Seven in Ten Carts Are Abandoned — And That Has Not Changed Since 2010 🛒

Seven in ten online carts are abandoned.

The exact figure is 70,22%, pooled from 50 separate studies (Baymard Institute, 2026).

Here is the part that should bother you more.

That number has barely moved since 2010.

Fifteen years of better websites, faster phones and smarter payments. Same result.

This article is not about the problem. It is about the roughly half of it that can actually be fixed. 🛒

🧾 Principais conclusões em resumo

Medir	Figura	Fonte
Average cart abandonment	70,22%	Instituto Baymard (2026)
Studies pooled for that figure	50	Instituto Baymard (2026)
Range across individual studies	55% to 84%	Instituto Baymard (2026)
Recoverable sales lost yearly	$260 billion (US + EU)	Instituto Baymard (2026)
Possible conversion gain	35.26%	Instituto Baymard (2026)
Form fields that could be removed	20% to 60%	Instituto Baymard (2026)
Quit because checkout was too long	Nearly 1 in 5	Instituto Baymard (2026)

🧊 The Half You Cannot Fix

Start here, because most advice skips it.

A large share of abandonment is not a failure at all.

People use carts as wish lists. They compare prices across shops. They check what delivery would cost.

Adding to a cart is often research, not intent to buy.

That behaviour is normal and permanent.

Why chasing zero is a trap

Some shops set a goal of cutting abandonment dramatically.

They then spend heavily on tactics that cannot work.

You cannot design away a shopper who was never going to buy today.

Aim at the friction share instead. That part responds.

What the 55% to 84% range tells you

Individual studies vary enormously (Baymard Institute, 2026).

That spread reflects different shops, products and prices.

A furniture shop and a takeaway app cannot share a benchmark.

So your own rate matters far more than the average.

💰 The $260 Billion Word: Recoverable

Estimativas Baymard US$ 260 bilhões em vendas recuperáveis are lost each year across the US and EU (Baymard Institute, 2026).

Recoverable is doing important work in that sentence.

It is not the value of every abandoned cart.

It is the portion attributable to checkout design that could be improved.

Where the recoverable number comes from

Baymard runs large-scale usability testing on real checkouts.

They watch people fail, then measure what fixing it is worth.

That is different from surveying opinions.

Observed behaviour beats reported behaviour, every time.

The 35.26% figure explained

Baymard found the average large ecommerce site could raise conversion by 35.26% through better checkout design.

Note carefully. That is a relative gain, not a jump to 35% conversion.

A shop converting 2% would move to roughly 2.7%.

On $200,000 of yearly sales, that is about $70,000 more.

Same traffic. Same products. Better checkout.

📝 The Form Field Problem

This is the single most actionable finding in the research.

Baymard reports most checkouts could cut 20% to 60% of their form fields (Instituto Baymard, 2026).

Every field is a small chance to stop.

Fields most shops do not need

Run through your own checkout and question each one.

Campo	Usually needed?
Company name	No, unless you sell to businesses
Address line 2	Can be optional or hidden
Phone number	Only if the courier requires it
Confirm email	No — show what was typed instead
Account password	No — offer guest checkout
How did you hear about us	Never at checkout

That last row appears more often than you would think.

Marketing questions belong after payment, not before it.

The billing address duplication

Many checkouts ask for shipping, then billing, separately.

For most orders they are identical.

Default to “same as shipping” and let people change it.

That one change removes several fields for nearly everyone.

🚪 The Forced Account Problem

Requiring an account before purchase is among the most costly choices a shop makes.

The reasoning is understandable. Accounts help retention.

But the timing is wrong.

You are asking for commitment before you have delivered anything.

What guest checkout actually costs you

Almost nothing, and this surprises people.

You still capture the email address for the order.

You can offer account creation on the confirmation page, after payment.

At that point the person has already trusted you.

The one-click follow-up

Offer to create the account with one tap after checkout.

They have entered everything already. Only a password is missing.

Conversion first, relationship second.

The reverse order loses both.

💳 Payment Options and the Typing Problem

Every card number typed on a phone is a risk of abandonment.

