Should Meta Ads start with broad targeting or interest targeting?
Quick Answer
Start broad with Advantage+ audience if you have 50+ pixel conversion events or a strong creative. Start with interest targeting if your pixel is new, your geographic area is small, or you need the targeting to do the qualification work your creative cannot yet do alone.
The shift to broad targeting and Advantage+ audiences has been one of Meta's most significant changes to its advertising platform in recent years. Meta's machine learning has improved to the point where, for many advertisers, letting the algorithm decide who to target outperforms manual interest stacking — particularly for accounts with significant pixel data. But broad targeting is not universally superior, and applying it without considering your specific situation can mean wasting your entire learning phase budget on entirely the wrong audience.
1. When Broad Targeting Outperforms Interest Targeting
Broad targeting (or Advantage+ audience) works best when Meta's algorithm has enough data to make intelligent decisions about who to target. The algorithm learns from your Pixel data, your previous campaign conversions, and the content of your creative — and uses these signals to identify users who are statistically likely to convert. When these signals are rich and reliable, the algorithm can find high-quality audiences that you would never have thought to target manually with interest categories.
- 50+ pixel conversion events: Once your Pixel has recorded at least 50 lead or purchase events, the algorithm has enough signal to optimise effectively with broad targeting.
- Strong, specific creative: If your creative explicitly names your target customer ("Attention Sydney homeowners..."), the content itself does the targeting work — broad audience settings are safe.
- Large geographic area: Cities with 500,000+ population within your service radius give the algorithm enough people to find your ideal customer within a broad audience.
- High-volume conversion services: Services with daily or weekly conversion events (e.g., restaurant bookings, fitness classes) generate enough data for broad targeting to work efficiently.
2. When Interest Targeting Is the Better Starting Point
Interest targeting is still the right starting configuration in specific circumstances — particularly when the algorithm does not yet have enough data to make good decisions on its own, or when your service is so niche that broad targeting would require the algorithm to find a very small signal within a very large noise pool. In these cases, manually guiding the algorithm with relevant interest categories gives it a head start and reduces the amount of budget wasted in the early learning period.
- New Pixel with no data: Without conversion history, the algorithm has no signal to guide targeting — interest categories provide a structured starting pool.
- Small geographic area: Targeting a single suburb or small town means your broad audience might only be 20,000–50,000 people — interest targeting can help concentrate spend within that limited pool.
- Generic creative: If your creative does not pre-qualify viewers through specific messaging, interest targeting must do the qualification work instead.
- Highly niche service: Services targeting very specific professional roles (e.g., commercial kitchen operators) benefit from interest or B2B targeting to avoid broad-audience waste.
3. Testing Broad vs. Interest: The A/B Framework
The most data-driven approach is to test both simultaneously using Meta's A/B test tool, which prevents audience overlap and provides a statistically valid comparison. Set up identical campaigns with the same creative, budget, and offer — varying only the audience type. Run the test for 14–21 days and compare CPL, lead quality, and ultimately cost per qualified appointment or sale. The winner becomes your primary targeting approach for the next 90-day campaign cycle, after which you test again as pixel data accumulates.
- Use Meta's A/B test tool: Prevents audience overlap contamination that occurs when running two manual campaigns to the same geographic area.
- Equal budget split: Allocate the same daily budget to both ad sets to ensure a fair comparison — do not bias the test by underfunding one variant.
- Primary metric: Evaluate on cost per qualified lead (or cost per booked appointment) rather than CPL alone — broad may produce more leads but at lower quality.
- Retest quarterly: As your Pixel data accumulates, broad targeting typically becomes increasingly competitive — retest every 90 days to capture this improvement.
4. The Hybrid Approach: Broad Prospecting + Interest Retargeting
Many advanced Meta advertisers use broad targeting for cold audience prospecting (letting the algorithm find new high-potential customers) while using interest and behaviour targeting for specific retargeting segments (reaching warm audiences defined by what they engaged with on Meta). This hybrid approach maximises the algorithm's strength in the prospecting phase while using manual signals to ensure retargeting ads reach the most relevant warm audiences — combining the best of both approaches in a single cohesive funnel structure.
- Cold prospecting: Advantage+ audience constrained to geographic service area — maximum algorithm flexibility for finding new customers.
- Warm retargeting: Manually defined custom audiences (website visitors, video viewers) with interest refinements if needed to improve relevance.
- Budget allocation: 70–80% prospecting (broad) + 20–30% retargeting (custom audiences) is a healthy starting split for most service businesses.
- Creative alignment: Broad prospecting creative should be specific and qualifying; retargeting creative should be trust-building and objection-handling — two distinct content strategies.
Frequently Asked Questions
Q:Is Advantage+ targeting the same as turning off all targeting?
Advantage+ audience is not "no targeting" — it still constrains your ads to your selected geographic area and uses all available algorithmic signals (Pixel data, creative analysis, historical performance) to find the most likely converters. The difference from manual targeting is that you are not manually specifying demographic or interest constraints — you are trusting the algorithm to find the best audience within your geographic boundary.
Q:Do interest categories still work in 2026 or has Meta made them obsolete?
Interest categories remain effective in 2026 for accounts with limited pixel data, small geographic targets, or niche service categories. They are less effective than they were in 2018–2020 due to iOS privacy changes reducing the accuracy of interest signals. However, they remain a valuable tool as a starting configuration and for audience research — helping you understand which interest clusters produce the best lead quality before transitioning to broad targeting.
Q:How much pixel data do I need before switching from interest to broad targeting?
A practical minimum is 50 optimisation events (lead form submissions, purchases, or contact completions) within the past 90 days at the campaign level. With fewer than 50 events, the algorithm does not have enough signal to outperform well-structured interest targeting. At 100+ events, broad targeting typically begins to match or beat interest targeting. At 500+ events, broad targeting with Advantage+ is almost always the higher-performing configuration.
Technical Terminology
Advantage+ Audience
Meta's AI-powered targeting configuration that removes manual interest and demographic restrictions, allowing the algorithm to find the best-converting users within a geographic boundary using all available data signals.
Pixel Maturity
The accumulated volume of conversion events recorded by the Meta Pixel on your website. Higher pixel maturity means more data for the algorithm to learn from, improving targeting accuracy and campaign efficiency.
Interest Targeting
A manual audience configuration in Meta Ads Manager that restricts ad delivery to users who have expressed interest in specific topics, pages, or activities based on their on-platform behaviour.