Ready to embrace a new era of task delegation?

HubSpot’s highly anticipated AI Task Delegation Playbook is your key to supercharging your productivity and saving precious time.

Learn how to integrate AI into your own processes, allowing you to optimize your time and resources, while maximizing your output with ease.

Get the free guide here.

Ah, the sweet, musky scent of ad-tech snake oil. You know it well—the kind that fills boardrooms with the promise of “next-gen” solutions that are really just an elaborate shell game played by guys in Patagonia vests who wouldn’t know real tech if it smacked them across the face with a line of Python. And yet, here we are again, peeling back another layer of this industry’s grand illusion.

So, a well-known ad exec—who, let’s just say, has enough industry scars to make a CTV fraudster sweat—sat down with me to drop a truth bomb that should have the entire ad-tech ecosystem squirming in their Aeron chairs. This guy isn’t just some LinkedIn loudmouth; he’s been in the trenches, working for some of the largest networks and is currently embedded deep in the CTV and streaming world. He knows where the bodies are buried because, well, he’s been in the rooms where they were dumped.

👉 Every so-called "contextual" or "brand safety" solution in CTV and online video is an absolute farce—a smokescreen wrapped in buzzwords, sprinkled with AI jargon, and sold to advertisers like it’s some kind of digital panacea.

You’ve heard the pitch before:

“Our cutting-edge, AI-powered, machine-learning-driven, next-gen contextual targeting solution ensures brand safety by analyzing video content in real time!”

Sounds fantastic, right? Almost too good to be true. And that’s because it is.

Let’s break this down, because the reality is far grimmer than the PowerPoint slides at some overpriced industry conference would have you believe.

🔹 These companies are NOT actually analyzing video content. There is no sophisticated AI sitting there like a diligent media studies major, analyzing every scene, breaking down the dialogue, identifying themes, and making intelligent decisions about whether an ad for baby formula should really be running next to a gruesome true crime doc.

That’s a fantasy.

What’s really happening? A glorified game of metadata bingo.

🔹 Instead of analyzing the content, these companies scrape publisher-declared metadata—information that publishers self-report about their content—and app bundle IDs passed into the bid stream by sellers and SSPs. Then, with a few clicks in the backend, they slap that junk data onto some pre-labeled "contextual segment" they cooked up.

So, let’s say a publisher marks their video content as “sports.” That could mean anything from a World Cup final to a 2005 YouTube clip of some guy getting hit in the groin with a football. But because that label says “sports,” these brand safety and contextual targeting vendors just roll with it, selling that classification as if it’s some deeply informed, algorithmically verified data point.

Spoiler alert: it’s not.

And let’s talk about that "pre-labeled contextual segment" nonsense.

These aren’t carefully curated, machine-learned categorizations. They’re often just blunt, oversimplified buckets that someone slapped together in a Google Sheet months ago.

🎭 “Comedy”
📰 “News”
📈 “Business”
🔪 “Violence”
⚠️ “Controversial”

That’s the level of granularity we’re dealing with.

And then advertisers are expected to make serious media-buying decisions based on this nonsense. So if you’re a brand, you might think you’re buying a nice, safe "family-friendly" content package.

But in reality?

📌 Your ad could be popping up in the middle of a chaotic MMA knockout reel
📌 Or a news clip about geopolitical chaos
📌 Or some random, low-quality, clickbait-filled dumpster fire

All because some spreadsheet jockey behind the scenes decided they all fall under the same broad category.

Who’s footing the bill for this mess?

👉 The buyers.
👉 The agencies.
👉 The brands.

They are forking over top dollar for what is, at best, an educated guess and, at worst, a flaming pile of nonsense.

You might as well ask a Magic 8-Ball to classify your media placements—it would be just as accurate and cost significantly less.

What’s worse is that this isn’t just inefficiency—it’s outright misrepresentation.

These companies—many of which love to tout their “proprietary” methodologies (which is just a fancy way of saying “don’t ask too many questions”)—are knowingly selling a lie.