Mobile abandonment runs at 73% to 75%, against 65% to 68% on desktop (Baymard Institute, 2026).

Typing difficulty is a large part of that gap.

Why digital wallets matter so much

They remove typing entirely.

The card details are already stored on the device.

Payment becomes a fingerprint or a face scan.

For mobile shoppers, wallet support is not a nice extra. It is the fix.

Which options to offer

Do not offer everything. Choice itself causes hesitation.

Offer cards, the two major device wallets, and one local method if your market expects it.

Four options is plenty. Twelve is a decision problem.

Where the payment step should sit

Last. Always last.

People will enter an address before they will enter a card.

Asking for payment early raises suspicion before trust is built.

💸 The Surprise Cost Problem

Unexpected extra costs are one of the most cited reasons for leaving.

Delivery charges are the usual culprit. Taxes and fees follow.

The issue is rarely the amount.

A cost revealed at step four feels like a trick. The same cost at step one is just information.

How to show costs early

Put delivery cost on the product page where you can.

Show a running total in the cart, not only at the end.

If delivery depends on location, ask for a postcode early.

People accept charges they saw coming.

The free delivery threshold

If you offer free delivery above a value, say so in the cart.

Show how far away they are from it.

This both reduces abandonment and raises order value.

It is one of very few changes that does both.

📉 Where People Drop Out, Step by Step

Abandonment is not one moment. It is a leak at every stage.

Mapping the funnel shows where your own money goes.

The biggest leak is usually first

Most shops lose the largest share between adding to cart and opening it.

That drop is mostly browsing behaviour, not friction.

The leaks worth fixing sit after checkout begins.

Someone who started entering an address wanted to buy.

Measure the last step first

If you track one thing, track the step from checkout start to payment.

That group showed real intent.

Losing them is expensive and usually preventable.

Baymard’s testing focuses on exactly this stage (Baymard Institute, 2026).

⏱️ Speed Is a Conversion Feature

Page speed rarely appears in checkout advice. It should.

Each extra second of load time gives someone a reason to give up.

On mobile data, pages that feel instant on office wifi can crawl.

What usually slows a checkout

Not the checkout itself. The things bolted onto it.

Common cause	Consertar
Tracking scripts on checkout pages	Remove all non-essential ones
Chat widgets loading early	Load after the page is usable
Large uncompressed images	Compress and resize
Review widgets in the cart	Not needed at this stage

The first row is the common one.

Analytics scripts collecting data about abandonment can help cause it.

If you need to shrink images, our image compression tool is free to use.

A quick way to test it

Open your checkout on a phone with wifi switched off.

Count the seconds until you can type into the first field.

Anything over three seconds is costing you money.

🔒 Trust Signals That Actually Work

Some shoppers leave because they do not trust the site with card details.

Most trust advice is decorative. A few things genuinely help.

Signal	Does it help?
Visible return policy at checkout	Yes, strongly
Real contact details and address	Sim
Recognised payment logos	Sim
Clear delivery timeframe	Sim
Generic security badges	Weak — everyone has them
Countdown timers	Often harmful

Why the return policy matters most

It answers the real question in the shopper’s head.

That question is not “will you steal my card”.

It is “what happens if this is wrong”.

A clear return policy removes the risk of buying.

Why timers backfire

Artificial urgency is now widely recognised.

Shoppers have seen countdowns reset when they reload the page.

Once someone suspects manipulation, trust drops sharply.

Genuine scarcity is fine. Invented scarcity is expensive.

📱 Fixing Mobile Specifically

Mobile is where most traffic and most losses are.

The fixes are physical rather than clever.

Make targets bigger

Buttons should be comfortably tappable with a thumb.

Small links placed close together cause mis-taps.

A mis-tap at checkout often ends the session.

Use the right keyboard

This is a small technical detail with a real effect.

A number field should bring up a number keypad.

An email field should show the @ symbol.

Wrong keyboards add seconds and irritation to every field.

Test on real conditions

Test on a real phone, on mobile data, not office wifi.

Test with one hand while standing up.

