There is no more qualified contact in an online store than someone who chose the product, added it to the cart, filled in details, and stopped. This person doesn’t need to be discovered, convinced about the category, or introduced to the brand — they already did all that themselves and got stuck on a specific doubt. This text addresses resolving that doubt via message: why abandonment happens, what to say at each touchpoint, what the channel’s rules allow, how to connect the store to support, and how much this is worth in a real operation.
In this article
- The hottest and cheapest contact
- The five real reasons for abandonment
- The first touch isn’t about discount
- The four-touch flow
- The 24-hour window and opt-in
- How the store communicates with support
- Where the human steps in and why they close more
- After recovery: the second purchase
- An operation in numbers
- Five mistakes in recovery
- Key metrics to track
- Frequently Asked Questions
The hottest and cheapest contact
The average cart abandonment rate in e-commerce ranges, depending on the segment and source, between 60% and 80%. This number is often treated as an industry fatality, when in practice it describes a queue of people who have already decided and stopped at the last step.
It’s worth comparing the cost of two customers. The first is unknown: they need to be reached by ads, attracted to the site, convinced of the product, price, and brand, and guided to the cart — all paid for. The second has already made this entire journey on their own and stopped for a reason that almost always fits in one sentence. Recovering the second costs a message.
That’s why cart recovery is the cheapest lever any online store has, and why it’s done poorly: handed over to a sequence of automated emails that most of the audience doesn’t open, when the same contact by message is read within minutes.
The five real reasons for abandonment
Before writing any message, it’s necessary to know what problem is being solved. Five causes dominate, and none of them is lack of interest:
- Shipping. The cost that appears at the end, the delivery time that doesn’t work, or the absence of a faster option. It’s the number one cause in practically every segment.
- Payment method. Insufficient installment options, absence of a specific method, declined card, or distrust in entering data.
- Product doubt. Size, measurement, compatibility, actual color, what comes in the box. The person stopped to ask and had no one to ask.
- Comparison. Opened another tab, went to check price elsewhere, and didn’t return. It’s not indecision — it’s a buying routine.
- Interruption. Life happened in the middle of checkout. This group is huge and the easiest to recover because there is no objection to overcome.
Note that four of the five reasons are solved with information, not price. This realization changes the text of the first contact.
The first touch isn’t about discount
The most costly and common mistake is to start recovery by offering a discount. It works enough to seem right and creates three problems.
The first is that it pays for those who were already going to buy. Many people who abandoned would return on their own, and giving coupons to all transfers margin to those who didn’t need it. The second is that it teaches behavior: customers who discover that abandoning the cart generates a coupon start abandoning always, and the store creates its own price erosion. The third is that it doesn’t solve the real reason — those who stopped because of delivery time remain stopped, now with a discount.
The correct first contact asks a question. Short, identifying the store and product, asking if there was any difficulty completing the purchase and if help is needed. It converts for two reasons: it resolves the doubt of those who had one and returns to decision those who were only interrupted. It also produces data the store didn’t have — the reason for abandonment, told by the person themselves, which over hundreds of conversations reveals what is broken in checkout.
Continue reading for free
Fill out once and unlock this and all other blog articles. We charge nothing — we just want to know who we’re writing for.
Takes 20 seconds. Prefer to talk directly on WhatsApp?
The four-touch flow
A well-constructed recovery flow has four contacts, spaced logically rather than habitually, each with a different objective.
- Touch 1, between fifteen and sixty minutes. The question. Short, personal, without aggressive payment links, with the product identified and an offer of help. It’s the highest converting touch in the entire flow and the most wasted when it becomes advertising.
- Touch 2, a few hours later. The likely objection, addressed. If the segment struggles with shipping, this is where delivery times and options are discussed. If it’s a technical doubt, this is where the missing information goes. This touch is written by product category, not the same for the entire store.
- Touch 3, the next day. Proof and reassurance. Reviews from buyers, exchange policy, warranty, real delivery time. Resolves hesitation of those unfamiliar with the store.
- Touch 4, with condition. Here, and only here, a benefit is introduced when it exists — shipping, gift, or discount — with a true deadline and an exit offered in the same message. If the person hasn’t responded to three contacts, the fourth is the last.
