Key Takeaway: The industry's response to signal loss has been better probabilistic targeting. More sophisticated algorithms, more inference, more guessing. The results speak for themselves: it's not working. The alternative isn't better guessing. It's deterministic verification.
Every proposed solution to the privacy-targeting crisis has one thing in common: they're all probabilistic.
Lookalike audiences model who might be similar to your best customers. Contextual targeting infers interest from content consumption. Google's Topics API assigns users to broad interest categories based on browsing behavior.
Each approach is more sophisticated than the last. Better algorithms, more signals, smarter inference. And yet conversion rates keep declining, CAC keeps rising, and marketers keep describing themselves as flying blind.
The problem isn't the sophistication of the guessing. The problem is that we're still guessing.
Probabilistic targeting works at scale. Target enough people who look like your customers, and some of them will convert. The math works out if you're comfortable with low precision and high volume.
But the math has gotten worse.
Lookalike match rates have dropped from around 80% to 40-50%. The seed audiences themselves are smaller and less representative. You're building models on thin data, generating lookalikes of lookalikes, each generation further from the original signal.
Contextual targeting tells you what content someone is consuming, not what products they buy. Someone reading a review of your competitor isn't necessarily a customer. They might be considering both options. They might be your customer researching alternatives. Context is a weak signal for intent.
The Topics API offers roughly 350 interest categories. Compare that to the behavioral granularity cookies enabled. Going from thousands of behavioral segments to 350 topics isn't refinement. It's a step function downgrade in targeting capability.
Each approach shares the same limitation: they infer rather than confirm. And inference has a ceiling.
Probabilistic targeting gives you confidence intervals, not certainties. There's a 60% chance this person is interested in your category. There's a 45% chance they're similar to your best customers. There's a 30% chance they're actively in-market.
These probabilities might be accurate. The targeting might be well-calibrated. But when you're spending money on each impression, "probably interested" is expensive.
The gap between click and conversion reflects this uncertainty. Platforms optimize toward engagement signals because that's what they can measure. You pay for reach. Whether that reach translates to qualification is a different question.
The more you spend, the more you're betting on probabilities. Some bets hit. Most don't. CAC is the weighted average of a lot of misses.
Deterministic targeting doesn't guess. It confirms.
A user isn't probably a competitor customer. They verified they have an active competitor account. They're not probably high-value. They proved their premium status tier. They're not probably in-market. They just completed a qualification flow.
The difference is fundamental. Probabilistic targeting says "this person is likely qualified based on signals we can observe." Deterministic verification says "this person is definitely qualified based on facts they've proven."
Conversion rates on deterministic audiences are structurally different from probabilistic ones. You're not reaching people who might be interested. You're reaching people who've confirmed they are.
Here's how deterministic qualification works in practice.
Instead of targeting people you think might be competitor customers, you create an offer for verified competitor customers. The targeting is open. The offer requires verification.
User sees the offer. Claims they're a competitor customer. Verification flow: they authenticate with their competitor account, status is confirmed, offer unlocks.
You've just converted a prospect who is definitionally qualified. They're not probably a competitor customer. They proved it. The acquisition is deterministic by construction.
This inverts the traditional funnel. Instead of targeting narrowly and hoping for qualification, you target broadly and let qualification filter. Everyone who converts is verified. There are no false positives.
The privacy changes that killed cookie-based targeting aren't reversing. The industry is moving toward more privacy, not less. Google's Privacy Sandbox, Apple's continued restrictions, expanding regulations worldwide.
Probabilistic approaches will keep getting more sophisticated. The algorithms will improve. The inference will get smarter.
But the ceiling is still there. Guessing has limits. And as targeting signals continue to degrade, even smart guessing produces worse results.
Deterministic verification sidesteps this entirely. You're not trying to infer from degraded signals. You're confirming directly with the user.
The irony is that verification produces more accurate targeting than cookies ever did. Cookies inferred behavior from browsing patterns. Verification confirms facts from actual accounts. It's not a workaround for signal loss. It's an upgrade.
The industry is bifurcating.
One path: keep investing in probabilistic approaches, accept declining precision as the new normal, optimize for scale over accuracy.
The other path: shift to deterministic verification, let users qualify themselves, and build acquisition funnels where every conversion is confirmed.
The brands on the second path are seeing structurally better unit economics. Not because they've solved probabilistic targeting. Because they've stopped relying on it.
Better guessing isn't the answer. Knowing is.
Want to build verified engagement into your campaigns?
Whether you're launching reward programs, partnership campaigns, or dynamic incentive strategies, we can help you eliminate fraud and target with precision.
Schedule a demo or learn more at burnt.com.