An online shop's revenue comes down to three numbers: visitors times conversion rate times average order value. To increase it, measure which of the three is weakest, then work on that one. A list of ten tips only helps after that. This article follows that order, with the metrics, our own sample of product pages, and a section on what you can still do before Black Friday on 27 November 2026.
Sources were read on 5 November 2026. Legal points here aren't legal advice, as we repeat where it matters.
The three levers at a glance
An invented example: 10,000 monthly visitors, a 2% conversion rate and a €60 average order give €12,000 in revenue. Raise conversion to 2.2% and it's €13,200. A €66 order or 11,000 visitors gets you the same. Which route is easiest depends on your shop, and the numbers tell you that.
| Lever | Metric | Typical actions | Effort (our judgement) |
|---|---|---|---|
| Visitors | Sessions per channel, clicks in Search Console | Category pages, Shopping ads, email to existing customers | Medium to high, some takes months |
| Conversion rate | Purchases per session, drop-off per step | Show shipping early, guest checkout, delivery date, payment methods | Low to medium, often single changes |
| Order value | Revenue per purchase, items per purchase | Free shipping threshold, relevant add-ons, bundles | Low, but it can cost margin |
| Repeat purchase | Share of returning buyers | Post-purchase email, loyalty programme | Medium |
The fourth row isn't a separate factor in the formula, since it works through visitors and conversion. We list it because you can measure it separately.
Measure first: what and where
E-commerce data doesn't flow into GA4 by itself. Google says Analytics doesn't collect e-commerce events automatically, so you have to set them up. It names one exception, Shopify, where some events are collected by themselves. Check this before reading any report. If an event is missing, a step looks like a collapse when only the measurement is absent.
| Step in the shop | GA4 event | What it can tell you |
|---|---|---|
| Saw a list | view_item_list | Which categories lead to product clicks |
| Saw a product | view_item | Which products are often viewed and rarely bought |
| Added to cart | add_to_cart | Whether the product page persuades |
| Viewed cart | view_cart | Whether visitors review the cart |
| Began checkout | begin_checkout | Whether the cart holds or costs come as a surprise |
| Chose shipping | add_shipping_info | How many continue after seeing delivery details |
| Chose payment | add_payment_info | Whether the payment method fits |
| Purchased | purchase | Revenue, order value, channel |
The right-hand column is our reading. For analysis there's the E-commerce purchases report and funnel exploration, where you can define up to ten steps. We suggest splitting by device and by new versus returning users. A drop that occurs only on phones is a different problem from one on every device.
Two sources of error are worth knowing. Without consent, data may be missing, so GA4 figures aren't your complete order numbers: reconcile purchase with your shop system. And if you only look at conversion across all channels, you can miss a change in the channel mix that has nothing to do with the shop getting better or worse.
Our sample: what shops show on the product page
We wanted to see how shops handle the points that appear in nearly every list of tips. On 5 November 2026 we hand-picked 20 shops, 10 for Germany and 10 for Spain, some of them from the top search results for common product terms. We checked only the product page and cart, with no login and no order, using an automated browser with no disguise.
The first result is the gap: we could measure only 6 of the 20. At 11 shops a block, a security check or a CAPTCHA stopped us, and we didn't get around any of them. The other 3 failed through maintenance, an empty page or a connection error. Of the 6 measured, 4 are shops for Germany and only 2 for Spain.
| Visible without logging in | Shops (of 6) | Note |
|---|---|---|
| Statement on shipping costs before the cart | 6 | A plain "free" counts. Only one of the 6 showed an amount, and it applied to delivery to Spain |
| Delivery date or time on the product page | 5 | One showed delivery to Spain, because we measured from there |
| Return period on the product page or in the cart | 5 | In one case only at the foot of the page. They stated 30 or 60 days or a fixed date |
| Payment methods | 4 | In 2 cases only in the footer, as logos |
| Free shipping threshold | 3 | €49 and €79 |
| Guest checkout visible in the cart | 0 | "Not seen" doesn't mean "doesn't exist": we deliberately didn't enter the checkout |
What this shows and what it doesn't. Six shops aren't a statistic. They were hand-picked, vary in size and are only those that let us through, so percentages would mislead. One observation remains: delivery time was mostly shown on the product, while guest checkout was visible at none of them before the checkout. A customer may therefore learn late whether it exists. That's an inference from a handful of cases.