Two rules accompany the flow. The first is that any response interrupts automation and passes to support, because the person who replied wants to talk to someone, not receive the next scheduled message. The second is that those who purchase exit the flow immediately — charging abandoned carts after payment is the failure that generates the most complaints in this type of operation.
The 24-hour window and opt-in
There are two restrictions that define what is possible, and ignoring them is the quickest way to lose the number.
The first is consent. Sending messages to those who haven’t authorized is not just a channel policy issue: it’s a data protection matter. The correct way is to obtain consent at checkout itself, with a clear box informing that the store may send messages about the order and the incomplete purchase, and to record this consent with date, time, and origin. Without this, there is no flow — there is risk.
The second is the support window. On Meta messaging platforms, after a person sends a message, a window — commonly 24 hours — opens during which the company can respond freely; outside it, sending is restricted to specific message types subject to channel rules. In practice, this means the first contact in the flow, made to those who never wrote, must follow the channel’s approved format, and every reply from the person reopens the conversation and enables normal support. This is another reason the first touch should be a question: a question generates a reply, and a reply opens the window.
How the store communicates with support
Recovery only works if support knows what happened. Four pieces of information need to come from the store system to the conversation:
- What was in the cart, with name, variation, and price — a generic abandoned cart message converts much less than one that mentions the product.
- At which step the person stopped, because those who stopped at shipping and those who stopped at payment have different problems and deserve different texts.
- Customer history, since someone who bought three times shouldn’t receive the same approach as someone who never bought.
- Real-time order status, so automation stops the moment payment is received.
This connection is made by integration between the store platform and the support system. Where no ready integration exists, an intermediate path — periodic export or a simple trigger sent by the store — solves most of it, and is infinitely better than the common alternative, which is someone looking at the dashboard and copying phone numbers by hand.
Where the human steps in and why they close more
Automation opens the conversation; the sale is closed by someone who responds. Three situations require a person and are precisely the highest value.
The first is a specific product doubt, where the right answer closes the sale immediately and a generic answer ends the conversation. The second is a payment problem, where the person wants to buy and something got stuck — this is the easiest and most profitable case to solve, and the most frequently abandoned due to lack of someone available. The third is an exception request: a deadline, a condition, an early exchange. A small decision made quickly by someone with autonomy converts a sale that standard policy would lose.
The effective design is simple: automation while there is no reply, a person the moment there is one. Those who respond to a recovery message are one sentence away from purchase, and response time is the factor most correlated with conversion at this stage.
After recovery: the second purchase
Cart recovery has a valuable and almost always ignored byproduct: it transforms an anonymous buyer into an active contact with history. This is worth more than the recovered sale.
Three uses leverage this asset. Order confirmation and delivery tracking via the same channel, which reduces the biggest source of support messages in any store — "where is my order?" The offer of a complementary product at the right moment, days after delivery, when the person is using the product and the recommendation makes sense. And the planned repurchase, for consumable products with a known cycle, notifying when it’s time to replenish.
None of these three require paid media, and all three depend on one thing that recovery already delivered: permission to converse.
An operation in numbers
An online home goods store received about 4,200 carts per month and completed 1,008 orders, with an average ticket of R$ 218 — a conversion rate of 24% and 3,192 abandoned carts per month. Recovery existed by email, with three automatic sends, and brought 84 orders per month, about 2.6% of abandonments.
The change was to replace the email sequence with a four-touch message flow, with opt-in collected at checkout, first contact within thirty minutes with a question instead of a coupon, human support triggered on any reply, and discount only on the fourth touch.
After three months: 41% of abandonments began receiving the flow — the rest had no consent or valid phone number. Of these 1,309 monthly contacts, 22% responded, and 311 orders were recovered per month, with an average ticket of R$ 232. That’s R$ 72,152 monthly versus R$ 18,312 from email, with an operational cost of two people working part-time shifts.
Two side effects appeared in reports. Responses to the first touch revealed that 38% of abandonments were due to delivery time in a specific region — information the store didn’t have and which led to hiring another carrier. And 27% of recovered customers bought again within ninety days, versus 14% of customers who bought without going through the flow, simply because they continued having an open channel with the store.
Five mistakes in recovery
- Starting with a discount. Pays for those who were already going to buy, teaches customers to abandon, and doesn’t solve the real reason.
- Sending without consent. Besides the risk of blocking, it’s a data protection issue — consent must be collected at checkout and recorded.
