🪦 Demand-Side Implosion: How the Entire DSP Category Became a Case Study in Overfunded Sameness and Underwhelming Innovation

How “Neutral” DSPs Became Just Another Margin Layer in a Tech Tax Pyramid Scheme

Pesach Lattin (ADOTAT)
April 29, 2025

Our Amazing Sponsor

🔓 The Great DSP Shrinkage

Or: Why the Ad Tech Industry Built a Hundred Versions of the Same Damn Thing and Can’t Seem to Figure Out Why They’re All Going Bankrupt

Let’s stop lying to each other.

What’s happening in the DSP space isn’t innovation, isn’t consolidation, and definitely isn’t the elegant “maturation” that overpaid consultants are trying to spin on stage at industry summits where half the audience is still drunk from the open bar. This is collapse. Slow, painful, avoidable, and utterly predictable collapse.

And yet the people responsible for it are out here acting like it’s a surprise.

For a decade, the demand-side platform became the go-to startup template for anyone orbiting the media-buying galaxy with a few engineering friends and just enough seed money to rent a coworking space with exposed brick. You didn’t need a new idea. You just needed a whitepaper, a well-lit headshot, and a claim that your platform offered “more transparency” than the last five that said the same thing.

They all told the same story:
“We’re rebuilding programmatic from the ground up.”
“We’re putting the buyer back in control.”
“We’re different because we use machine learning.”

No. You weren’t. You didn’t. And you don’t.

What you actually built was a slightly re-skinned version of MediaMath, maybe with a fancier dashboard, maybe with a few cents shaved off the margin, but fundamentally no different. You plugged into the same SSPs. You chased the same bidstream. You ran the same campaigns for the same brands with the same leaky attribution models and declared victory because you labeled your pie charts “AI-Driven.”

And that’s the real problem: no one built anything truly new.

Sure, a few of you leaned into native. A few leaned into CTV. Some of you pretended to care about SPO. A brave few tried to wrap it all in identity graphs built on shaky email matchbacks and third-party cookies that were already halfway into retirement. But the result was always the same: a shiny dashboard for media buyers to fiddle with while the same broken pipes pumped junk inventory into the same overpriced auctions.

What followed was inevitable.

When things were frothy, no one noticed—or cared—that most DSPs were just middlemen with delusions of algorithmic grandeur. Budgets were flowing. CPMs were rising. Everyone was too busy buying panels at Dmexco and hiring evangelists to ask whether anything underneath the hood actually worked.

But then margins started to tighten.
Brand dollars stopped growing.
Measurement got more complicated, not less.
And suddenly the grown-ups—finance, procurement, ops—started asking questions the industry had no real answers for.

Questions like:

“Why are we using six DSPs when they’re all bidding on the same inventory?”
“What exactly are we paying for in these opaque tech fees?”
“Why is our ‘attention-optimized’ campaign still delivering 90% of impressions to games, trivia apps, and niche streaming channels that no human being watches on purpose?”

And just like that, the smoke cleared.

The emperor had no clothes. Or rather, the emperor had seven different DSP logins and no clothes.

Now, as we stumble into 2025, the consequences are finally catching up. One by one, DSPs are folding, merging, or slipping quietly into irrelevance. Some are being absorbed into holding companies desperate for a proprietary stack to sell to clients who stopped caring five quarters ago. Others are shutting down with a LinkedIn post, a Medium essay, and a list of “lessons learned” that carefully avoids mentioning unpaid vendors or the team they laid off over Zoom.

The ones that remain? They’re not necessarily better—they’re just bigger, or more entangled in the infrastructure of the walled gardens that now dominate everything. Google, Amazon, The Trade Desk—these are no longer just platforms. They’re gravity wells.

If you’re a buyer, you go through them not because you want to—but because at some point, you stopped being able to avoid it. Your campaigns touch DV360 because your video budget goes through YouTube. You use Amazon DSP because your brand sells toothpaste on Prime and they won’t share conversion data with anyone else. You tolerate The Trade Desk because they’ve wrapped enough feature sets in enough acronyms to convince your CTO it’s “strategic.”

