Ecommerce Customer Retention for Stores on Google Ads

A store that grows on Google Ads brings buyers back mostly outside the ad account: order follow-up emails, service, loyalty perks and products worth reordering. Inside Google Ads, it can mark past buyers with Customer Match lists and use lifecycle goals in Performance Max. Moving budget from new buyers to past buyers is rarely the first move. In our data, return on ad spend (ROAS) barely moved as spend grew, and ad rank rather than budget capped 110 of 119 stores (GetProfit data, June 2025 – June 2026).

Each section of this guide answers one question briefly and links to the article that covers it in full.

What does customer retention mean for an online store?

Customer retention is a store’s ability to get people who bought once to buy again. A subscriber leaves on the day they cancel. In a store without subscriptions, a buyer quietly stops ordering, and you see it only in hindsight.

Shopify’s guide gives the common formula for the customer retention rate: customers at the end of a period, minus new customers acquired in it, divided by customers at the start (14 Customer Retention Strategies for Ecommerce). In a store, “customers at the end” depends on how long a buyer may go without an order before you count them as lost. So the window you choose sets the number, and two stores can compare results only when they use the same window.

That is why you read retention through several numbers, each answering its own question:

Only your order data can fill three of the four rows. That is the first thing to know about retention in a store that sells through ads.

What your ad data shows about returning buyers, and what only your orders show

Google Ads records a purchase only when it follows an ad interaction within the conversion window. According to Google Ads Help (About conversion windows), the default window is 30 days, and you can extend it to 90 days at most, depending on the conversion source. Google Ads leaves out any conversion that comes after the window.

Take a buyer who clicked an ad in March and ordered again in July after your newsletter. For your store, that is a returning buyer. For Google Ads, the July order isn’t linked to the March click.

Google can split purchases into new and returning buyers, but only after you set it up. The “New vs. returning customers” segment appears once you activate a new customer acquisition goal and optimise for purchase conversions (Measure your lifecycle goals campaigns). With Google’s auto-detection, a new customer is someone who hasn’t purchased in the last 540 days.

Google Analytics 4 (GA4) answers a different question. In the GA4 reference (API dimensions and metrics), a returning user is anyone with one or more previous sessions. A visitor who came back three times and never bought counts as “returning” there.

Our own data has the same blind spot. We build it from the Google Ads accounts of the stores we analyse, so it holds no repeat purchases and no lifetime value. That is why this guide gives no benchmark for the repeat purchase rate: any such number from us would be invented.

To read retention over time, use a cohort table, which groups buyers by the month of their first order. The guide to cohort analysis for online stores shows how to build one from an export and how to tell a seasonal cohort from weak retention.

Is keeping a customer always cheaper than winning a new one?

That is the usual argument for retention. A 2014 Harvard Business Review piece (The Value of Keeping the Right Customers) says acquiring a new customer is five to 25 times more expensive than retaining one. It adds a caveat: “depending on which study you believe, and what industry you’re in”. The same piece cites research by Frederick Reichheld of Bain & Company: raising retention rates by 5% raises profits by 25% to 95%.

Neither number tells you what to do with the next unit of your Google Ads budget. For a store on Google Ads, the useful question is narrower: can you still buy more orders at a return you can afford? If you can, every amount you move to past buyers comes out of orders you could have had.

Most stores in our data still had room to grow ad spend

Two observations from our data point that way.

More spend went with steady ROAS. Across 1,360 store-months, the rank correlation between the change in spend and the change in ROAS was +0.026, which is practically zero. We measure each change against the median store in the same calendar month, so season and holidays cancel out (GetProfit data, June 2025 – June 2026). For the full breakdown, see budget increases across 1,360 store-months.

Clicks got more expensive as spend grew, but revenue grew with spend, and ROAS barely moved.

Ad rank capped far more stores than budget did. Across 125 stores with at least 120 days of data, the median store won 45.8% of the search impressions it was eligible for, which is its search impression share. The median share lost to ad rank was 50.2%, and to budget 13.6%. Of the 119 stores where we could measure both losses, ad rank capped 110 (92%) and budget capped 9 (GetProfit data, June 2025 – June 2026).

One limit applies to these figures. Impression share describes only the auctions your ads were already eligible for, and your catalogue and budget shape that pool. It is not the size of the market, and for Performance Max it is incomplete.

Read all of this as observations, not an experiment. A store that raises spend may be one where sales were already picking up. Our ROAS also counts revenue, not profit, because we have no product costs.

What follows for the budget question:

• For a campaign that loses impressions to ad rank, the lever is the target and the bid, as set out in lost impression share: budget versus rank. Shifting money between new and past buyers leaves that cap in place.
• A store where more spend keeps ROAS steady has room to grow in the ads it already runs. Any budget it moves to past buyers comes out of orders it could still buy.

There is also rarely a separate budget to move. Performance Max runs in 141 of 146 stores in our data and takes a median 95.7% of their ad budget (GetProfit data, June 2025 – June 2026). One campaign serves new and past buyers, and the real choice is how much it should bid for each of them. That is a setting, not a budget split.