- Not stopping the flow when the person buys. Charging abandoned carts after payment is the failure that generates the most complaints in this type of operation.
- Leaving replies unattended. Those who reply are one sentence from purchase, and subsequent automation destroys the moment.
- Using the same message for the entire store. The second touch needs to address the category’s objection, and different categories stall for different reasons.
Key metrics to track
Five indicators drive this operation: flow coverage, i.e., the percentage of abandoned carts that actually receive contact, usually the biggest bottleneck and dependent on opt-in; response rate to the first touch, measuring whether the message sounds human or automated; recovery rate over contacts sent, compared by touch to see which is working; declared reason for abandonment, tabulated, which is a free checkout diagnosis; and ninety-day repurchase rate of recovered customers versus others, which is the long-term return almost nobody measures.
Frequently Asked Questions
Why does recovering carts via WhatsApp convert better than email?
Because the channel is read. The average abandonment rate in e-commerce ranges between 60% and 80% depending on the segment, and this queue is composed of people who already chose the product, filled in details, and stopped at the last step — the most qualified contact a store has. Recovering this customer costs a message, while acquiring a new one costs ads, traffic, brand persuasion, and price. The problem with email isn’t the text: it’s that much of the audience doesn’t open it, and an automatic sequence of three sends usually recovers around 2% to 3% of abandonments. The same sequence by message is read within minutes, allows immediate response, and turns recovery into conversation. In a real operation, the switch increased recovery from 84 to 311 monthly orders without increasing media investment.
What to write in the first abandoned cart message?
A question, not a discount. Starting with a coupon is the most costly and common mistake, for three reasons: it pays those who would return on their own, transferring margin to those who didn’t need it; it teaches customers to always abandon because they discover abandonment generates a discount; and it doesn’t solve the real reason, since those who stopped due to delivery time remain stopped, now with a discount. The correct first contact is short, identifies the store and product, asks if there was any difficulty completing the purchase, and offers help. It converts because it resolves the doubt of those who had one and returns to decision those who were only interrupted — and produces data the store didn’t have, the reason for abandonment told by the person themselves, which over hundreds of conversations reveals what is broken in checkout.
How many messages to send and in what timeframe?
Four touches, spaced logically rather than habitually. The first between fifteen and sixty minutes, with the question — it’s the highest converting touch in the flow and the most wasted when it becomes advertising. The second a few hours later, addressing the likely objection of that category: delivery time and shipping where the problem is delivery, technical information where the doubt is about the product. The third the next day, with proof and reassurance, including reviews, exchange policy, and real delivery time. The fourth with condition, the only place where a benefit enters, with a true deadline and an exit offered in the same message. Two rules accompany this: any reply interrupts automation and passes to human support, and those who buy exit the flow immediately, because charging abandoned carts after payment is the failure that generates the most complaints.
Do I need customer authorization to send abandoned cart messages?
Yes, and this is not just channel policy — it’s data protection. The correct way is to obtain consent at checkout itself, with a clear box informing that the store may send messages about the order and incomplete purchase, recording consent with date, time, and origin. Without this, there is no flow, only risk. There is also the support window rule: on Meta messaging platforms, after the person sends a message, a window commonly of 24 hours opens during which the company can respond freely, and outside it sending is restricted to specific message types subject to channel rules. In practice, the first contact to those who never wrote must follow the approved format, and every reply reopens the conversation — another reason the first touch should be a question.
What does the store need to integrate for the flow to work?
Four pieces of information need to come from the store system to the conversation. What was in the cart, with name, variation, and price, because a generic message converts much less than one mentioning the product. At which step the person stopped, since those stuck at shipping and those stuck at payment have different problems and deserve different texts. Customer history, because someone who bought three times shouldn’t receive the approach of someone who never bought. And real-time order status, so automation stops the moment payment is received. Where no ready integration exists between platform and support system, a periodic export or a simple trigger sent by the store solves most of it — and is much better than someone copying phone numbers from the dashboard by hand.
The abandoned cart is the cheapest lever your store has
VeyloCRM integrates your store with support, triggers the recovery flow with the product identified, stops automatically when payment is received, transfers to a person the moment the customer replies, and keeps WhatsApp, Instagram, and Messenger in the same panel with a unified history — so the conversation that recovered the sale remains valid for the next purchase.