They know their technology doesn’t do what they claim.
They know they’re just repackaging whatever scraps of data they can pull from the bid stream.
And yet, here they are, trotting out their junk products at Cannes, patting themselves on the back at CES, and pretending like they’re at the forefront of digital innovation.

Meanwhile, the DSPs aren’t just tolerating this—they’re actively complicit.

They integrate these very same vendors into their stacks, slap their own branding on the data, and push it as a premium solution.

Why?

Because there’s a fat margin on it. It’s just another layer of the ad-tech tax that keeps bloating this industry while delivering little to no actual value.

But hey, as long as everyone gets their cut, who cares if the data is garbage?

The exec’s million-dollar question, then, is this:

How are the solutions you’re using for targeting and measurement actually built?

Are they analyzing the actual video content that audiences are watching?
Or are they just making educated guesses based on recycled metadata?

And before anyone gets too comfortable dismissing this as industry paranoia, let’s be clear—

👉 This is the exact same garbage playbook verification companies ran on social networks.
👉 They “graded” content based on platform-set guidelines instead of actual independent evaluation.
👉 Now, that same sleight of hand is buried deep in CTV and streaming audio, and no one is talking about it.

So, Mike Shields hit me with a good question: Does context actually matter for ad effectiveness? You know, the classic adland assumption that if you’re feeling a certain way—say, on the edge of your seat watching a crime thriller—you’ll be more likely to buy something that aligns with that moment. Like, “Oh wow, this car chase is so intense; I suddenly need a Red Bull.” 🏎️⚡

Sounds a little… forced, right? But let’s not go full conspiracy theorist just yet. Instead, I decided to actually do some research (shocking, I know) and see if there’s any real data proving that the environment where an ad appears has any tangible impact on whether people pay attention, remember it, or—heaven forbid—actually buy something.

The Science Says… It’s Complicated 🧠📊

A Journal of Advertising Research meta-analysis spanning 50 years of studies (because apparently, people have been debating this forever) found a weak correlation between media context and ad memory. In other words, yeah, context might matter… but not as much as marketers love to claim. So if you’re sinking millions into making sure your product is advertised just right within certain types of content, you might want to rethink your priorities. 🚨💰

But hold on—before you torch all your contextual targeting strategies, there’s some nuance here.

🔹 A study from Integral Ad Science (IAS) found that when ads are placed in content that emotionally matches them, recall goes up by 40%. So, if you’re watching something heartwarming and an ad for a feel-good brand pops up, you might actually remember it. This makes sense—humans are emotionally driven creatures, not robots mindlessly clicking through banner ads.

🔹 Meanwhile, a study by RMT Solutions found that context does influence consumer responses, but not in a one-size-fits-all way. If an ad matches the tone and theme of the program—like a car ad in an action movie—it can lead to better recall. But if it feels forced or completely out of place (think: a funeral insurance ad in the middle of a slapstick comedy 🎭⚰️), you’re likely just annoying your audience.

Contextual Targeting vs. Behavioral Targeting: The Ongoing Battle ⚔️🎯

So where does this leave us? The ad industry has been shifting away from behavioral targeting (RIP cookies) and leaning more into contextual advertising, under the logic that serving ads within relevant content is more ethical and more effective.

But is it really?

GumGum (not a joke, it’s a real company) and Dentsu ran a study showing that contextual targeting outperformed behavioral targeting when it came to engagement and attention. More accuracy, less creepy tracking, and—surprise!—it might actually work better.

The Verdict: Context Matters… Sometimes? 🤷‍♂️🔍

Look, context isn’t totally useless—it can boost recall and engagement, especially when done right. But is it the holy grail of ad effectiveness? Probably not.

The real lesson here?
How you place an ad matters more than just where you place it.

So, Mike, good question.
The short answer? Context matters, but not nearly as much as ad execs want you to believe.
The longer answer? Well, it depends—on the ad, the audience, and whether or not people actually care about what’s being shoved in front of them.

And that, my friends, is why the advertising world is always overcomplicating things. 🚀🎭