Broad Targeting on Meta: Let Your Creative Do the Targeting
Interests and lookalikes matter less than they used to. Learn how broad targeting works on Meta, when it beats detailed targeting, and how creative becomes your real targeting lever.
Five years ago, a big part of a Meta media buyer's job was audience research: stacking interests, testing 1% versus 3% lookalikes, excluding the right segments. Today, many of the strongest accounts run with almost no targeting at all. Country, maybe an age floor, and that is it.
This is broad targeting. It works because Meta's delivery system has become better at finding buyers from conversion signals than a human is at guessing them from interest categories. But broad targeting only works if you give the system something to work with, and that something is your creative.
What broad targeting actually means
Broad targeting means you set only the minimum constraints, usually location, and sometimes age and gender if your product truly requires it, then let Meta decide who sees the ad. No interests, no behaviors, no lookalikes.
Advantage+ audience, Meta's default for many objectives now, works in a similar way. You can add "audience suggestions" such as interests or a customer list, but Meta treats them as hints and will go beyond them when it finds better prospects. Advantage+ sales campaigns go further and remove most audience controls entirely.
In practice, broad targeting and Advantage+ audience are close cousins. Both hand targeting decisions to the algorithm.
Why broad often beats interests
The algorithm sees more than you do
Meta knows who has clicked, bought, watched and scrolled past ads like yours. Interest categories are a coarse summary of that data, often based on page likes and loose signals. When you narrow to "interested in yoga," you exclude many likely buyers who never signaled that interest in a way Meta categorized.
Signal loss made detailed targeting weaker
Tracking limits reduced the data behind interest and lookalike audiences. Broad delivery leans on conversion events reported through the Pixel and Conversions API, which you control. If your conversion data is clean, broad has more to learn from than a shrinking interest graph. Our Meta Pixel and Conversions API guide covers getting that data right.
Bigger audiences mean cheaper auctions
Narrow audiences force Meta to compete for the same small group of people, which often pushes CPMs up. A broad audience gives delivery room to find cheaper impressions among people who still convert.
How creative does the targeting
Here is the idea that makes broad work: the ad itself filters who engages. Meta watches who stops, clicks and buys, then finds more people like them. The creative decides who those first responders are.
An ad that opens with "If you're a runner with knee pain after 5K…" will be ignored by most people and stopped on by runners with knee pain. Meta learns quickly who responds and delivers to more of them. You did not select "runners" as an interest. The hook did.
This changes how you plan creative. Instead of one polished ad shown to five interest audiences, you want five different ads, each speaking to a different buyer, all shown to one broad audience.
| Old approach | Broad approach |
|---|---|
| 1 to 2 ads, many audiences | Many distinct ads, 1 audience |
| Targeting done in ad set settings | Targeting done in hooks and visuals |
| Test audiences | Test angles and personas |
| Narrow lookalikes and interests | Country plus age floor |
Meta's recent delivery changes (often discussed under the name Andromeda) reward exactly this: genuinely different creative concepts, not small variations of one ad. Ten color swaps of the same image count as roughly one concept to the system.
Building a broad-ready creative set
Think in personas and angles. For a hypothetical $45 ergonomic pillow, you might build:
- Side sleeper with neck stiffness. Hook: "Waking up with a stiff neck? Your pillow height might be the problem."
- Couple with different needs. Hook: "One pillow, adjustable fill. Finally ends the 'too flat, too thick' argument."
- Gift buyer. Hook: "The gift for the person who complains about sleeping badly."
- Back pain sufferer, skeptical. Hook: "Tried three pillows already? Here's what's different about this one." (Keep claims verifiable and avoid promising medical outcomes.)
- Hot sleeper. Hook: "Flip it at 2 a.m. The cool side actually stays cool." (Only if true.)
Each concept gets its own visual style too: a talking-head UGC clip, a static comparison graphic, a product demo, a review screenshot. Variety in format helps delivery find different pockets of buyers.
If you struggle to generate distinct angles, use the ad angle generator or the customer persona generator to list the buyer types first, then write a hook for each.
The best source of angles is what already works in your niche. Ads that have been running for months in the Meta Ad Library are a strong clue about which messages convert, as we explain in why long-running ads are your best creative research.
When broad targeting fails
Broad is not a universal answer. It struggles in a few situations.
- Weak conversion data. If your Pixel fires unreliably or you get only a handful of purchases a week, the algorithm has little to learn from. Consider optimizing for a higher-volume event (add to cart) temporarily, or fix tracking first.
