What settings should I use for an Advantage+ Shopping campaign?
Short answer
One Advantage+ Shopping campaign per product, budget set at campaign level at $50 a day, highest-volume bidding with no cost cap, country-only targeting, existing-customer budget cap at 0-10%, all Advantage+ placements on, Purchase as the conversion event, and 4 to 6 creatives in the single ad set. Then no edits for 72 hours, because every change restarts the learning phase.
| Setting | Choose | What happens if you do something else |
|---|---|---|
| Campaign type | Advantage+ Shopping | A manual sales campaign needs audience work that no longer beats the algorithm. |
| Budget level | Campaign, $50/day | Ad-set budgets split your data across ad sets that never exit learning. |
| Bid strategy | Highest volume, no cap | A cost cap on a cold campaign starves delivery and hides whether the product sells. |
| Targeting | Country only, 18-65+ | Interests and lookalikes under 1,000 buyers narrow the pool without improving it. |
| Existing-customer cap | 0-10% | Higher, and the campaign flatters itself by re-selling people you already have. |
| Placements | All Advantage+ on | Turning off Audience Network raises CPMs more than it raises quality. |
| Conversion event | Purchase | Optimising for Add to Cart buys you carts, and carts do not pay for anything. |
| Attribution | 7-day click, 1-day view | The default. Changing it makes your history incomparable, not more accurate. |
| Creatives | 4-6, different hooks | Ten creatives on $50/day gives each $5, which is below a read. |
| Edits after launch | None for 72 hours | Any edit restarts learning and discards the data you just paid for. |
Every setting in an Advantage+ Shopping campaign for a single-product test, and what changing it costs you. Checked August 2026.
The structure is deliberately boring
One campaign. One ad set. Broad targeting. Four to six creatives. $50 a day at campaign level.
That is the whole framework, and its purpose is to avoid splitting a small budget across ad sets that each collect too little data to exit the learning phase. Complexity is how people spend $500 and learn nothing.
The settings that actually matter
Budget at campaign level, $50/day. Ad-set budgets fragment delivery. One pot, one decision.
Highest volume, no cost cap. A cap on a campaign with no history throttles delivery, and you end up judging a product on 40% of the impressions you paid for. Find the real cost per purchase first.
Country only, 18-65+. Advantage+ deliberately limits detailed targeting, and that is fine — the algorithm finds the buyer faster than audience research does. Adding a lookalike built on fewer than 1,000 buyers narrows the pool without improving it.
Existing-customer budget cap at 0-10%. This one is quietly important on a test. Left high, the campaign spends on people who already bought from you, the ROAS looks excellent, and you have learned nothing about whether the product acquires new customers.
Purchase as the conversion event. Optimising for Add to Cart produces carts. Carts do not pay for shipping.
Creatives: four to six, genuinely different
Not four edits of one video. Four different arguments — problem, result, unboxing, comparison, objection — each with its own first three seconds.
At $50 a day across six creatives, each gets about $8.33 a day, or $41.65 over five days. That is above one $35.20 break-even CPA per creative, which is the threshold at which “no sales” means something. Ten creatives puts each at $25 over five days, which is below it. The full arithmetic is in how many creatives to test a product.
Then leave it alone
Every edit to budget, audience or creative restarts the learning phase and throws away the data you just bought. New campaigns look terrible for 48-72 hours, which is exactly when people panic and change something.
The reading schedule:
- Day 3: outbound CTR and cost per add-to-cart. Kill individual creatives below 0.8% CTR. Do not touch the campaign.
- Day 5: cost per purchase against your $35.20 break-even CPA. Now you decide.
- Days 6-10: let winners spend, add two new hooks.
The thresholds are in what CTR is good on Facebook ads, and the full 14-day budget cycle is on the full ad blueprint.
Set up the columns before you launch
Half the mistakes on this page are reading errors, not setup errors. Configure the ads manager view once:
- Outbound CTR and cost per outbound click, not the all-clicks versions
- Cost per purchase, which you compare against your $35.20 break-even CPA
- Cost per add-to-cart, for the day-3 read
- Breakdown set to ad level, because campaign averages hide both your winner and your three failures
Save it as a preset. At 2am on day 5 you want one screen with four numbers on it, not a dashboard you have to interpret.
What people get wrong
Launching five campaigns for five products. $50 a day split five ways is $10 a campaign, which tests nothing. One product at a time.
Turning off Audience Network because the placement has a bad reputation. It raises your CPMs more than it raises your quality.
Duplicating the campaign on day 2 because it looks bad. It always looks bad on day 2.
On Google Shopping there is no campaign structure to get wrong, because the feed is the ad.
Scaling by editing the budget. When it does work, duplicate at 2x and leave the original running — the sequence is in when to scale a winning product.
Related questions
Should I use a cost cap or a bid cap?
Do I need a product catalogue connected?
One campaign per product, or all products in one?
This is one question out of a much longer guide. The full breakdown lives onthe full ad blueprint, and you can run your own numbers in thebreak-even calculator.