The rest? They're clinging to niche use cases, trying to pivot into retail media, or just waiting to be acquired by a company that still believes “multi-touch attribution” is worth paying for.

This isn’t consolidation. It’s triage.

And here’s the dirty secret: most of these platforms weren’t needed in the first place. They were a response to overfunding, not demand. A fever dream of investors chasing martech multiples, and founders chasing exits.

The market didn’t ask for dozens of DSPs.
It didn’t need them.
It just didn’t notice until the money stopped flowing.

And now?
Now we’re left picking through the rubble.
Trying to remember which logos are still active, which were acquired, and which were just white-labeled clones of someone else’s tech stack to begin with.

So if you’re still clinging to the idea that there’s some great comeback story coming for independent DSPs, I have a freshly abandoned seed round and a job board full of ex-CTOs to show you.

The reckoning isn’t on the horizon. It’s already happened.

And if your DSP isn’t dead yet... someone is already measuring the casket.

🚨 The DSP Collapse Wasn’t a Mystery—It Was a Correction

🔍 The Accusation: They all claimed to be snowflakes. Turns out, they were ice cubes. For over a decade, Demand-Side Platforms screamed “differentiation,” while quietly building the same reskinned ad tool with a slightly different PowerPoint deck. And now? The thinning of the herd isn’t a tragedy. It’s overdue housecleaning.

📜 The Evidence:

Let’s rewind to the DSP golden age—2015. Over 40 platforms, all promising:

✨ “AI-Driven Optimization” (read: bid slightly higher on the users who clicked last time)
✨ “Full Transparency” (except for that 15% markup they buried in ‘tech fees’)
✨ “Omnichannel Access” (translation: we buy 90% of our media from Google, Amazon, and Facebook and pray you don’t notice)

The quiet irony? Real differentiation wasn’t in the tech. It was in the pipes.

Who had actual first-party, walled-garden, can’t-get-it-anywhere-else data?

Google: YouTube + Search + Display
Amazon: Prime Video + Retail Intent + DSP/Marketplace
Microsoft: LinkedIn’s B2B targeting + Netflix’s ad inventory
TTD: UID2 + OpenPath + the cult of Jeff Green

Everyone else? Just remixing IAB protocols with a prettier dashboard and hoping for an exit.

⚠️ The Catch: The “AI” pivot was a smokescreen. As things got tight, desperate DSPs started slapping “AI-powered” on their sales decks like expired USDA stickers on day-old sushi.

But the real machine learning magic requires data loops, not guesswork:

🧠 Amazon sees intent and purchase.
🧠 Google captures the query and the click.
🧠 TTD has identity through UID2—and can stitch media paths from bid to conversion.

The rest? Fancy macros in Excel.

🔥 The Big Question: Why did anyone ever think there would be 40 sustainable DSPs in the first place?

VCs funded lookalike clones because they thought adtech was SaaS with CPMs.
Agencies wanted “neutrality” to avoid being Google’s unpaid interns.
Brands were told they could “build their own DSP” like it was a Wix project. They could not.

🎤 Industry Vibe Check: Let’s talk losers:

Agencies: Built trading desks, then collapsed them back into the same 3 platforms. Now watching DSPs go direct while explaining to clients why they pay more for less.
DSPs without data: The me-too platforms are either begging for M&A or quietly pivoting to “media curation solutions” (read: another margin layer).
Resellers & Middleware: Congratulations, you now sit on top of a tech tax pyramid scheme, praying no one audits the stack.

💤 Sleeper Story: Microsoft While everyone was busy watching the usual gladiators, Microsoft quietly turned Xandr from a half-dead AppNexus carcass into a streaming and B2B beast. Netflix access? Check. LinkedIn targeting? Checkmate.

And unlike others, they didn’t need to shout about AI. They just built it into Azure, fed it intent data, and didn’t brag until it started printing money.

