What the Signal Era Means for You

Hi there,

The word "signal" is being used constantly within performance marketing and advertising.

"We need better signal." "Our signal is thin." "Signal is the new targeting."

Signal has also been the hot topic in all my recent meetings too, so I want to take a step back.

What is signal? How did we end up here? And why does the quality of signal decide how efficiently your ad dollars get spent?

The collapse of the pixel

From 2007 to 2020, advertisers ran their own targeting. You dropped a pixel on your site, it followed people around the web through cookies and mobile ad IDs, and you built your own audiences by hand. You pulled the lever, and you could say with real confidence which ad drove which sale.

Then iOS 14.5 shipped in April 2021 with App Tracking Transparency, and a permission prompt landed on every iPhone. Roughly 88% of people said no [1]. Layer the slow death of the third-party cookie on top of that, and the entire measurement foundation cracked at once. Conversions from opted-out users stopped getting reported. Attribution windows shrank. Retargeting audiences shrank right along with them. Meta told investors the fallout alone would cost the company roughly $10 billion in 2022 revenue [2].

Here's the part I think gets flattened when people tell this story fast, like it was one clean break. It wasn't. The algorithm didn't step in and take over the next day. From 2021 to 2023, most brands spent two full years trying to patch the old playbook back together: new attribution windows, workaround pixels, first-party data hacks, all of it built on the hope that granular control would come back. It didn't. It just kept eroding, a little more every quarter.

It wasn't until 2024 that the shift actually finished. Meta, Google, and TikTok all converged on the same kind of tool: one workflow that automates targeting, creative, and bidding together, whether they call it Advantage+, Performance Max, or Smart+ [3]. Most advertisers aren't even thrilled about the trade. 60% say they don't have real visibility into how the algorithm is making its decisions [3].

The algorithm runs targeting now. Not the marketer.

Good signal needs four specific things

Here's where I think most brands get stuck. They know they need "better data." Nobody's told them what that actually means in practice.

Signal is the conversion data you send back to a platform after someone buys, telling it who bought, what they bought, and that it was real. In the pixel era, the pixel handled this automatically. Now that the algorithm is doing the targeting, that data is the one lever you still control.

1. Good signal is complete. It reflects your actual sales, not just the ones that happened to run through your website.
2. It's deterministic, meaning it's tied to a real identifier your customer gave you, not a modeled guess about who probably bought.
3. It's server side, sent directly platform to platform through something like Meta's Conversions API, so a browser privacy setting can't quietly kill it.
4. And it's well matched. Meta actually scores this and calls it Event Match Quality, a 0 to 10 rating based on how cleanly your data ties back to a real account [4].

None of those four things are about creative. None of them are about copy. They're about the plumbing behind the ad, and the plumbing is the job now.

Most brands are training their algorithm on one slice of their business

If most of a brand's sales happen at retail or on a marketplace, and only a small slice happens on the brand's own site, the algorithm has never seen most of that brand's customers.

In nearly every operator conversation I've had this quarter, it's the same story. Website checkout is treated like the whole business, when for a lot of these brands it's the minority of it. The algorithm optimizes against whatever slice it can see, decides that's who the customer is, and goes looking for more people who look exactly like that slice.

People assume their targeting is broken. It's not broken. It's blind. Nobody told the algorithm the other side of the business exists.

This is the thing I think about constantly running Brij. A brand that only ever shows the algorithm its DTC checkout is asking that algorithm to go find more of a customer who might not even look like the one actually buying the product off a shelf.

You can rewrite your ad copy fifty times, and it won't fix a signal problem. The lever isn't creative anymore. It's what you're actually feeding the machine.

Talk soon,

Kait

A few extras:

Some of my favorite brands are hiring…

1. BRUNT Workwear - Search Manager, Growth Marketing
2. Homecourt - Director, E-Commerce
3. Therabody - Sr. Product Manager, Growth
4. Bobbie - Creator Partnerships Coordinator

Sources:

[1] Brij, "What Is Signal?: A 2026 Guide."

[2] Forbes, "Apple, Meta And The $10 Billion Impact Of Privacy Changes."

[3] eMarketer, "FAQ on AI Media Buying: Platform Tools, Agency Strategy, and How to Win in 2026."

[4] Meta for Developers, "Conversions API Dataset Quality API (Event Match Quality)."