That is how people actually shop.

Most shop owners have never done this once.

🧪 How to Test Changes Without Fooling Yourself

Improvements are easy to imagine and hard to prove.

Regra	Por que
Change one thing at a time	Otherwise you learn nothing
Wait at least two weeks	Weekly patterns distort short tests
Compare to the same period last month	Controls for seasonality
Split mobile and desktop	They behave differently
Write the prediction down first	Stops after-the-fact stories

That last rule is the most useful and the least used.

If you predict the result before testing, you cannot fool yourself afterwards.

Watch five real people instead

If your traffic is small, statistics will not help.

Watch five people try to buy something while you say nothing.

Baymard’s whole method is built on this kind of observation.

You will find more in thirty minutes than in a month of dashboards.

📋 The Checkout Audit, In One Sitting

Here is a practical sequence that takes about an hour.

Etapa	O que fazer
1	Buy something from your own shop, on a phone
2	Count every form field
3	Mark each one required, optional or removable
4	Check when delivery cost first appears
5	Check whether an account is forced
6	Check which wallets are offered
7	Note where the return policy is visible

Most shops find three obvious fixes in that hour.

The tooling side is covered in our guia de software para pequenas empresas.

Fix in this order

Remove fields first. It costs nothing and cannot break trust.

Then show costs earlier. Then add guest checkout.

Then add wallets, which usually needs developer time.

Cheapest and safest first, always.

🔬 Why Baymard’s Data Is Worth Trusting

Not all ecommerce statistics deserve equal weight.

Recurso	Por que isso importa
Pools 50 separate studies	Not one survey’s quirks
Based on observed testing	Watches behaviour, not opinions
Publishes the range, not just average	Admits variation
Consistent method over 15 years	Trends are real
Sells research, not checkout software	No product to flatter

The stability of the 70% figure is itself evidence of quality.

A number that refuses to move despite pressure to look better is usually honest.

Why observed testing beats surveys here

Ask people why they abandoned a cart and they will give a tidy answer.

Watch them abandon one and you often see something else entirely.

People misremember, and they explain their behaviour after the fact.

They rarely notice the small friction that actually stopped them.

That is why usability testing finds problems surveys miss.

The limits worth naming

Testing skews toward larger sites in developed markets.

The 35.26% uplift is what is possible, not what is typical.

And usability testing uses small groups by design.

Those are reasonable trade-offs, not flaws, but they shape how you should apply the numbers.

📧 Recovery Emails: What They Can and Cannot Do

Most shops reach for abandonment emails first. They are useful but limited.

They work on the group that intended to buy and got interrupted.

They do nothing for the browsing majority.

A recovery email is a reminder, not a repair.

Why fixing checkout beats emailing

An email asks someone to return and face the same broken checkout.

If the reason they left is still there, they will leave again.

Fix the cause first. Then the reminder has somewhere good to send people.

What a sensible sequence looks like

Timing	Contente
About 1 hour	Simple reminder, no discount
About 24 hours	Address a likely objection
About 3 days	Optional, only if margin allows a discount

Note the first row carries no discount.

Discounting immediately trains people to abandon carts on purpose.

That is a habit you can create by accident and struggle to undo.

The consent point

You can only email people who gave you an address and agreed to contact.

Rules differ by country and are enforced more than they used to be.

Collecting the email early in checkout is fine. Emailing without consent is not.

🌍 Why Local Payment Habits Change Everything

Checkout advice written for one market often fails in another.

Payment preference is strongly regional.

Market pattern	What it means for checkout
Card-dominant markets	Wallets matter most on mobile
Bank-transfer markets	Cards alone will lose sales
Cash-on-delivery markets	Trust concerns dominate
Instalment-heavy markets	Price framing changes behaviour

Global averages hide all of this.

If you sell across borders, your abandonment rate is really several different rates.

How to find your own pattern

Split your checkout data by country before drawing conclusions.

One weak market can drag an otherwise healthy average down.

That looks like a checkout problem when it is a payment-method problem.

Our growth analysis covers how uneven regional markets are.