- Tiny budgets. At $20 a day with a $100 product, you may get one purchase every few days. Learning takes a long time. Broad still often beats narrow here, but expect patience.
- Niche or regulated products. Products with legal age limits need hard constraints. Very niche B2B-like offers may waste impressions on irrelevant people early on.
- Generic creative. If your ad could be for anyone ("Shop our new collection"), broad has no filter and delivery spreads thin. This is the most common failure.
A broad targeting setup to start with
For a consumer brand with working Pixel and Conversions API:
- Campaign: Sales objective, Advantage+ sales campaign or manual sales campaign with Advantage campaign budget.
- Audience: Your shipping countries. Age 18+ (or your legal minimum). No interests. If using manual setup, enable Advantage+ audience with no suggestions, or one customer-list suggestion.
- Exclusions: Recent purchasers if this is an acquisition campaign (or use the existing-customer budget cap in Advantage+).
- Creative: Five to ten distinct concepts across at least two formats.
- Budget: Enough for roughly 50 conversions per week per ad set if possible, which is Meta's commonly cited threshold for exiting learning. If that is not realistic, consolidate into fewer ad sets.
Worked example of that last point: if your cost per purchase is around $40, 50 purchases a week is $2,000 a week, or about $285 a day. If your budget is $100 a day, you will not hit that in one ad set, so keep everything in a single ad set rather than splitting it. Check your own numbers with the CAC calculator.
How to read results from a broad campaign
Since you cannot compare audiences, compare creative. Look at each ad's:
- Hook rate (3-second video views divided by impressions) for video.
- Click-through rate on the link.
- Cost per purchase and purchase volume.
- Spend share. Meta will concentrate spend on what it thinks works. Ads that get almost no spend after a week are losing the internal auction.
Check the age and gender breakdown too. It tells you who your creative is actually reaching, which can inform new concepts. If a pillow ad meant for young couples is mostly converting women over 55, that is a new persona worth building ads for.
For testing structure, see how to test creative on Meta without burning budget.
Broad targeting and ChatGPT ads
The broad-targeting lesson carries over to newer channels. ChatGPT ads do not use interest lists either. They use context hints: plain-language descriptions of the situation someone is in, such as "side sleeper asking how to choose a pillow for neck stiffness." The ad shows when the conversation matches.
That is the same principle stated differently. On Meta, your hook tells the algorithm who to find. On ChatGPT, your context hint and headline describe the moment you belong in. Both reward specific buyer situations over generic brand messages. Our guide to ChatGPT ad context hints goes deeper on that.
Where SecondWin helps
SecondWin takes the messages that have survived longest in Meta's Ad Library for your niche, the hooks the algorithm has already proven find buyers, and turns them into original ChatGPT ads aimed at the questions those buyers ask. It writes the headline, description, image concept and context hints, checks every ad against OpenAI's policies, and runs campaigns in your own OpenAI ad account.
See which buyer messages and questions it finds for your store with a free SecondWin site analysis, then compare the monthly plans on the SecondWin pricing page.
FAQ
Is broad targeting better than interest targeting on Meta?
For many consumer brands with working conversion tracking, yes. Broad lets Meta use conversion data to find buyers rather than relying on coarse interest categories. It tends to lower CPMs and scale more easily. Interest targeting can still help with very small budgets, brand-new pixels or highly niche products, so test both if you are unsure rather than assuming either wins.
Does broad targeting work for a new ad account?
It can, but expect slower learning. A new pixel has little history, so Meta has less to go on. Make sure purchase events fire correctly through both Pixel and Conversions API, consolidate into one ad set, and use creative with a clear, specific hook so early responders are the right people. If purchases are rare, optimize for add to cart temporarily.
How many ads should I put in a broad ad set?
A common range is three to ten distinct concepts. More than that splits spend too thin on smaller budgets, and fewer gives the algorithm little variety to work with. The key is that each ad is genuinely different in angle, persona or format, not small edits of the same image or headline.
Should I still use lookalike audiences?
Lookalikes are less important than they were, and under Advantage+ audience they act more as suggestions than limits. Some advertisers still see value from a value-based lookalike on smaller budgets. Test a lookalike ad set against a broad one with the same creative and budget, and keep whichever produces lower cost per purchase over two to three weeks.
What age and gender settings should I use with broad targeting?
Usually only your legal minimum age, with gender left open. Let the data show who responds. If a breakdown shows strong skew, build creative for that group rather than restricting the audience. Hard restrictions make sense only when the product is truly limited to one group or legally restricted by age.