From $1M to $2M a month, you need a new-customer engine.

Why does a brand doing $1.2M a month at a 4.0 MER feel like it is working harder every quarter for the same growth?

Usually because the 4.0 is not describing what the founder thinks it describes. At this size, the brand has tens of thousands of past customers. They open the emails, search the brand name, click the retargeting ad, and buy. The ads get credit for all of it.

That is not waste, exactly. Those customers are the business. But it means the headline number is increasingly a measure of the customer base, and the part of the account that is supposed to grow the base is buried inside it.

The short version: run acquisition as its own business

The rule for this rung: stop judging acquisition on blended numbers. Measure new-customer CAC, ad spend divided by genuinely first-time buyers, against your own customer list. Give it a ceiling. Then use the platform settings built for this, and design offers that bring new people in without paying your loyal customers to do what they would have done anyway.

How returning customers turn a $60 CAC into a $24 cost per order

Here is an illustrative month. This is arithmetic, not a client result. The brand makes $1.2M: $750,000 from 7,500 returning-customer orders at $100, and $450,000 from 5,000 new customers at a $90 first order. It spends $300,000 on ads. Contribution margin is 45%.

Same spend, same month, two different businesses. The blended view says every order costs $24. The new-customer view says each new customer costs $60 and starts $19.50 underwater. Only the second one tells you what growth costs. The full method for building this split is in what a new customer really costs.

Where the 5,000 new customers actually come from

Split the same month by campaign type and the picture gets sharper. Still illustrative arithmetic, and the reported ROAS figures are each platform’s own claim, which overlap, as the second-channel rung explains.

The campaigns with the best reported ROAS buy the most expensive new customers. Retargeting and brand search look brilliant because most of the people they reach have bought before. Prospecting looks weakest and does most of the acquiring. Rank this account by ROAS and you would cut the engine and feed the reminder.

The decision rule: a $62.50 new-customer CAC ceiling

A new-customer CAC means nothing without a ceiling. Build yours from two numbers: first-order contribution, plus the contribution your own cohort data shows new customers add within the payback window you can finance.

In the example, first-order contribution is $40.50. Say the brand’s cohorts show an average new customer adding $22 of contribution in repeat orders within 90 days, and the business can finance a 90-day payback. The ceiling is $62.50. Shopify’s customer cohort analysis report is a practical place to read how each month’s new customers come back, and the CAC payback calculator turns the result into months.

• →Below the ceiling: the campaign is acquiring at a price the business can carry. Give it budget in steps and watch the marginal cost.
• →Above the ceiling, and it is an acquisition campaign: cap its budget until creative, offer, or targeting brings the cost down.
• →Above the ceiling, and it is retargeting or brand: do not judge it as acquisition at all. Judge it on whether it is incremental, which needs a holdout test, not a CAC.

Prospecting at $51 clears the $62.50 ceiling. Retargeting and brand at $100 do not, and that is fine, because they are not the engine. The mistake is letting their blended ROAS decide how much the engine gets.

One more discipline: the ceiling applies at the margin. Prospecting that averages $51 can still be buying its last thousand customers at $75. When you add budget, divide the extra spend by the extra new customers, the same step test used in the doubling arithmetic, and stop adding when that marginal figure reaches the ceiling, even if the average still looks comfortable.

What $2M a month asks of the engine: 11,667 new customers

Run the rung’s target through the arithmetic of doubling revenue. Suppose returning revenue grows from $750,000 to $950,000 by the target month. The other $1.05M has to come from new customers at a $90 first order: about 11,667 of them, 2.3 times today’s 5,000.

At the $62.50 ceiling, that is up to about $729,000 a month of acquisition spend. The engine has to more than double its output while every extra customer costs more than the last. That is why this rung is about acquisition as a system, and why the rung above it is about margin and AOV: once the engine is running, the next constraint is how much each customer is worth.

Google: three customer acquisition settings

Google Ads has customer lifecycle goals built for exactly this. Under customer acquisition, Google lists three settings, each available in a different set of campaign types among Search, Performance Max, Shopping, and Demand Gen.

• →New Customer Value: bid higher for new customers than existing ones. It uses value-based bidding, Target ROAS or maximize conversion value, and needs at least one purchase conversion goal. Google recommends it for all advertisers with purchase conversion goals.
• →High Value New Customer: bid higher still for new customers who resemble your most valuable ones, in Performance Max and Search.
• →New Customer Only: bid only for new customers. Google recommends it only if you have strict acquisition budgets, and advises a separate campaign to reach existing customers when you use it.

All three depend on how well Google can tell your customers apart. Google detects new customers automatically by default, and its documentation lists signals you can add, including an uploaded customer list and a conversion tag that reports whether the buyer is new. A longer, cleaner list means fewer returning customers mistaken for new ones.

Meta: split new and existing customers into two ad sets

Meta used to offer a simple dial for this: an existing customer budget percentage that capped how much of an Advantage+ shopping campaign could be spent on people you had defined as existing customers. Check your setup against the current documentation. Meta’s developer guide now says that field is not available for new Advantage+ campaigns.

