Roughly 70% of online shopping carts get abandoned. It’s one of the most quoted stats in ecommerce, and most retailers have quietly made peace with it — treating cart abandonment as a fixed cost of doing business online.
It shouldn’t be. And once you run the math, it stops being one.
What the cart abandonment rate actually is
The Baymard Institute’s ongoing meta-study of 49 separate cart abandonment studies puts the documented average at 70.19%. Statista’s tracking lands in the same neighborhood, and the number climbs higher on mobile — typically 80–85% depending on category. In some verticals like luxury and travel, mobile checkout abandonment routinely crosses 87%.
Translation: for every ten shoppers who add to cart, seven leave without buying. On mobile, closer to eight.
The industry’s default fix arrives too late
Walk into almost any ecommerce stack and you’ll find the same recovery playbook: a retargeting ad that follows the shopper around the internet for a week, an abandoned cart email three days later with a 10% discount code, maybe an SMS nudge on top. Every element of that stack is designed to win the customer back after they’ve already left.
It works — a little. Klaviyo’s public benchmarks put abandoned cart email recovery rates at roughly 3–11% depending on industry, with the median clustered around 10%. That’s not nothing. But it’s treating the symptom, not the cause. And it’s doing it with a discount you often didn’t need to give.
Why shoppers actually abandon carts
Baymard’s survey data on why shoppers walk away is the part most retailers skip past. The top reasons, in order:
- 48% — extra costs (shipping, taxes, fees) too high
- 26% — the site wanted them to create an account
- 25% — didn’t trust the site with card info
- 23% — delivery was too slow
- 22% — checkout process was too long or complicated
- 18% — couldn’t see or calculate total cost up front
- 17% — return policy wasn’t satisfactory
Read that list carefully. Almost every one of those reasons is a question the shopper couldn’t get answered in the moment. What’s the actual delivered price? Do I have to make an account? When will this arrive? Can I return it if it doesn’t fit? Is this site legit?
A discount email three days later doesn’t answer any of those. It just offers a bribe to come back and hope the question resolves itself.
The real cause is usually one unanswered question
Does this ship to Canada? Will this fit a 6’2” frame? What’s the return window — 14 days or 30? Is this actually in stock in navy, or just the color swatch? Is the subscription cancellable after one month?
These are small questions with small answers. But when nobody answers them at the moment the shopper is deciding, the shopper doesn’t wait around. They close the tab. If you’re lucky, one of those retargeting ads brings a fraction of them back three days later — at a discount you didn’t need to give.
The math: recovering just 5% pays for the agent in week one
Let’s put realistic numbers on it. Take a mid-size ecommerce store:
- Average order value: $50
- Sessions reaching checkout each month: 10,000
- Cart abandonment rate: 70%
That’s 7,000 abandoned carts per month, worth $350,000 in stalled revenue — every single month.
You don’t need to recover most of that to make the case. Recover just 5% of those abandoned carts — 350 orders — by answering the hesitation in real time instead of three days later, and you’ve captured $17,500 in monthly revenue an AI Brand Agent had a direct hand in closing. That’s $210,000 a year, on the conservative end, from the shoppers you were already paying to acquire.
Now scale it up. A store doing $150 AOV on 25,000 checkout sessions is sitting on $2.6M in monthly abandoned revenue. Recovering 5% is $131,000 a month. Recovering 10% — still well below what real-time answering achieves in most deployments — is $262,000.
Against the cost of running a Brand Agent, that isn’t a marginal improvement. That’s the entire business case, in one month.
Why real-time beats retroactive, every time
The evidence on in-the-moment engagement is consistent across categories. Forrester’s research on live chat and conversational commerce has documented 10–30% conversion rate lifts when shoppers engage with a real-time assistant during checkout. Salesforce’s State of the Connected Customer data shows shoppers who interact with AI during a session convert at roughly 3–4x the rate of shoppers who don’t.
The reason is simple: an abandoned-cart email is trying to reconstruct intent that has already cooled. A Brand Agent is preserving intent that’s still hot.
This is a conversation problem, not a marketing problem
The fix isn’t a better discount email. It’s making sure the question gets answered before the shopper decides to leave, not after. A Brand Agent sitting at the point of decision — on the product page, in the cart, at checkout — can answer the shipping question, confirm the fit, clarify the return policy, and keep the customer moving toward the purchase they already wanted to make.
And unlike a chatbot bolted onto a support widget, a real Brand Agent shows up in the brand’s voice, pulls in the actual product video, remembers the returning visitor, and captures every abandoned interaction as a first-party signal about what’s costing you sales.
Retargeting ads and abandoned cart emails aren’t going away. They still have a job to do for the shoppers who leave for other reasons — got distracted, comparison shopping, weren’t quite ready. But for the majority of abandonments — the ones caused by one small, answerable question — the fix was never a follow-up campaign.
It was just answering the question the first time.
Frequently asked questions about cart abandonment
What is a good cart abandonment rate?
The Baymard Institute’s documented industry average is 70.19%. Anything below 65% is strong. Above 75% (or above 82% on mobile) suggests specific checkout friction worth investigating.
Do abandoned cart emails actually work?
Yes, but modestly. Klaviyo’s public benchmarks put average recovery at around 10%, meaning 90% of abandoned carts stay abandoned even with the best email flow in place. Emails also usually require offering a discount, which cuts into margin on the sales you do recover.
What’s the difference between a chatbot and a Brand Agent?
A chatbot is typically a scripted or FAQ-matched deflection tool sitting on a support page. A Brand Agent is trained on the brand’s actual products, content, and voice, shows up at the point of decision, uses rich media (video, product cards), and captures every conversation as structured data. It’s built to convert, not deflect.
How quickly does a Brand Agent affect cart abandonment?
Real-time answering affects conversion in the same session it’s deployed. Most brands see measurable lift within the first two weeks of installation, before any tuning.
See what real-time answering does to your abandonment rate →