🚫 What This Data Does Not Tell You

It does not give you a target. The 55–84% range is too wide for that.

It does not account for your prices. Expensive items are abandoned more.

It does not cover every market. Payment habits vary hugely by country.

It cannot separate browsing from friction precisely. Nobody can.

The uplift figure is a ceiling. Most shops will capture part of it.

🏁 Versão curta

70.22% of carts are abandoned, and that has held since 2010 (Baymard Institute, 2026).

Roughly half of that is ordinary browsing you cannot remove.

The rest is friction, and friction is fixable.

Cut form fields, show costs early, allow guest checkout, support wallets.

Baymard puts the possible gain at 35.26% and the lost value at $260 billion a year.

None of the fixes are clever. All of them are boring.

That is why they still work fifteen years later. 🛒

❓ Perguntas frequentes

What is a good cart abandonment rate?

There is no single good number. The average is 70.22%, but studies range from 55% to 84% (Baymard Institute, 2026). Compare against your own past performance.

Why is mobile abandonment higher?

Mobile runs 73–75% against 65–68% on desktop. Typing is harder, screens are smaller, and people often shop while doing something else.

What is the single best fix?

Removing unnecessary form fields. Baymard finds most checkouts can cut 20–60% of them, and it costs nothing to do.

Should I force customers to create an account?

No. Offer guest checkout and invite account creation after payment, when trust already exists.

Do trust badges help?

Generic security badges do little because every site has them. A visible return policy and real contact details help far more.

Are countdown timers effective?

Often harmful. Shoppers recognise artificial urgency, and suspicion damages conversion more than urgency helps it.

How much money is at stake?

Baymard estimates $260 billion in recoverable sales lost each year across the US and EU.

How long should I run a test?

At least two weeks, comparing to the same period previously, with mobile and desktop measured separately.

Do abandonment recovery emails work?

They help with people who were interrupted, not with people who were browsing. Fix the checkout first, or the email sends them back to the same obstacle.

Should the first recovery email include a discount?

No. Discounting immediately teaches shoppers to abandon carts deliberately, which is a habit that is hard to undo.

Why does my rate differ from other shops?

Price, product type and country all move it. Bank-transfer and cash-on-delivery markets behave very differently from card markets.

Does page speed affect abandonment?

Yes. Tracking scripts and chat widgets on checkout pages are a common cause, and they are usually safe to remove.

How many people should I watch testing my checkout?

Five is enough to find the obvious problems. Usability testing deliberately uses small groups.

Onde posso ler a pesquisa original?

Baymard publishes its cart abandonment statistics openly. The link is in the references below.

📚 Referências

Instituto Baymard. (2026). Estatísticas da taxa de abandono de carrinhoConsultado em 8 de agosto de 2026, em https://baymard.com/lists/cart-abandonment-rate

Departamento do Censo dos EUA. (2026). Vendas trimestrais de comércio eletrônico no varejo: primeiro trimestre de 2026Consultado em 8 de agosto de 2026, em https://www.census.gov/retail/ecommerce.html

Shopify. (2026). Estatísticas e tendências globais do comércio eletrônicoConsultado em 8 de agosto de 2026, em https://www.shopify.com/enterprise/blog/global-ecommerce-statistics

IBM. (2026). Custo de um relatório de violação de dados em 2026Consultado em 8 de agosto de 2026, em https://www.ibm.com/reports/data-breach

Instituto Stanford para Inteligência Artificial Centrada no Ser Humano. (2026). Relatório do Índice de IA de 2026Consultado em 8 de agosto de 2026, em https://hai.stanford.edu/ai-index/2026-ai-index-report

Leituras relacionadas neste site

Our ecommerce growth analysis puts these figures in market context. For the software side of running a shop, see the guia para pequenas empresas. Shops holding customer data should also read our breach cost analysis.

Sobre esta análise

Figures come from Baymard Institute, which pools 50 studies and conducts its own large-scale usability testing. Where a split between fixable and unfixable abandonment is shown, it is labelled as indicative, because Baymard reports causes rather than a single ratio. Figures were checked on August 8, 2026.