DTC Audience Targeting in 2026: What Still Works and What's a Waste of Budget

If you built your Meta targeting strategy in 2019 or 2020, it's probably costing you money. The detailed interest stacks that used to find niche audiences, the narrow lookalikes that outperformed everything else, the elaborate exclusion layers built to prevent audience overlap — most of that is now theater. It creates the appearance of precision while actually constraining the algorithm from doing the thing it's now very good at: finding your buyers within large, broad audiences.

The shift didn't happen overnight. It came in waves — iOS14 degraded audience signal, algorithm improvements made broad targeting more viable, Meta's push toward Advantage+ consolidated audience control toward the machine. But the net result in 2026 is clear: audience targeting as a competitive advantage is mostly gone. Creative is now doing the work that targeting used to do.

That doesn't mean audience strategy is irrelevant. It means the strategy has changed fundamentally, and the brands still running the 2019 playbook are leaving real money on the table.

The Death of Interest Targeting (And Why It Happened)

Interest targeting was always a proxy. When you targeted "people interested in fitness" for a supplement brand, you weren't actually reaching your buyers — you were reaching everyone who liked a fitness page, which includes athletes, casual gym-goers, physical therapists, fitness influencers, and people who followed a fitness page three years ago and never unfollowed it. The overlap with "people who will buy your supplement" was real but imprecise.

In the pre-algorithmic-optimization era, that proxy was useful because the algorithm needed your help. You knew more about who your buyer was than Meta did, and detailed targeting helped Meta find them. That advantage eroded as Meta's optimization capability grew. By 2022–2023, for most mainstream DTC products, Meta's algorithm could outperform human-specified interest targeting when given sufficient audience breadth and enough conversion events to learn from.

The practical evidence is compelling. When DTC brands at scale run A/B tests between their best interest-stacked campaigns and broad targeting with identical creative, broad wins more often than not — and the gap widens with spend. At $50K/month, interest targeting can still competitive. At $200K+/month, broad is usually dominant because the algorithm needs volume to optimize and interest stacking starves it of that volume.

There are exceptions. Products with very small, clearly definable audiences (B2B-adjacent DTC, professional tools, highly specialized verticals) can still benefit from interest targeting as a qualifying layer. The test is simple: does the interest you're targeting genuinely identify purchase intent, or does it identify a demographic that happens to include some purchasers? If it's the latter, you're probably better off with broad.

Lookalikes: Still Alive, But Not What They Were

Lookalike audiences haven't died — but they've been weakened by the same forces that degraded retargeting. The quality of a lookalike is a function of the quality of the seed audience. If your customer list is built from pixel events and iOS opt-outs mean a significant portion of your buyers are untracked, your seed data is incomplete. Meta is modeling your best customers from an imperfect sample.

The result: 1% lookalikes from customer purchase lists still show value for some brands, but the performance gap over broad targeting has narrowed significantly. The brand we ran two years ago that showed 40% better CAC on 1% LAL vs. broad? Now it's closer to 8–12%. Still meaningful, but not worth the structural complexity of building your entire targeting strategy around it.

Where lookalikes still earn their keep:

• As seed-enhancement for broad: Using your customer list to inform Meta's understanding of who to find, even without a formal LAL campaign, through value-based optimization signals
• For specific high-value cohorts: A lookalike of your top 10% LTV customers (modeled from first-party data, not just pixel purchases) can still surface meaningful efficiency gains
• For new account scale-up: When an account has limited conversion history, a LAL gives the algorithm a starting point while it builds its own optimization data
• As a testing baseline: Comparing LAL vs. broad performance on a regular cadence is still useful signal for understanding how your algorithm optimization is developing

First-Party Data: The Actual Moat

While the interest targeting debate has dominated conversation, the most durable targeting advantage in 2026 is first-party data — your customer list, your email subscribers, your high-intent site visitors. This data isn't degraded by iOS because it doesn't depend on Meta's pixel. You own it. You upload it. Meta matches it against its own user base at the hashed email level.

This is where the targeting strategy should actually live for mature DTC brands. Not in interest stacks or even lookalikes, but in systematically building and segmenting your own audience data and activating it through Meta's custom audience tools.

The quality of these custom audiences compounds over time. A brand with three years of customer data, properly segmented and maintained, has a targeting infrastructure that a new entrant can't replicate quickly. This is the real moat — not creative, not budget, but the depth and quality of your first-party audience stack.

Campaign Structure When Audience Isn't the Primary Variable

If the audience isn't doing the work it used to do, how do you structure campaigns? The answer is to structure around the work that IS being done — creative differentiation by awareness stage.

The old model: multiple campaigns targeting different audience segments with similar creative, optimizing for who you reach. The new model: fewer campaigns with broader audiences, differentiated by creative that targets different points in the buyer's awareness journey.

This looks like three core campaign buckets:

Cold prospecting (broad): New-to-brand audiences. Creative is hook-first, problem/solution framing, brand introduction. This audience doesn't know you exist — your creative has to earn their attention in the first three seconds.

Warm consideration (engagement + email LAL): Audiences who've had some brand exposure. Creative can go deeper — product education, proof points, differentiation from alternatives. These people know what you do; help them understand why you're the right choice.

Hot conversion (site visitors + cart abandoners): Audiences with active purchase intent. Creative addresses final objections — risk removal, social proof at scale, specific incentives. Close the deal, don't re-introduce the brand.

The structure discipline here is important. Running conversion-focused creative to a cold audience is one of the most common and expensive mistakes we see. The message doesn't match where the audience is in their journey. You're asking someone to buy from a brand they've never heard of. No amount of targeting optimization compensates for that mismatch.

The Creative-Targeting Relationship: What This Actually Means

The fundamental insight of the targeting landscape in 2026 is that your creative is now your targeting. The hook, the framing, the emotional entry point of your ad — these determine who self-selects and engages, which determines who Meta's algorithm shows your ad to next. The audience finds you through your creative, not the other way around.

This is a genuinely different way of thinking about paid media. Pre-2021, a media buyer could be the competitive advantage — the person who could build a better audience than anyone else. Today, the creative team is the competitive advantage. A brand with better creative than its competitors will outperform at audience acquisition regardless of how similar their targeting structures are.

The practical implication: media buying decisions and creative decisions have to be made together. A media buyer who doesn't understand awareness stage differentiation in creative will build a structurally sound campaign that still underperforms. A creative team that doesn't understand that different audiences need different entry points will produce great ads that get deployed in the wrong places.

What to Stop Doing and What to Start

Based on where things actually stand in 2026:

Stop:

• Building interest-stacked targeting layers for mainstream DTC products at scale
• Creating separate campaigns for every audience segment you can think of
• Running identical creative across awareness and retargeting campaigns
• Treating lookalikes as the primary prospecting audience for accounts with $100K+/month budget
• Fighting the algorithm's audience consolidation by over-constraining with exclusions

Start:

• Building and maintaining your first-party data stack systematically
• Differentiating creative explicitly by awareness stage (not by audience segment)
• Testing broad targeting against your current interest-stacked setup
• Using customer list quality as a targeting moat — segment by LTV, purchase frequency, product line
• Structuring campaigns around creative hypothesis testing, not audience hypothesis testing

The brands that are scaling efficiently in 2026 have made this transition. They've accepted that audience targeting is a commodity — that Meta's algorithm is better at finding buyers than they are when given sufficient breadth. And they've redirected the competitive energy into the place where it actually creates advantage: building a creative system that can produce differentiated, awareness-stage-specific content at volume and velocity.

---