Cart abandonment falls when surprise costs, unnecessary steps and payment doubt disappear, not when you add another discount popup.

That figure gets quoted often enough to sound like a fixed law of ecommerce, but treating it as a target to eliminate misreads what it actually measures. A meaningful share of "abandoned" carts were never purchase attempts at all, someone comparing three browser tabs, saving an item to think it over on payday, or browsing on a lunch break with no real intention to buy today.

None of that is a problem your checkout can fix, and chasing it with more urgency messaging or another popup usually just annoys a buyer who was never going to complete the order this visit. The abandonment worth spending effort on has a specific, identifiable cause sitting at a specific step: a shipping cost revealed too late, a signup wall, a payment method nobody recognises, or a mobile page too slow to finish before patience runs out.

A useful filter for telling the two apart: did the visitor reach the payment or delivery-detail step at all? Someone who leaves a product page without adding to cart is browsing. Someone who reaches checkout, hits a surprise, and leaves, that is the abandonment this article targets, and it almost always traces to one specific moment rather than a vague loss of interest.

Of every cause on this list, a late shipping-cost surprise is probably the single most avoidable one. Put an estimate on the product page, show it again in the cart, or state plainly "shipping calculated at checkout, typically RXX, RXX to major centres", a buyer expecting a small surprise abandons far less often than one given no warning at all.

If exact shipping genuinely cannot be confirmed until an address is entered, because your courier pricing is weight- or region-based, say so explicitly and give a realistic range rather than leaving the field blank until the last screen. Vague is fine. Silent is what costs you the sale.

Forcing account creation adds a decision point at the exact moment a buyer wants the fewest possible steps between them and a confirmed order. Make guest checkout the visible default, and invite account creation only after payment, once it carries zero friction risk.

Some platform themes bury the guest option behind a smaller, less obvious link than "create account." Check yours specifically, a technically-available guest option that a buyer never notices does no good at all.

Name, delivery address, contact number, payment detail. Anything beyond that list is friction with no offsetting benefit. An optional "how did you hear about us" field belongs in a post-purchase survey, not in the path to payment, move it, or delete it entirely.

Audit your current checkout field by field and ask, honestly, whether fulfilment or customer contact genuinely requires it. Most stores find at least one or two fields that have been sitting there since setup for no real reason.

Card, plus at least one instant EFT or wallet option such as Ozow, SnapScan or PayFast. Displaying only unfamiliar or card-only options at the final step raises hesitation right when commitment should be highest, particularly for buyers who would rather not enter card details online at all.

Test the checkout end to end in the currency and language your actual customers use. A gateway redirect to an unfamiliar third-party page with no visible link back to your branding can itself trigger last-second abandonment, even for a payment that would have gone through perfectly safely.

Most South African store traffic is mobile, often on variable data rather than reliable office WiFi. A cart or checkout page that takes several seconds to load, or that requires constant pinching and scrolling to complete a form, loses buyers who were otherwise ready to pay right now.

Compress checkout-page images, strip unnecessary scripts from the payment step specifically, and test the full flow on an actual mid-range Android device on real mobile data, not a fast laptop on office WiFi. Review widgets, chat tools and marketing pixels are common, overlooked contributors to a slow checkout; each one adds its own load time on top of the page.

• Stock counts shown are the real, current number, not a rotating "only 3 left" script.
• Countdown timers, when used, reflect a genuine deadline such as a real promotion end date.
• Delivery estimates match what customer service actually tells buyers when they ask.
• Reviews and ratings shown are real customer reviews, not manufactured or purchased ones.

1. Get consent to contactOnly email or SMS buyers who provided an address or number and did not opt out, POPIA applies to this data like any other.
2. Send the first reminder within a few hoursA factual note ("your cart is saved") with a direct link back to checkout, no pressure tactics needed yet.
3. Follow with a helpful second messageAnswer a likely doubt: confirm shipping cost and timeframe, or note the return policy, rather than repeating the same reminder.
4. Consider a modest, real incentive as a final touchA small, honestly time-limited discount can recover marginal carts, but should not become the default expectation for every visitor.

Retargeting ads reminding a visitor about a product they viewed can genuinely recover intent-to-buy traffic. Keep the message factual and specific to what they actually looked at, and cap frequency so the same buyer is not shown the same ad a dozen times a day.

Pair retargeting with the fixes above rather than treating it as a substitute for them, spending on ads to bring someone back to a checkout that still has the exact friction that lost them the first time just makes the leak more expensive to watch happen again.

• Shipping cost or a clear estimate shown before the final checkout step.
• Guest checkout available as the default path.
• Checkout fields limited to what fulfilment and contact genuinely require.
• At least one locally trusted payment method displayed clearly.
• Mobile checkout tested on real mobile data, not just office WiFi.
• Stock, delivery and urgency messaging are all factually accurate.
• Abandoned-cart emails or SMS only sent with proper consent, in a factual tone.

Track abandonment rate against your own baseline over time, not an industry-wide figure, it varies hugely by category, average order value and traffic source, and chasing a generic benchmark tends to produce the wrong fixes for your specific store.

When you introduce a recovery sequence, measure recovered revenue against the cost of any incentive offered, not just how many people clicked back to checkout. A "recovered" cart that only completed because of a 20% discount may not be profitable once real margin is accounted for.

Segment by device and traffic source where your volume allows it, a mobile-specific checkout fix often moves the overall number more than a general template change to the recovery email. Give any single change a fair, fixed comparison window against the prior period, and account for obvious seasonality before judging the result.

A small South African skincare brand selling on Shopify had reasonable traffic and a checkout abandonment rate that had bothered the owner for months, with no clear explanation. Session recordings showed a consistent pattern: customers filled their cart, entered their address at checkout, saw the calculated shipping cost for the first time, and closed the tab within seconds.

The product page and cart had never mentioned shipping at all. Adding a simple line to the product page ("Shipping from R65, calculated at checkout") and repeating the estimate in the cart before the address step reduced that specific drop-off point noticeably within the first two weeks, based on the store’s own checkout funnel report. No discount, no urgency banner, no design change, just removing a surprise that had been sitting there the whole time, unaddressed.

Week 1: complete a real checkout test on mobile data and note every point where a surprise, extra field, or unfamiliar payment option appears.

Weeks 2 to 3: fix shipping-cost transparency and guest checkout first (they are usually the highest-impact, lowest-effort changes) then set up a basic two-touch abandoned-cart email sequence with proper consent.

Week 4: review abandonment rate and recovered revenue against your prior baseline, and decide whether a modest, honestly-time-limited incentive is worth testing for the remaining gap.