Scaling Winners by New Customer Share: Give the Most Budget to Ads That Find New Buyers

Short answer: Sort your winning ads by what share of their purchases come from brand new customers, then place them into three Ad Sets inside one scaling campaign: 70 to 100% new, 40 to 70% new, and under 40% new. Keep the targeting identical in all three so budget becomes your steering wheel, and give the top bucket the most money. Then prove it is working with a net new reach report on that campaign.

The brands in the examples are made up.

What does scaling by new customer share mean?

Scaling by new customer share means giving the most budget to the ads whose purchases come mostly from people who have never bought from you, and less to ads that mostly resell to warm buyers.

Here is the problem. Two ads can show the same ROAS (return on ad spend). One finds new buyers. The other mostly closes people who already know you. Scale both the same way and much of the budget goes to people you would likely have sold to anyway.

Here is the solution. Sort each winner by how much of its buying comes from new people, and let that decide where the money goes. It is the thinking behind new vs returning delivery, turned into a campaign structure.

Why not just scale whatever has the best ROAS?

Because ROAS does not tell you who bought. A high ROAS ad that mostly sells to existing customers is often harvesting demand your email, organic posts or other ads already created.

Ads that find new customers often show a higher cost per purchase, because cold people are harder to convince. That is normal, so we judge them on cost per new customer instead, which ties into marginal CAC (the cost of the next customer you buy, not your average).

Step 1: How do you find which ads drive new purchases?

Sort every winning ad by its new purchase share, using the Audience segments breakdown in Ads Manager. This step is required.

  1. Set up your audience segments first. Meta can only split purchases into new and existing if you have told it who your engaged audience and existing customers are. This lives in your ad account settings. Meta renames menus often, so check its current setup.
  2. Pull purchases at the ad level. On the Ads tab, pick a date range with enough data and add the Audience segments breakdown.
  3. Work out new share. New purchase share is new audience purchases divided by all purchases. If an ad had 50 purchases and 40 came from the new audience segment, its new share is 80%.
  4. Sort your winners from highest new share to lowest.

This is Meta's own read of who is new, so check it against Shopify or a third party tool when you can. Ads with only a handful of purchases swing a lot, so set a minimum before you trust the number.

What is the One vs The Rest exclusive reach check?

It is an optional check that shows how much of an ad's reach no other ad in the account is getting. We use it before handing an ad the biggest budget, or to settle an ad sitting on a bucket line.

Exclusive reach is account reach minus the reach of everything else. The quickest way to pull it: in Ads Manager, set your date range and read the account's total reach. Then filter to campaign (or ad) name does not contain the one you are checking, and read the reach on the total row. Account reach minus that number is the reach only that campaign or ad delivered. Divide it by its own reach to get its exclusive reach share. One slice on its own tells you little, so run the same read for two or three campaigns or ads side by side, in the same window, and skip anything that has been live for less than two weeks, since a newer ad has not built up enough reach to read fairly.

Two more reads round it out. Overlap across the account: add up each campaign's reach, subtract the account's unique reach, and divide the difference by that sum. Our rule of thumb is under 20% is healthy, 20 to 40% is acceptable at scale, and above that your campaigns are paying to reach the same people. Like the goals below, these lines can shift by brand and product. Overlap inside the scaling campaign: add up the three Ad Sets' reach and compare it to the campaign's unique reach the same way. A low number means the buckets mostly reach different people, even with identical targeting. Our guide to Ad Set overlap walks through the math.

One watch out: geography. A campaign that runs in a country no other campaign targets will show exclusive reach from geography alone, so compare campaigns that target the same places.

New purchase share Exclusive reach share What it means
High High A true new customer finder. Give it the biggest budget.
High Low Sells to new people inside a pool other ads also reach.
Low High Opening new ground before it converts. Watch it.
Low Low Serving warm people again. Under 40% bucket.

Do not cut the low and high ad too fast: it may become a top bucket ad once that ground starts buying. Our creative reach report goes deeper on reach at the ad level.

Step 2: How do you set up the scaling campaign?

One Sales campaign using ABO (Ad Set budget optimization, where you set the budget on each Ad Set instead of letting Meta split it), with three Ad Sets that differ only in the ads inside them.

Ad Set New purchase share Job
70 to 100% Most buyers are new Main growth engine. Biggest budget.
40 to 70% Mixed Middle tier. Watch for ads ready to move up.
Under 40% Mostly warm buyers Closes people already in the funnel. Smallest budget.

Give all three Ad Sets identical broad targeting, with past buyers and subscribers excluded. When the targeting is the same, the only difference is the ads, so budget is how you steer. Each ad lives in one Ad Set only. The exact budget split depends on your margins, but the top bucket should get clearly the most.

This campaign is where winners land after testing. For the full path, see promoting a winner to Scaling.

What does this look like on an example?

Say Nightfold, a weighted sleep mask brand, has set up the three bucket campaign. Every number below is illustrative, made up to show the method: $500 a day in 70 to 100%, $200 a day in 40 to 70%, and $80 a day in under 40%.