A question for your own shop. That 11 of 20 stopped us says nothing about Googlebot. It does raise the question of who else your rules catch: price comparison services, audit tools, AI shopping assistants. That may be intended. Check your firewall logs to see whether it was actually decided that way.
The product page
This is where the decision to add to cart is made. Here are the points, in the order they answer questions:
- What exactly is it? Specifications, measurements and variants as text, not only in images. That helps search engines too.
- Will it fit? Images from several angles, plus a size guide or comparison where wrong purchases are likely.
- When will it arrive? A date is more concrete than a range. Five of the six shops we measured gave one or the other here.
- What does it cost in total? Price with shipping, or at least a note on shipping costs.
- What if it isn't right? The return period and who pays for return shipping.
In Germany, shipping isn't only a conversion question. Section 6 of the price indication regulation (PAngV) requires distance sellers to state whether shipping costs apply and how much, as far as they can reasonably be calculated in advance. Other countries have their own rules, so have the details checked.
For search and Google's shopping surfaces, product schema markup is also worthwhile. Google documents specific fields for shipping and returns, in Merchant Center as well.
Cart and checkout
The Baymard Institute keeps a list of 50 studies on cart abandonment from 2006 to 2025. The unweighted average, as of 28 October 2026, is 70.22%, with a range of 55% to 84%. That's a mean across very different studies and not a target for your shop. Your own funnel tells you more.
The reasons are more useful. Baymard surveyed US online shoppers and set aside the answer "just browsing" (42%). The remaining reasons, each as a share. The page gives no sample size or base, so read the percentages as a direction:
| Reason for abandoning (Baymard, US) | Share | What you can check |
|---|---|---|
| Extra costs too high (shipping, tax, fees) | 40% | Show shipping early, total price in the cart |
| Delivery too slow | 20% | State a date, offer express |
| Didn't trust the site with card details | 19% | Familiar payment methods, company details, trust marks |
| Had to create an account | 18% | Offer guest checkout and show it early |
| Checkout too long or complicated | 17% | Count the fields and steps |
| Errors or crashes | 17% | Errors per step and device |
| Returns policy unsatisfactory | 13% | State period and cost clearly |
| Total cost not visible up front | 12% | Calculator or estimate in the cart |
| Card declined | 10% | Offer an alternative method |
| Not enough payment methods | 9% | Check methods for your market |
These figures are for the US, and payment habits vary by country. Pay-by-invoice in Germany or Bizum in Spain aren't on this list, yet they may matter there. The overall direction is still useful: costs and forced accounts are things you control.
Recovering abandoned carts
Reminder emails are a common tool, but legally sensitive in Europe. Have it checked before you start, because much depends on the individual case.
Two examples. In Germany, Section 7 of the Unfair Competition Act (UWG) treats email advertising without prior express consent as an unacceptable nuisance. The exception in paragraph 3 requires, among other things, that you got the address in connection with a sale, advertise your own similar goods, that the customer hasn't objected, and that you point out the right to object at collection and with every use. Whether an abandoned cart qualifies isn't answered in the text. In Spain, Article 21 of the LSSI law is similar: promotional email needs a request or express permission, except with an existing contractual relationship, lawfully obtained contact details and similar products of your own, and recipients must be able to object simply and free of charge. It doesn't expressly settle the abandoned cart either. In other countries, check your own rules.
What you can do without email: save the cart for returning visitors, remind them in the shop, and use remarketing through channels with consent.
Site search and filters
Check in GA4 whether your site search is being recorded, and read the terms with no results. They show what customers expect: missing products, different words, typos. Filters should match the questions customers ask, such as size, delivery time and price. For SEO, each filter raises its own question: should this combination be an indexed page or not? There's more in our guide to SEO for e-commerce.
