You open Ads Manager, head to the audience section, and start typing interests one by one: “food & dining,” “modest fashion,” “stay-at-home moms,” “online shopping,” “quitting your job to start a business.” The list keeps growing, and it feels like the more interests you add, the more “on target” your ad becomes. The reality is often the opposite: your Rp50,000 daily budget disappears, your reach stays narrow, and you can count the number of closed sales on one hand.

Meanwhile, another seller you know sets nothing but location and age, leaves the interests blank — and their sales just keep rolling in. So in an era where Meta is getting smarter with AI, do you still need to hand-pick interests one by one? Or are you better off handing the job to the system? Let’s break it down completely, with a setup recommendation you can put to work today.

First, Get the Definitions Straight: Broad vs Interest Targeting

To avoid confusion, let’s agree on what each term actually means.

  • Interest targeting. You tell Meta specifically who you want to reach based on interests, behaviors, or demographics — for example, people who like “skincare,” behaviors like “frequent online shoppers,” or women aged 25–34.
  • Broad targeting. You barely restrict interests at all. You set location, an age range, and gender, then let Meta’s algorithm decide who is most likely to buy.

Here’s an analogy. Interest targeting is like personally pointing out potential buyers in a crowd. Broad targeting is like handing a photo of your product to a guide who knows every face in the city and saying, “Find me the ones most ready to buy.” How smart that guide has become is exactly what has changed dramatically over the past few years.

Why Interest Targeting Used to Be King

Interest targeting used to be the main weapon, and for good reason. Meta’s data wasn’t as rich as it is now, the pixel hadn’t collected many conversion signals, and the algorithm wasn’t nearly as intelligent. So the advertiser’s job was to “spoon-feed” the system clues through interest selection so it wouldn’t aim at the wrong people.

But this approach carries traps that still catch beginners today:

  • Over-narrowing. Stacking too many interests until the audience becomes so small that Meta has no room left to learn and optimize.
  • Audience overlap. One person can fall into several interests at once, so your ads end up competing against your own ads — and your CPM climbs.
  • Guesses that may be wrong. Someone who likes a “cooking” page won’t necessarily buy your cookware. Interests often signal curiosity, not purchase intent.

What Changed: Meta Is Far Smarter Now

This is the heart of the shift. Meta now relies on machine learning that processes millions of signals: who clicks, who adds to cart, who checks out, at what time of day, and from what kind of content. Armed with conversion data from the pixel and the Conversions API, the system often finds buyers more accurately than our manual interest guesses ever could.

Meta is even steering advertisers toward Advantage+ Audience, where the interests you enter are no longer a rigid fence but merely an “audience suggestion.” If the system finds potential buyers outside the interests you chose, it’s still free to reach them. The signal is clear: the era of hand-picking 20 detailed interests is over.

If your ads are burning budget even though your interests look “perfect,” the root cause usually isn’t the targeting — it’s your creative and your pixel data. We dig into that in our other guides on why ROAS runs low and how to fix it over on the blog.

Broad vs Interest: When Does Each One Win?

Honestly, neither wins outright. What actually happens is that each one wins under certain conditions. Here’s the guide.

Broad targeting has the edge when:

  • Your pixel is mature and has collected enough conversion data — say, dozens of purchases per week.
  • Your product is mass-market: fashion, food, skincare, gadgets, general services. The market is wide, so let Meta do the searching.
  • Your budget is large enough to give the system room to learn — at least around Rp50,000–Rp100,000 per day per campaign.
  • You’re trying to scale and need the largest possible audience without hitting fatigue too quickly.

Interest targeting is still useful when:

  • Your pixel is still “cold” — newly installed with little data. A few broad interests can give it early direction.
  • Your product is very niche: specialist fishing gear, gear for a particular hobby, tightly segmented B2B products.
  • Your budget is very small, so you want to limit spend to the most relevant segment first.

A Practical Setup for Meta’s AI Era

This is the framework we use when managing client accounts. Follow it in order.

  1. Check your pixel first. If it has no conversion data at all, don’t jump straight to full broad. Help Meta along with one to three genuinely relevant broad interests.
  2. Mass-market product with pixel data? Start broad. Set the location (Indonesia or your target cities), a realistic age range, then leave interests empty. Don’t forget to exclude past buyers so budget isn’t wasted on people who already purchased.
  3. Don’t stack 20 interests. If you still use interest targeting, three to five broad interests combined into one large audience is plenty — ideally a potential reach above one million people — rather than splitting into tiny slivers.
  4. Turn on Advantage+ Audience and treat your interests as a suggestion, not a hard limit. Give the system freedom to search.
  5. Pour your energy into creative. In the broad era, creative is the real targeting. A strong three-second hook, copy that names your audience’s problem, and a clear offer will filter who stops scrolling far more powerfully than any interest selection.
  6. Allow 3–7 days to learn. Don’t kill a campaign just because day one produced no sales. The system needs to exit the learning phase first.

A Budget Illustration

Say you allocate an ad spend of Rp10,000,000 per month — roughly Rp330,000 per day. That’s usually more than enough for one or two broad campaigns with healthy room to learn. On the other hand, if your budget is only Rp30,000 a day and you split it across five narrow interest ad sets, each one gets just Rp6,000 — far too little for Meta to learn anything. You’re better off pooling the budget.

Common Mistakes That Waste Ad Budget

  • Stacking too many interests until the audience is narrow and expensive.
  • Splitting a small budget across many ad sets, so none of them gets enough to exit the learning phase.
  • Killing campaigns too early, before the system has had 3–7 days to optimize.
  • Blaming targeting for weak creative. When the hook and offer are flat, no interest list can save the campaign.
  • Forgetting to exclude past buyers, quietly spending on people who already converted.

So, Which One Wins?

The honest answer: broad targeting is winning the race for most mass-market SMBs, provided your pixel has data and your creative earns attention. Interest targeting isn’t dead — it’s a useful starter for cold pixels, niche products, and tiny budgets. But it’s no longer the main event. Set your foundation, hand the searching to Meta, and spend your best hours making creative that stops the scroll.

Want an Expert to Run It For You?

If dialing in targeting, creative, and tracking sounds like a full-time job, that’s because it is. Aira Tech runs data-driven Meta Ads for Indonesian SMBs — you keep full control of your ad account and budget, and we handle the strategy, creative, and optimization. Explore our services, see transparent pricing, or configure a package that fits your goals. For more practical playbooks, browse the rest of our blog.