Meta’s suggested replacement is structural. Build a custom audience that defines your existing customers. Set the budget at campaign level with Advantage campaign budget. Then create two ad sets: one for existing customers, and one that excludes that audience so it reaches only people who have not bought. The split is now visible by ad set, which is more useful than a percentage anyway, because you can read new-customer cost directly.

The exclusion is only as good as the audience behind it. Build it from your full customer list, not only from pixel purchasers, and refresh it regularly. The practical detail of exclusions and retargeting is in Meta audience exclusions and retargeting.

Creative for people who have never heard of you

Most brands at this size make creative that works on their existing customers: new launches, restocks, the latest drop. It performs well in the platform because loyalists respond. It does very little for strangers, who do not know why they should care.

Acquisition creative starts one step earlier. It names the problem before the product, explains what makes the product different from the obvious alternative, and carries proof a skeptic will accept. And it has to be genuinely different from ad to ad, because similar ads compete for the same people. The case for difference over volume is in Meta creative diversity, and the production side is on the creative supply rung.

A quick check on your own account: once new and existing customers sit in separate ad sets, compare which concepts earn the most spend in each. If the same ads lead both, your creative is probably speaking to people who already know you, and the new-customer ad set is borrowing its winners from the loyal one.

First-order offers that do not pay loyal customers $45,000

Offers are the fastest lever on new-customer CAC and the easiest to get wrong. A sitewide code is a discount for everyone, and at this size most of everyone is returning customers.

Illustrative arithmetic: if returning customers use a 15% sitewide code on $300,000 of orders during a promotion, that is $45,000 of margin given to people who would largely have bought anyway. The new customers the promotion brought in now carry that cost too.

Count the first-order offer inside new-customer CAC. A $15 discount on a first order is $15 of acquisition cost, even though it never appears in the ad account. A campaign at $51 with a $15 welcome discount is really acquiring at $66, which in the example sits just above the ceiling.

• →Make first-order offers first-order only. Tie the code to new customers at checkout, rather than publishing it sitewide.
• →Prefer value over discount. A first-order bundle raises AOV, which lowers the new customers you need, instead of cutting the price of the order you already had.
• →In Google Shopping, Merchant Center promotions have an audience attribute with a new customers value, which Google says is shown only when it judges a shopper is likely to be a new customer.

What common advice gets wrong: cut the lowest-ROAS campaigns

When blended efficiency falls at this size, the standard response is to move budget from the weakest-ROAS campaigns to the strongest. It feels disciplined. On a mature account, it is backwards.

Look at the second table again. The strongest ROAS belongs to the campaigns reaching people who already know you. Moving budget toward them raises the headline number and lowers the number of new customers, which is how a brand ends up with a better MER and a shrinking customer base. The metric improves and the business stalls, and nobody notices for months because returning revenue takes time to fall.

The opposite overcorrection is wrong too. Excluding existing customers from everything throws away real value: loyal customers reached by an ad at the right moment do buy more. The fix is not to ignore them. It is to judge each campaign on its own job, acquisition on new-customer CAC and retention on incrementality, and never let one number referee both.

When this advice is wrong

• →Your customer list is short or starts late. If your history only reaches back a year, many returning customers will look new, and the split will flatter acquisition. State when your history begins and treat older unmatched buyers as unknown.
• →You are a replenishment or subscription brand. When reorders are the business model, reaching lapsed customers can be the cheapest growth available. Google’s retention settings exist for this, and the engine should sit alongside a reactivation program, not replace it.
• →Cash is tight. A ceiling that includes 90 days of repeat value means financing each customer for three months. If you cannot, set the ceiling on first-order contribution alone, and grow more slowly.
• →Your first-order margin is already negative before ads. Then no ceiling will work, and the constraint is product economics, not acquisition.
• →Most of your growth is wholesale or marketplace. A store customer list then misses a large part of who your customers are.

What to do this month

1. Export your full customer list with first-order dates, and write down the date your history begins.
2. Build the split for last month: new-customer CAC by campaign type, with the unknown share beside it.
3. Set your ceiling from first-order contribution plus measured repeat value inside the payback window you can finance.
4. On Google, turn on New Customer Value for one Shopping or Performance Max campaign already using value-based bidding, with a fresh Customer Match list.
5. On Meta, build the existing-customer audience and split your main sales campaign into new and existing ad sets. If you are inside five weeks of Black Friday, test the split on one campaign now and restructure the main campaign after Cyber Monday.
6. Audit your offers: find every code that returning customers can use, and work out what it cost last quarter.

This is the third rung of the ecommerce growth ladder. The split it depends on starts with your customer list. When we take on an account, we ingest the customer list and separate new customers from returning ones, so new-customer CAC is measured against the people who have actually bought from you, not against what a pixel happens to remember. Our DTC growth engagements are built around that split. If you want an outside read on your ad account first, start with the free audit.