Existing customers
We didn't find a source for this article for the often-quoted rule that a new customer costs many times more than an existing one, and it may not hold for your shop. Measure it yourself: what share of buyers place a second order, and after how many days at the median? That gives you the right moment for a post-purchase email. Loyalty programmes make most sense where you can expect repeat purchases at all, and mattresses aren't coffee. More in our entry on email marketing.
Organic traffic
In our view, category pages come before the blog for a shop, because that's where buying intent sits. In Search Console, see which categories have impressions at lower positions, then give them text that helps people choose, plus internal links. This takes months and doesn't belong in the week before a campaign. If you can't carry it yourself, there's our e-commerce SEO service.
Paid channels
Google Shopping and search ads bring visitors quickly, and the accounts show whether it pays, provided measurement is right. The return signal decides: do purchases reach Google Ads cleanly and without duplicates? Compare conversions in the account with purchase in GA4 and with orders in your shop system. The three won't match exactly. Large unexplained gaps should be cleared up before a campaign. Our Google Ads service can help.
Three weeks before Black Friday
In 2026, Black Friday falls on 27 November and Cyber Monday on 30 November. For SEO it's too late: what goes online today will hardly rank by then. What makes sense now acts immediately and carries little risk:
- Measure the funnel. Check all events and note today's drop-off per step, so you can compare the campaign period.
- Check shipping and delivery dates. Do they hold under load? A broken promise costs more than an honest longer time.
- Test checkout on a phone. Make a real payment with each method.
- Settle capacity. Talk to your hosting and payment providers, and freeze larger changes before the date.
- Prepare the campaign page. If you have one, use the same address every year and update it, not a new one.
- Check price claims. Reduced prices come with rules you should clear legally before the campaign.
- Use your email list. It's the channel you can still reach reliably, within your consents.
This list is our judgement, not a measurement.
Frequently asked questions
What is a good conversion rate for an online shop?
We don't give a general figure, since it depends on sector, price, channel, device and what's being counted. A comparison with your own earlier period, split by channel and device, is more useful. Check too whether missing consent means you're undercounting sessions or purchases.
What is the average cart abandonment rate?
The Baymard Institute gives 70.22% as an unweighted mean of 50 studies from 2006 to 2025, with a range of 55% to 84%. For your shop, your own rate per step in GA4 matters. A comparison with the mean says little, because the studies measure differently.
Does a shop have to offer guest checkout?
We haven't checked a general legal duty for this article. At Baymard, 18% of surveyed US shoppers name forced accounts as a reason to abandon. If you require one, test removing it and show guest checkout early. In our small sample, it wasn't visible before the checkout at any shop.
How long is the right of withdrawal for online purchases in the EU?
According to the Your Europe portal, you generally have 14 days to change your mind on online purchases in the EU, with no reason needed. Shops in our sample stated 30 or 60 days, more than that period. Exceptions and details should be checked legally.
Is free shipping above a minimum worth it?
That depends on your margin and your current order value. In our sample, 3 of 6 shops named a threshold, €49 or €79. Test a threshold just above your current average and measure order value and margin, because a higher order may only cover the shipping.
Sources
Read on 5 November 2026.
- Baymard Institute: Cart Abandonment Rate Statistics (updated 28 October 2026)
- Google Analytics: E-commerce purchases report
- Google Analytics: Funnel exploration
- Google: E-commerce events in GA4
- Google Merchant Center: Shipping
- Google: Merchant listing structured data
- Germany: Section 7 UWG (in German)
- Germany: Section 6 PAngV (in German)
- Spain: Law 34/2002 (LSSI), Art. 21 (in Spanish)
- Your Europe: Returns and the right of withdrawal
How this article was made
This article was first published in November 2016. It was rewritten in November 2026 with AI assistance, from the sources linked above and our sample, and reviewed before publication. We don't name the shops, and we hold the full sample with screenshots.