A few weeks in, the team pulls the Audience segments breakdown again:

Ad (illustrative) Ad Set now Spend Purchases New share Belongs in
Side sleeper review (UGC video) 40 to 70% $3,200 64 76% 70 to 100%
Gift for a light sleeper (static) Under 40% $900 15 55% 40 to 70%
Blackout test in daylight (static) 70 to 100% $2,800 47 91% Stays
Pressure on the eyes explained (video) 70 to 100% $1,100 16 85% Stays
Partner snoring skit (UGC video) 70 to 100% $2,100 33 82% Stays
Restock reminder (static) Under 40% $600 14 29% Stays

All figures illustrative. Meta reported purchases, default attribution.

Two ads are one bucket too low. The side sleeper review sells mostly to new people but is stuck in the middle, earning less budget than it deserves. Both it and the gift static move up. Ads drift in both directions, which is why you resort.

Step 3: How do you prove it is working?

Run a net new reach report on the scaling campaign. The Audience segments breakdown is the sorter. Net new reach is the scoreboard.

Net new reach is the number of people your ads reached in a period who had not been reached before. If this method works, net new reach climbs as budget moves into the top bucket. Check 1, 3, 6 and 12 month windows to see the short and long term trend.

Working: spend goes up and net new reach goes up with it.

Not working: spend goes up and net new reach stays flat. The "new" pool is being served over and over.

The goals we start with: cost per 1,000 net new reach under $100 to $200, and a net new reach share of 10 to 20% or higher. Treat these as starting points, not rules. They can change a lot by brand and product: a low priced impulse buy, a high priced considered purchase and a brand that is already well known will each land in different places, so set your own bar from your account's history once you have a few months of data.

Our guides to net new reach and cost per 1,000 net new reach cover how to pull both and set a goal from your own history.

What rules and watch outs matter most?

  1. One ad, one Ad Set. The same ad in two buckets bids against itself and muddies your read.
  2. Add ads in batches, not a flood. Adding ads can push an Ad Set back into the Learning phase (the unstable stretch while Meta gathers data). Add a large number at once and several Ad Sets can get stuck in Learning together.
  3. Verify new share outside Meta. Meta's split depends on your customer lists being current. Check it against Shopify or third party new customer counts.
  4. Judge the top bucket on cost per new customer. A higher cost per purchase in the 70 to 100% Ad Set can be fine.
  5. Keep targeting identical. If one Ad Set drifts, budget stops being a clean steering wheel.

What do you need to decide before you start?

Three choices depend on your account size and buying cycle:

  1. Minimum purchases before an ad gets bucketed. Too low and new share is noise. Too high and good ads wait too long.
  2. How often you resort. Often enough to catch drifting ads, not so often that you keep pushing Ad Sets into Learning.
  3. The budget split. How much more the top bucket gets, based on margins and what you will pay for a new customer.

Key takeaways

  • ROAS does not tell you who bought. Sort winners by new purchase share instead.
  • Use the Audience segments breakdown, after setting your engaged audience and existing customer lists.
  • Build one ABO scaling campaign with three Ad Sets (70 to 100%, 40 to 70%, under 40%) and identical broad targeting, so budget does the steering.
  • Prove it with net new reach on the scaling campaign: more spend should mean more net new reach.
  • Judge the top bucket on cost per new customer, not cost per purchase.

FAQ

Where do I find new purchase share in Meta? At the ad level in Ads Manager, add the Audience segments breakdown. It splits purchases into new and existing, but only after you set your engaged audience and existing customer lists. Divide new audience purchases by total purchases.

Why use ABO instead of letting Meta move budget? ABO lets you decide how much goes to the new customer finders. With campaign level budgets, Meta may shift money toward the cheapest conversions, which are often warm buyers.

What if an ad has great ROAS but low new share? Keep it in the under 40% bucket with a smaller budget. It still closes warm buyers well. It just should not get the money meant for growth.

How do I know if the setup is working? Run net new reach on the scaling campaign over several windows. If it rises as budget moves into the top bucket, it works. If spend rises and net new reach stays flat, you are recycling the same people.

Want to know which of your ads actually find new buyers?

We'll look at where your scaling budget goes and who it reaches, for free. You get a full account audit plus net new reach and cumulative reach reports. Request it on our contact page.

Check what you learned

Tap an answer to see if you got it right.

1. What should you sort winners by instead of ROAS?

ROAS doesn't tell you who bought. New purchase share does.

2. How is the scaling campaign set up?

Same broad targeting in each, so budget does the steering.

3. How do you prove it's working?

If you're really finding new customers, net new reach grows with spend.

Ready for the full test? Take the Scaling exam and get certified. Want this run on your account? Get my free audit.

About the author

Jake Mehani, founder of Creative Analytics

Jake has spent 15 years running Meta and paid social ads for consumer brands. Meta published his RBX Active Advantage+ case study, he taught the DMAnc Meta and TikTok ads masterclass, and he spoke at Motion's 2024 Creative Strategy Summit. His senior team runs media buying, creative and measurement like an in house team.

Jake on LinkedIn  |  Case studies  |  Free Meta audit

Previous
Previous

Meta Static Ad Playbooks by Category: Fashion, Jewelry, Bedding and Food

Next
Next

When Should You Use Amplified Budgets, and How Do You Judge Them?