Facebook Ads Targeting: Broad vs Detailed

Facebook ads targeting is how you tell Meta who may see your ads on Facebook, Instagram and WhatsApp. Broad targeting sets only basic limits, such as location, age and language, and lets Meta's AI find buyers. Detailed targeting narrows the audience by interests, demographics and behaviours. Custom and lookalike audiences start from your own customer data.

• Broad: Location, age and gender at most. The creative and the delivery system do the rest.
• Detailed: Adds interests such as "coffee" or behaviours such as "engaged shoppers" to narrow who is eligible.
• Custom audience: People who already know you: site visitors, app users, video viewers or a consented customer list.
• Lookalike audience: New people who resemble a custom audience.
• Advantage+ audience: Meta's setting that treats your choices as suggestions, not hard limits.

This lesson follows one example: a D2C filter coffee brand from Chennai that sells coffee powder and a monthly subscription on its website. It ships across India and wants new subscribers outside Tamil Nadu.

Quick Answer

For most accounts that already record purchases through the Pixel and Conversions API, start with broad targeting and strong creative, because the AI ad delivery system usually finds buyers better than a hand-picked interest list. Use detailed targeting for a new account with little data, a niche product, or a test of a new region. Use custom audiences for retargeting and lookalikes to widen reach from your best customers. Whatever you choose, the ad itself does much of the targeting: a clear message about South Indian filter coffee attracts the right people.

Facebook Ads Targeting Comparison Table

When to Use Broad Targeting

• The Pixel records purchases: The system has enough events to learn who buys.
• Proven creative: An ad that clearly says "South Indian filter coffee, ground fresh, delivered across India" filters viewers by itself.
• Scaling: When a campaign works and needs more reach, broad audiences give it room to grow.
• Advantage+ campaigns: Meta Advantage+ campaigns are designed around broad audiences.

When to Use Detailed Targeting

• A new ad account: With no purchase history, interests give the system a starting point.
• A niche offer: A brass coffee filter set for collectors reaches a narrow group, and an interest in South Indian cooking helps.
• Testing a region or group: To learn whether Pune or Kolkata responds, narrow to that city and compare.
• Language groups: A Tamil-language ad can be limited to people who use Facebook in Tamil.

When to Use Custom and Lookalike Audiences

• Retargeting: Show a subscription offer to people who added coffee to the cart but did not buy. This needs events from the Meta Pixel and Conversions API.
• Seed for lookalikes: Upload a list of subscribers who agreed to marketing use, and let Meta find similar people.
• Exclusions: Keep current subscribers out of an offer meant for new customers.
• Consent first: Only use lists collected with permission, never bought or scraped ones, as explained in first-party data strategy.

Example: Targeting for a Chennai Coffee Brand

• Month 1, detailed: The new ad account starts with interests in coffee and South Indian food, in Bengaluru, Hyderabad, Pune and Mumbai. Sales come in, and the Pixel starts recording purchases.
• Month 2, custom: A retargeting ad set shows a first-order offer to cart abandoners from the last 30 days.
• Month 3, broad: A new ad set targets all of India outside Tamil Nadu, ages 24 to 55, with no interests. The creative does the work: a video of decoction dripping through a steel filter.
• Compare fairly: The team runs the broad and detailed ad sets with the same ads and similar budgets, and compares cost per new subscriber, as shown in creative testing on Meta.
• Decide with real data: It checks new subscribers in its own store records, not only Meta's reports, before moving budget.

How AI Changes Facebook Ads Targeting

What AI Automates Now

Meta's AI now finds people inside large audiences, expands beyond detailed targeting when expansion is on, and builds lookalikes. In many campaigns, the audience you enter is only a suggestion. The idea behind lookalikes, finding people whose data looks similar, is close to the embeddings used in AI search.

What Still Needs a Human

People set the hard limits that matter to the business, such as delivery regions and age, decide which customer data may be used, and write creative that speaks to one clear buyer.

Risk to Watch

Uploading customer lists without consent breaks trust and the law. Also watch for broad campaigns spending on people who already buy from you; check new versus returning customer numbers in your own records.

Do It with AI

Use this prompt to plan an audience test. It works in ChatGPT, Claude or Gemini.

1. Write down your tracking status and what customer data you have consent for.
2. Run the prompt and pick one audience test.
3. Build both ad sets with the same ads and run them for at least a week.
4. Compare new customers in your own records, not only in Ads Manager.

Check Before You Use It

• Facts: Confirm current targeting options in Ads Manager; Meta removes and renames them.
• Brand fit: Keep the ad message true to the product, since the creative now does much of the targeting.
• Compliance: Use only consented lists, follow the DPDP Act, and never target by sensitive traits such as health or religion.