Retargeting for Indian D2C Brands: Costs, Results and Meta's Controls

What do retargeting ads cost and return in India?#

Meta retargeting costs about ₹235 per 1,000 impressions and reports a ROAS of about 7.75x, while Google remarketing reports 6.1x. These benchmarks come from ProfitBox360 research, 2025 (client data and shopper surveys).

Retargeting costs ₹25 more per 1,000 impressions than the Meta average, because small audiences cost more to reach. In return, every impression goes to someone who already knows the brand.

How retargeting and prospecting add up to 3.65x#

Weight the two ROAS figures by spend and they blend back to the all-Meta average.

All-Meta ROAS = retargeting share × retargeting ROAS + prospecting share × prospecting ROAS

21% × 7.75 + 79% × 2.56 = 1.63 + 2.02 = 3.65x

Per ₹1,00,000 of Meta spend, that looks like this:

Retargeting's ₹21,000 reports ₹1,62,750, not far short of the ₹2,02,240 from nearly four times the spend. That's why moving budget from prospecting to retargeting looks attractive on the dashboard. The next section shows why it usually isn't.

Size the budget to the audience#

Match the retargeting budget to the number of people you can reach, because extra money past that point buys repeat impressions rather than new people. Meta's frequency cap and target frequency settings are documented for reservation campaigns. In a sales campaign, you control frequency mainly through audience size and budget.

Weekly retargeting budget = people reached × frequency × CPM ÷ 1,000

Illustrative example: a brand retargets its website visitors and cart adders from the last 30 days.

• Meta reaches 40,000 people from this audience each week
• Frequency and CPM follow the benchmarks: 4.25 a week and ₹235

40,000 × 4.25 × ₹235 ÷ 1,000 = ₹39,950 a week

Spend more than that on the same audience and frequency climbs instead of reach. If the same budget reached only 20,000 people, each would see the ads 8.5 times a week.

Put most of the budget on the newest visitors. In a field experiment with an online home-improvement seller, 33% of the first week's retargeting effect came on the first day.

Google remarketing#

Indian D2C brands see an average Google-reported ROAS of 6.1x on remarketing campaigns. Google and Meta credit sales under different attribution rules, so don't rank the two platforms on these figures alone.

How much of retargeting's ROAS is really incremental?#

About 30%: lift tests on Indian D2C brands' Meta retargeting find that a reported 7.75x is worth about 2.3x in sales the ads actually caused.

Incremental ROAS = reported ROAS × incremental share

7.75 × 30% = about 2.3x

Of every ₹1,00,000 of revenue Meta credits to retargeting, about ₹30,000 came because of the ads. The other ₹70,000 came from people who would have bought anyway. On the ₹21,000 of retargeting spend above, ₹1,62,750 of reported revenue shrinks to ₹48,825 of extra sales.

The 2.3x sits below even the 2.56x that prospecting reports. Prospecting's figure also includes sales that would have happened anyway, so test both before moving money. Then compare each campaign's incremental ROAS with your break-even ROAS: 1 ÷ CM2, on the revenue you keep.

The gap is large because retargeting reaches people who have already visited, added to cart or bought. Many of them were coming back anyway. Meta's default attribution credits an ad for a purchase within 7 days of a link click or 1 day of a view, whether or not the ad changed anything.

What published experiments found#

Controlled experiments outside India show that retargeting does cause extra visits and sales. They also show that measures without a control group can overstate what advertising does.

Read these lifts as an addition to what people would do anyway. A 10.5% lift in purchases means the ads added about 10 purchases for every 100 the same people would have made without them.

The 30% benchmark measures something different: the share of Meta-credited revenue that the ads caused. Use it when you judge a retargeting campaign's ROAS.

Meta's incremental attribution column#

Use Meta's incremental attribution for a quick estimate between lift tests. Meta describes it as a model that predicts whether a conversion was caused by an ad.

• To view it: in Ads Manager, open the Columns: Performance menu, select Compare attribution models, choose Incremental and click Apply. Results aren't available for dates before 1 April 2025.
• How to read it: Meta designed it to show relative incrementality between Meta campaigns. Put your retargeting and prospecting campaigns side by side, both in this column.
• To optimise for it: in the ad set, choose Website as the conversion location, then Maximise number of conversions or Maximise value of conversions. Click Show more options and set Attribution model to Incremental. Bid controls aren't available, and the model can't be changed after publishing.

It's a model's prediction, not an experiment. Confirm large budget moves with a lift test.

How do you test retargeting lift yourself?#

Run Meta's Conversion Lift if your account qualifies. If it doesn't, hold back half of a customer list and compare store orders from each half.

Meta Conversion Lift#

Conversion Lift splits people into a test group that can see your ads and a control group that can't. The difference in conversions between the two groups is the lift.

• You create the test in Experiments, and the test itself costs nothing extra.
• As a guide, the ad account needs a campaign started in the past year with at least USD 5,000 of spend and 500 conversions.
• The account must also meet a signal-quality rule, such as sending Conversions API events with an Event Match Quality score above 5.

Google Ads offers Conversion Lift too, based on users or on geography. It isn't available to every account, so ask your Google account representative.

Test with a customer-list holdout#

1. Export past buyers with their phone numbers and emails. Split them into two equal halves at random, for example by odd and even customer ID.
2. Upload each half as its own customer-list custom audience.
3. Target half A in a retargeting ad set. Exclude half B from every Meta campaign, including Advantage+ sales campaigns.
4. Run for four weeks without other changes aimed at either half.
5. Count store orders from each half by phone number or email, using Shopify rather than Meta.

Incremental orders = orders from half A − orders from half B

Illustrative example: a four-week holdout on a customer list.

