back; the more it reads like a sales pitch, the more people unsubscribe. Send the first email, run it for a week, look at the four numbers, and then decide what comes next.

Views 14Comments 0

Comments (0)

ME
0/1000
Loading comments...
Browse projectsPublish your project

Related articles

A young woman seals a cardboard shipping box with a red tape gun in a bright home office, shelves of shipping boxes behind her, a laptop on the desk
News
The Code Is Done. The Business Is Just Starting: Replit Agent Can Now Buy Supplies and Run Your Ads

On 12 November 2026, Replit gave Replit Agent two new jobs: buying supplies for your business and running your Google Ads. Purchases go through Stripe's Link wallet and Visa's rails; ads go through Google Ads. The common thread: the agent can only prepare the work — money moves only when the founder approves, with checks enforced in code, not left to the model.

Payments & MonetizationProduct LaunchAI Coding
Real photo: a bank of light gray metal filing cabinets with one drawer pulled open, revealing rows of hanging file folders inside
Guide
A real database inside the browser: IndexedDB local storage, explained plainly

localStorage only stores strings and has no indexes. IndexedDB is a real database in the browser: structured data, far larger capacity, indexes, and transactions. Following MDN's Using IndexedDB, IDBDatabase, and IDBObjectStore: the five-step routine, opening, upgrades, keys, indexes, transactions, add versus put, get, cursors, ranges, and multi-tab upgrades.

Tool TipsData WorkflowsFrontend Engineering
Real photo: hands typing on a MacBook Pro showing JavaScript code that changes HTML content, with a white mouse, papers, and binders on a white desk
Guide
Functions without the confusion: JavaScript functions, parameters, and arrow functions, explained plainly

You already call functions every day; this guide names what you have been doing. Following MDN's Functions, Arrow function expressions, and Rest parameters: declarations versus expressions, parameters and arguments, default values, rest parameters, spread syntax, return values, arrow function shorthand, the this trap, callbacks, and scope.

Tool TipsFrontend EngineeringAI Coding