The essentials
- Why: a returning visitor costs 5 to 7 times less than a new one acquired through advertising, and converts 2 to 3 times better.
- Measurement: GA4 supplies the Retention report, the share of returning visitors and the cohorts; the right metric depends on the site (media, e-commerce, SaaS, brochure).
- 2026 levers: renewed content, email and notifications, a customer account, personalisation, speed and accessibility, measured remarketing.
- Benchmarks: 20 to 35% returning visitors on a media site, 40 to 60% on a mature e-commerce site, 15 to 25% on a B2B brochure site.
Retaining a website's users is the cheapest growth lever and the least worked on. The budgets go to acquisition, when a returning visitor costs 5 to 7 times less than one acquired through advertising and converts 2 to 3 times better. This article gives the method: measuring retention in GA4, choosing the levers suited to the type of site, setting 2026 benchmarks and avoiding the classic mistakes. The general metrics are detailed in digital marketing KPIs.
Why retention weighs more than acquisition
Three mechanisms explain retention's weight in a site's profitability. The cost of acquiring a paid visitor rose 30 to 50% between 2021 and 2026 on Google Ads and Meta, while the cost of an email or a notification stays close to nothing. A returning visitor has already passed the discovery and trust stages: their conversion rate is higher and their average order 15 to 30% larger in e-commerce. Finally, the engines and generative AI read engagement signals (direct visits, duration, returns) as quality clues; a site with a loyal audience ranks better and gets cited more often.
| Metric | A new visitor | A returning visitor |
|---|---|---|
| Average cost of the visit (advertising vs email) | €0.50 to €2 | €0.01 to €0.10 |
| E-commerce conversion rate | 1 to 2% | 3 to 6% |
| Pages per session | 1.5 to 2.5 | 3 to 5 |
| GA4 engagement rate | 45 to 60% | 65 to 80% |
Measuring retention in GA4
GA4 offers several complementary views. The Retention report (Life cycle) shows the share of users returning at day 1, day 7 and day 30 after their first visit. The "New / Returning" dimension separates the two populations across every acquisition and engagement report. Cohort exploration follows a group of users who arrived in the same week and shows whether they come back in later weeks. Audiences let you create reusable segments such as "visitors who returned 3 times in 30 days" for advertising; the procedure is described in creating a custom audience in GA4.
| Type of site | Retention metric to track | 2026 benchmark |
|---|---|---|
| Media, blog, content | Share of returning visitors, visits per user per month | 20 to 35% returning; 2 to 4 visits a month |
| E-commerce | Repeat purchase rate at 90 days, share of revenue from returning customers | 20 to 30% repeat purchase; 40 to 60% of revenue |
| SaaS, web application | Retention at day 7 and day 30, monthly active users | 40 to 60% at day 30 depending on the sector |
| A B2B brochure site | Returning visitors, returns to the pricing and contact pages | 15 to 25% returning |
| A local site (shop, tradesperson) | Direct visits and visits through Google Business Profile, repeat calls | Few returns to the site; loyalty measured offline |
Two precautions. Visitors who refuse consent or change device appear as new: the returning share is understated by 20 to 40% depending on the consent rate. And retention is read over long periods: a weekly report is too short to judge a retention lever.
The levers that bring visitors back
- Renewed content: a reason to come back every week or every month (features, updates, tools, data). An editorial calendar kept up produces more loyalty than one much-shared piece.
- Email and notifications: a newsletter with a fixed appointment, welcome sequences, personalised alerts. The average open rate sits between 25 and 40% in 2026, and email remains the leading return channel for media and B2B.
- A customer account or personal area: history, favourites, preferences, documents. On an e-commerce site, a customer with an account returns twice as often as a guest buyer.
- Personalisation: recommendations based on pages viewed, content adapted to the sector or the location, resuming where they left off. It must stay legible and respect consent.
- Speed and accessibility: a slow or inaccessible site loses visitors at every return; the Core Web Vitals and accessibility standards are retention levers as much as SEO ones.
- Measured remarketing: retarget engaged visitors with a follow-on message, cap the frequency and exclude converters; see remarketing.
- Community and interaction: comments, reviews, online events, social media that sends people back to the site with an appointment.
Adapting the strategy to the business
Some sites aren't meant to bring their visitors back. A plumber, a solicitor or a kitchen retailer answers a one-off need: the issue is brand loyalty (reviews, recommendations, direct contact at the next need), not returns to the site. For those businesses, tracking covers the number of reviews, direct visits and brand searches, measured with Google Business Profile and Search Console. Conversely, a media site, an e-commerce site or a software product live on recurrence, and retention there deserves as much budget as acquisition.
The mistakes that cost you visitors
- Measuring loyalty by bounce rate: a loyal visitor who reads one page and leaves is a success, not a bounce; the bounce rate isn't a retention metric.
- Asking too soon: a sign-up pop-up in the first second, multiple banners, notifications requested before any value is given.
- Changing the landmarks: a redesign that upends the navigation, URLs changed with no redirect, content deleted.
- Not closing the loop: no email after a sign-up, no follow-up after a download, no response to comments.
- Retargeting without limits: remarketing with no frequency cap wears people out and damages the brand.
A 90-day retention plan
Weeks 1 to 2: turn on the measurement (the Retention report, returning-visitor audiences, sign-up and repeat-purchase events, see GA4 key events). Weeks 3 to 6: set up a content appointment and a welcome email sequence. Weeks 7 to 10: fix the speed and the friction points of the most revisited pages. Weeks 11 to 13: launch remarketing targeted at engaged but unconverted visitors, and compare the cohorts before and after. A reasonable objective over a quarter is a gain of 5 to 10 points in the share of returning visitors.
How GreenRed helps
Rather than juggling several tools, GreenRed's GA4 performance and traffic tracking brings these metrics together in a single dashboard, compares them over time and tells you which actions come first. You can try it free, with no card, from the Pricing.
Frequently asked questions
How do you measure a site's visitor retention?
In GA4, use the Retention report (returns at day 1, day 7, day 30), the New / Returning dimension and cohort exploration. On an e-commerce site, add the repeat purchase rate at 90 days and the share of revenue from returning customers. Read that data over at least a quarter.
What share of returning visitors is normal in 2026?
Between 20 and 35% for a media site or a blog, 40 to 60% for a mature e-commerce site, 15 to 25% for a B2B brochure site. Those figures are understated by consent refusals and device changes; what counts is the trend over several months.
Which retention levers give the fastest results?
Email (a welcome sequence, a newsletter with a fixed appointment) and remarketing targeted at engaged visitors produce returns within weeks. Renewed content, the customer account and speed act over several months but build sturdier loyalty.
Should a local service brochure site work on retention?
Rarely on the site itself: the need is one-off. The effort goes into brand loyalty: Google reviews, recommendations, direct contact, presence on Google Business Profile. The relevant metrics are direct visits, brand searches and the number of reviews.
Does the bounce rate measure loyalty?
No. A loyal visitor who reads one page then leaves isn't a failure. Bounce rate describes a session, not a user; loyalty is measured by returns over time, visits per user and the conversions of returning visitors.