• 40,000 past buyers, split into two halves of 20,000
• ₹40,000 of retargeting spend on half A
• Every order is worth ₹1,000
• Meta reports 280 purchases from the campaign
• Shopify shows 1,210 orders from half A and 1,140 from half B

So ₹2,10,000 of the ₹2,80,000 Meta credited would have come without the ads.

A gap of a few dozen orders can come from chance. If the difference is small, run the test again before you cut or raise spend.

This method works only for people you can identify in store data. For anonymous website visitors, use Conversion Lift.

How much control do Advantage+ campaigns give over retargeting?#

Enough to exclude past buyers and cap retargeting spend, but not with one setting: Meta has retired the existing-customer budget cap.

The cap let Advantage+ campaigns limit the share of budget spent on existing customers. Meta's Business Help Centre now says it's "no longer available" and shows how to do the same thing manually. Meta's developer notes say the cap isn't available with Advantage+ campaigns. Since 18 February 2026, Meta's Marketing API can no longer create the old Advantage+ shopping campaigns that the cap belonged to.

Advantage+ sales campaigns spend a large share on people who already buy from you. When nothing limits spend on existing customers, 32.5% of Advantage+ sales campaign spend reaches them. That's ₹32,500 of every ₹1,00,000. It's retargeting you didn't set up, and its sales sit inside the campaign's ROAS.

Meta's controls that matter for retargeting#

Meta defines the segments this way:

• New audience: people who haven't interacted with your products or services.
• Engaged audience: people aware of your business or who have interacted with your products, but haven't bought.
• Existing customers: people who have bought or signed up.

You build the engaged and existing segments from custom audiences, such as website visitors, customer lists and catalogue activity.

Cap retargeting the way Meta now recommends#

Replace the cap with one campaign that has two ad sets, following Meta's own steps.

1. Create a sales campaign and choose Campaign budget as the budget strategy.
2. In ad set 1, open Ad set spending limits and set a daily or lifetime maximum as a percentage.
3. Under Audience, select Further limit the reach of your ads. Include your existing-customer custom audiences and untick Use as a suggestion.
4. Duplicate the ad set, keeping the same creative.
5. In ad set 2, swap the inclusion for a custom audience exclusion of the same audiences, under Controls, and leave it without a spending limit.

A percentage limit scales with the campaign budget; a fixed amount stays the same. To spend nothing on existing customers, run ad set 2 on its own. Set the limit from your lift test. Until you have one, keep it at or below the 21% benchmark share.

Google remarketing in Performance Max and Demand Gen#

Use Demand Gen when you need strict remarketing, because Performance Max treats your lists only as hints.

• Your data segments: this is Google's name for remarketing lists, covering website visitors, app users, customer lists and YouTube users. A segment needs at least 100 active users in the last 30 days to serve. Membership lasts 30 days by default and up to 540 days.
• Performance Max: your data segments work only as audience signals. Google says Performance Max may show ads to people outside your signals when they're likely to convert. So a Performance Max campaign can't be limited to remarketing.
• Keeping existing customers out: use the customer acquisition goal set to "Only bid for new customers", or "Bid higher for new customers than existing ones". It works in Performance Max, Search, Shopping and Demand Gen. Define existing customers with a customer list and the website tag.
• Lapsed customers: the customer retention goal lets you bid higher to re-engage lapsed customers you define, using Customer Match lists.
• Demand Gen: you can target your data segments directly. Optimised targeting is on by default and looks beyond the segments you pick, so turn it off in ad group settings for strict remarketing. It won't show ads to people you've excluded with customer data segments.
• Display campaigns: from June 2026, eligible advertisers can move Display campaigns into Demand Gen with a migration tool. Later, new campaigns can only be created in Demand Gen, and remaining ones will move automatically.

FAQ#

How long should a retargeting window be?#

Keep most retargeting spend on people who visited in the last 7 days, and spend less on older visitors. In a field experiment with an online home-improvement seller, 33% of the first week's retargeting effect came on the first day. The effect faded as time since the visit grew. Extend the window only if a holdout test supports it.

Should past buyers be excluded from retargeting?#

Yes, until they're due to reorder; then test whether ads bring them back sooner. In eBay's paid search experiments, frequent users' buying wasn't changed by ads, while new and infrequent users responded. Meta always treats a custom audience exclusion as a hard control, so it reliably keeps recent buyers out of an ad set.

Are product ads better than brand ads for retargeting?#

Not always. A field experiment at an online travel firm found that ads showing the exact products people had viewed worked worse, on average, than generic brand ads. Product-specific ads caught up once shoppers showed signs of narrowing their choice, such as visiting review sites. Test both creative types on browsers and cart abandoners separately.

What retargeting frequency is too high?#

Indian D2C brands average 4.25 retargeting impressions per person every 7 days on Meta. A weekly frequency far above that, with falling results, usually means the budget is too big for the audience. Lower the budget, widen the audience window or add new creative, rather than keep paying for repeat impressions.

Does excluding past buyers stop Advantage+ from retargeting?#

No. Excluding existing customers removes only buyers, so an Advantage+ sales campaign can still reach website visitors and cart abandoners. To keep a campaign on new people, also exclude the custom audiences behind your engaged audience, such as website visitors. Then check the audience segments breakdown to see where the spend went.

Is Meta's incremental attribution the same as a lift test?#

No. Incremental attribution is a Meta model that predicts which conversions an ad caused. Conversion Lift is an experiment with a control group that can't see the ads. Use the incremental column to compare retargeting with prospecting week by week. Use a lift test to confirm the result before moving large budgets.

Can a small brand test retargeting without Conversion Lift?#

Yes, with a customer-list holdout. Meta's guide for Conversion Lift asks for a campaign with USD 5,000 of spend and 500 conversions in the past year. Below that, split your customer list at random, show retargeting to one half only and compare store orders from each half over four weeks.