When does budget for past buyers make sense?

1. Acquisition is at its limit. Your ROAS on new orders is close to break-even, so more spend buys orders at the edge of profit.
2. Repeat orders are frequent enough to change the maths. Your export shows that a new buyer is worth noticeably more than the margin on their first order, because buyers come back.
3. You can name the people. You hold emails or phone numbers for enough lapsed buyers to build a Customer Match list.
4. You can measure the result separately. You can count re-engaged buyers on their own, apart from the campaign’s total ROAS.

Google presents its own retention setting as an addition to acquisition. Re-engagement works only in a Performance Max campaign that already uses the New Customer Value setting. Google’s best practice is to increase the budget by at least 20% at the start to make room for lapsed customers (About customer retention).

How can Google Ads bid differently for new, returning and lapsed buyers?

Google calls these settings customer lifecycle goals. According to Google Ads Help (About customer lifecycle goals), they work like this:

Two details matter for a store. Google recommends New Customer Only just for strict acquisition budgets or non-purchase goals such as leads. If you use it, Google says to run a separate campaign for existing customers. Re-engagement relies on you to name lapsed buyers: in that mode, Google identifies them only from lists you upload (How customer lifecycle goals detect customer segments).

You also set the extra value yourself. For re-engagement, Google asks what a lapsed customer is worth to your business. It notes that this value should usually sit between your typical new customer value and existing customer value. That number should come from your order data: what a returning buyer adds in margin over the following months.

The alternative is to leave the goals off and lower your target ROAS to account for repeat orders. LTV-adjusted targets versus the retention goal compares the two and shows how to tell whether either paid off.

To judge a retention campaign, Google suggests the re-engagement ratio: lapsed customers re-engaged, divided by all customers the campaign brought. The “Re-engaged customers” column appears once the setting is on.

Do you need a separate remarketing campaign next to Performance Max?

Not necessarily. Performance Max serves across all Google Ads inventory from one campaign, including Search, YouTube, Display, Discover, Gmail and Maps. Audience signals guide its learning (About Performance Max campaigns).

A separate remarketing campaign earns its place only if it brings sales you wouldn’t get anyway. Proving that takes a test, not a report, as explained in remarketing next to Performance Max.

Two narrower cases have their own guides:

• Ads that show past visitors the products they viewed need a working tag and a clean feed. For setup, and for why ads sometimes show the wrong products, see dynamic remarketing for online stores.
• In Search and Standard Shopping, you can bid up past visitors or exclude recent buyers with audience lists. For how long to keep people on each list, see RLSA and excluding recent buyers.

How does Customer Match turn your buyer list into a signal?

Customer Match lets you upload customer data you have collected and reach those people on Search, the Shopping tab, YouTube, Gmail and Display (About Customer Match). For a store, it links your order data to your ad account.

What to know before you upload anything:

• Lists age out. A membership added or refreshed more than 540 days ago stops counting. A list stays eligible only if it has at least 100 members added or updated in that time. Google recommends refreshing lists regularly.
• The route has changed. Google recommends Data Manager or the Data Manager API for Customer Match and advises against the Google Ads API for new integrations.
• Small lists are weak. For lifecycle goals, Google warns that a list with fewer than 1,000 members may hurt the campaign’s ability to run new customer acquisition.

The guide to Customer Match for online stores covers the requirements, small lists and how Customer Match differs from enhanced conversions.

Which buyers should you bring back first, and which products start repeat orders?

Some past buyers are worth more effort than others. Three tools help you choose.

1. RFM segments. You split buyers by how recently they bought, how often they bought and how much they spent. A spreadsheet is enough, and the result tells you what to send each group and whom to put on which list: RFM analysis for online stores.
2. Cohorts. You see whether buyers from recent months return more or less often than earlier ones, which tells you whether retention is improving.
3. First-order products. Some products start long relationships, others bring one-off buyers. Google Ads reports each product’s results from the orders it records, so a product with a modest ROAS may still be worth its spend if its buyers return. How to find those products and campaigns in your data: first orders that lead to repeat purchases.

What brings buyers back outside the ad account?

Most retention work happens outside Google Ads. Shopify’s list of 14 strategies starts with customer accounts, better service, a loyalty programme and lifecycle emails. It goes on to returns, subscriptions, referrals and buy now, pay later.

Two of them connect directly to your ads:

• Email and messages. They reach past buyers, and you pay nothing per click. Which flows to set up first is in email and SMS for online stores. The same guide shows how to check that they add orders rather than relabel orders your ads brought anyway.
• Loyalty and referral programmes. Points, tiers, paid membership and referral rewards all cost margin. Members can also feed the Loyalty Program Members setting in Google Ads. Loyalty and referral programmes sets out what each format costs and what your margin can carry.

How repeat orders change what a new buyer is worth

Example store, not client data.

A tableware shop spends 40,000 a month on Google Ads, which reports 300 orders and 180,000 in revenue. Its average order value is 600 and its cost per order is 133. The gross margin is 35% of revenue, so each order leaves 210 in gross profit, before delivery and other costs.

