Meta Wants You to Describe Your Audience in Plain English. That Is Not a Convenience Feature.

Meta rolled out a text box.

That's the most reductive way to describe what happened. Inside Advantage+ targeting, there's now a tab called "Describe Your Audience" where you type a plain-English description of who you want to reach. Up to 2,000 characters. Meta's AI reads it, cross-references its behavioral data, and builds an audience.

Most of the coverage I've seen frames this as a UX improvement. Easier targeting. Fewer dropdowns. And that's technically accurate, but it misses what's actually going on.

You're Training Meta's Targeting AI. For Free.

Every description you write becomes a data point in how Meta's system understands the relationship between human intent and audience characteristics. When you type "mid-career professionals who just got promoted and are thinking about upgrading their wardrobe," you're not just creating an audience. You're teaching Meta's Lattice model what that phrase means in behavioral terms.

Jon Loomer walked through the feature when it rolled out, and the interface is deceptively simple. You type. It builds. You refine. But underneath, Meta is mapping natural language to its own behavioral graph at scale. Thousands of advertisers describing millions of audience concepts in their own words, every day.

That's not a targeting tool. That's a training dataset.

I'm not saying that's a reason to avoid it. I think you should use it. I'm saying you should understand what you're contributing to while you do.

The Dropdown Was Already Dying

The context that makes this feature land differently: Meta's detailed targeting has been in decline for two years. The algorithm increasingly treats your interest selections as suggestions, not instructions. If you've run Advantage+ campaigns, you've probably noticed your reach extending well beyond whatever audiences you thought you defined.

Summit's analysis of the 2026 Meta updates put it bluntly: the system is often smarter than our manual selections. That tracks with what I've seen. The best-performing Advantage+ campaigns in most accounts are the ones where the advertiser got out of the algorithm's way. Describe Your Audience formalizes what was already happening. Instead of fighting the algorithm with dropdown filters it mostly ignores, you're giving it a brief.

Think of it like giving driving directions to someone who already has navigation running. The old way was turn-by-turn instructions the driver mostly ignored. The new way is telling them where you want to end up and letting the system find the route.

The Testing Economics Are What Actually Changed

This is the part that hasn't gotten enough attention.

Under the old system, testing a targeting hypothesis required building an ad set. Sometimes a whole campaign. You'd stack interests, set exclusions, split budget, wait for statistical significance. Testing four audience concepts meant four parallel ad sets competing for the same budget.

With Describe Your Audience, testing a targeting hypothesis costs you 30 seconds of typing. Write three different descriptions, run them against the same creative, compare CPA after a week. The barrier to testing just collapsed.

And the part that gets genuinely interesting: you can now describe audiences that were impossible to build with dropdowns. "Parents of kids starting travel sports who are about to spend a lot on equipment" is not an interest category. "Small business owners who just lost their bookkeeper and need software fast" is not something you can build with demographic filters. But Meta's AI can interpret both and find behavioral matches in its data.

The test I'd run this week: take your best-performing Advantage+ campaign. Duplicate it three times. Leave the original with whatever targeting it has. For the three copies, replace targeting with three different audience descriptions, each describing a different buying moment. Same creative, same budget, 7 days. Compare CPA across all four. I think you'll find at least one description outperforms your current targeting, and I think the gap will be wider than you expect.

What Most Teams Will Get Wrong

The lazy version of this feature is typing "people interested in fitness" into the box. That gives Meta roughly the same signal as selecting the Fitness interest from the dropdown. You've changed the input format without changing the input.

The slightly better version is describing demographics: "Women aged 25-40 in urban areas who earn above $75K." That's more specific, but it's still constraining the algorithm with the kind of filters it already wants to override.

The version that actually matters is describing the moment. Not who the person is, but what they're about to do. "Someone who just moved into a new apartment and needs furniture but doesn't want to spend IKEA time assembling it." "A marketing manager whose agency just got fired and needs to bring media buying in-house within 60 days."

These descriptions work because they give Meta a behavioral signal, not a demographic one. And Meta's behavioral data, from what I've seen, is considerably better than its demographic data at predicting purchase intent.

I wrote about this dynamic when Meta started pushing its three-app subscription model. The pattern is consistent: Meta keeps shifting control away from the advertiser and toward the algorithm, then provides new tools to make the handoff feel collaborative rather than forced.

Your Creative Brief Just Became Your Targeting Brief

The implication that Alex Neiman highlighted is worth sitting with: if the algorithm handles audience finding, competitive advantage shifts entirely to creative and offer. Two advertisers targeting the same description will reach similar people. The one with better hooks wins.

This has been true directionally for a while. Describe Your Audience makes it structurally true. When targeting is a text box anyone can use, there's no targeting edge left. The only edge is what you say to the audience the algorithm finds for you.

For teams that have been investing in creative testing and iteration, this is confirmation. For teams that have been relying on clever audience segmentation to compensate for mediocre creative, this is a problem that was honestly coming regardless. This feature just moved the timeline up.

12 Months Until the Dropdown Disappears

I'd put real money on detailed interest targeting being deprecated for most campaign objectives within a year. Not removed entirely, probably available as a legacy option somewhere. But the default experience will be Describe Your Audience plus Advantage+ automation plus creative as the primary targeting signal.

If that happens, the advertisers who spent the next 6 months learning how to write effective audience descriptions will have a meaningful head start over the ones who kept selecting interests from a dropdown until Meta took the dropdown away.

Right now, your descriptions are teaching the system. Within a few months, it won't need the lesson. The question is whether you'll have learned anything from the process by then.