The shop’s order export adds two facts that Google Ads doesn’t hold:

• 240 of the 300 orders came from new buyers. If you charge the whole spend to them, each new buyer costs 40,000 ÷ 240 = 167. That is an upper bound, because part of the spend also brought the 60 repeat orders.
• 20% of new buyers place a second order within 12 months, mostly after an email, at about the same order value. Over a year, each new buyer then brings 210 + 0.2 × 210 = 252 in gross profit.

Repeat orders nearly double what each new buyer leaves after ad cost, from 43 to 85. Google Ads reported the 300 orders, but a second order placed months later after an email often falls outside its conversion window. The number that would justify paying more for a new buyer lives in the shop’s export. The guide to what an LTV:CAC ratio means works out how big a loss on the first order that number can cover.

Check what Google counts before you set up any retention setting

Most retention settings in Google Ads run on your purchase conversions. New Customer Value and Re-engagement need at least one Purchase conversion goal. Google’s auto-detection of returning buyers relies on the purchases you track. If Google counts an order twice, or every order carries the same fixed amount, both the new-versus-returning split and the extra value you assign rest on wrong numbers.

The portal’s Conversion check looks at a 90-day window. It shows whether your ads learn on a purchase, whether the real order amount arrives with each order and whether two goals count the same order. It also lists orders whose amount is many times your usual order value.

What your ads are really learning on. An order counted twice. An amount substituted for the real one. A payment several times larger than your usual basket. The portal checks what stands behind them. The portal changes nothing without your consent.

Check your conversions →

What to do this quarter

1. Export your orders for the last 12–24 months with a customer ID, date and amount. Every retention number in this guide comes from that file.
2. Calculate the repeat purchase rate and the median number of days between the first and second order. These two numbers tell you whether retention is worth working on.
3. Build a cohort table by first-order month, so you see whether retention improves.
4. Write down what a returning buyer adds in gross profit over 12 months. This is the value you can later give Google.
5. Check your conversion tracking. Make sure each order counts once, with its real value, and that your shopping campaigns learn on a purchase.
6. Find your ceiling. If campaigns lose impressions to ad rank rather than budget, the lever is the target, not a shift of money to past buyers.
7. Start retention where you pay nothing per click: email and messages to past buyers.
8. Test re-engagement only with a list and a value. Use a Customer Match list of lapsed buyers in a Performance Max campaign, and judge it by re-engaged buyers, not by the campaign’s total ROAS.
9. Refresh your Customer Match lists regularly. Memberships older than 540 days stop counting.

Sources

• About customer retention — re-engagement only in Performance Max; prerequisites: Performance Max, the New Customer Value setting, Customer Match lists of lapsed customers; lapsed customer value between new and existing customer value; the re-engagement ratio and “re-engaged customer” column; the best practice of a budget increase of at least 20%. Checked 2 October 2026.
• About customer lifecycle goals — the settings, eligible campaign types, bid strategies and purchase goal requirements; New Customer Only for strict acquisition budgets or non-purchase goals, with a separate campaign for existing customers; Loyalty Program Members. Checked 2 October 2026.
• How customer lifecycle goals detect customer segments — lapsed customers identified only from uploaded lists; auto-detection over up to 540 days of tracked purchases; lists under 1,000 members. Checked 2 October 2026.
• Measure your lifecycle goals campaigns — the “New vs. returning customers” segment, available only with an activated lifecycle goal on purchase conversions; new customers under auto-detection. Checked 2 October 2026.
• About Customer Match — where Customer Match serves; the 540-day membership limit and the 100-member minimum; Data Manager as the recommended route. Checked 2 October 2026.
• About conversion windows — the 30-day default window, up to 90 days by source, conversions after the window not recorded. Checked 2 October 2026.
• About Performance Max campaigns — one campaign across all Google Ads inventory; audience signals guide learning. Checked 2 October 2026.
• Lifecycle goals, Google Ads API — customer acquisition, customer retention and loyalty retention goal types; a retention goal requires a Performance Max campaign and a Customer Match list. Checked 2 October 2026.
• API dimensions and metrics, Google Analytics Data API — the “New / returning” dimension: returning users have one or more previous sessions. Checked 2 October 2026.
• 14 Customer Retention Strategies for Ecommerce — the customer retention rate formula; the list of 14 retention strategies. Checked 2 October 2026.
• The Value of Keeping the Right Customers — acquisition five to 25 times the cost of retention, with the author’s caveat; Reichheld’s 5% retention and 25–95% profit figure. Checked 2 October 2026.
• GetProfit data: month-to-month changes in 1,360 store-months, June 2025 – June 2026, each measured against the median store in the same month — spend, revenue, ROAS and cost per click.
• GetProfit data: 125 stores with at least 120 days of data, June 2025 – June 2026 — impression share and impressions lost to rank and budget.
• GetProfit data: 146 stores, June 2025 – June 2026 — share of budget and number of stores by campaign type.
• GetProfit portal, Conversion check — the checks it runs and its 90-day window.