🧠 Thought Bubble: The great DSP shakeout wasn’t a downturn. It was a reset. A reminder that in adtech, those who own the identity and control the inventory don’t just survive—they write the damn rules.

🧹 TL;DR: The DSP graveyard is littered with platforms that thought a better UI could replace proprietary data. It couldn’t.

The winners?

Google (because it is the internet)
Amazon (because shopping is identity)
Microsoft (because B2B is finally sexy)
TTD (because cults scale faster than CRMs)

Everyone else?

Middlemen with a logo.

🪦 "DSPs Are Dead. Judy Brought the Shovel." How Judy Shapiro Skewered the Illusions of Adtech—and Didn't Even Break a Sweat.

When Judy Shapiro talks about DSPs, it’s not with the reverence of someone mourning a fallen comrade. It’s with the clinical detachment of someone who’s seen this movie before — and knows exactly how it ends.

Her view on the so-called “future” of DSPs? Forget the euphemisms. According to her, “the future of DSPs — including the dominant ones — is on life support.” And not the good kind where they’re waiting for a miracle. The kind where everyone’s just too polite to pull the plug.

She doesn’t tiptoe around the failures, either. DSPs didn’t just fall out of favor — they scorched the earth on the way down. “DSPs have burned their reputation beyond redemption due to fraud, lack of transparency, and accountability,” she said. That’s not a critique. That’s a final ruling. Game over. Turn off the lights.

And while everyone else is pretending AI is going to save the day, Judy calls it what it is: smoke and mirrors. “AI is neither artificial nor intelligent,” she quipped, before gutting the entire optimization industry with this one: “AI gives marketers a false sense of security, thinking the best optimization that can be done is being done. That is far from the truth.”

That sound you just heard? Half the C-suite realizing their dashboards are little more than glossy placebo buttons.

Judy also isn’t buying the idea that consolidation is some grand strategy. In her words, “weakened companies will gather together like strangers standing under an awning during a thunderstorm.” They’re not building the future. They’re just huddling until the rain passes. And spoiler: it’s a monsoon.

Even retail media — the golden goose of 2025 — isn’t safe from her scrutiny. DSPs hoping to sneak in through that side door? Judy sees right through it. “Nothing about DSPs now suggests growth or vitality,” she said flatly, noting that “eCommerce and retail marketers won’t use DSPs because the user experience of digital ads is not conducive to conversion campaigns.” In other words, if you think you’re going to pivot into retail media with the same old junk, you’re the punchline.

At the end of the day, Judy doesn’t need to scream to be heard. Her words are a scalpel, not a sledgehammer — and she’s performing adtech autopsies while the rest of the industry claps for the corpse.

Her diagnosis is clear: DSPs aren’t adapting. They’re dying.
AI isn’t rescuing anyone. It’s repackaging the wreckage.
And unless you're building for real discoverability, real performance, and real brand goals—
you’re not the future. You’re just another line on the liquidation spreadsheet.

Thanks for the clarity, Judy. The rest of us were still squinting through the fog machine.

You're missing ADOTAT+ because you’re still reading everyone else’s ad tech coverage like it’s a polite book club—and not the knife fight it actually is.

Let’s call it what it is:

Most of the newsletters you’re subscribed to are either:

💤 Regurgitated press releases dressed up as analysis.
💼 Vendor-friendly softball interviews that wouldn’t ruffle a feather in a TikTok influencer house.
🧼 Corporate sanitization stations, scrubbing every sharp edge off the truth until it’s just safe enough to send to procurement.

But ADOTAT+? We’re not here for safe. We’re not here for polite. We’re not here to protect your favorite DSP’s feelings.

We're here to:

Expose the real reason half the industry is playing musical chairs with LinkedIn titles.
Name the platforms that are circling the drain before they start ghosting clients.
Decode what those panic-buying PE firms are actually trying to build behind the curtain (spoiler: it’s not innovation).
Track layoffs, acquisitions, PR whitewashing, and stealth shutdowns in real-time—not three months later when the trades finally catch on.

And let’s be honest: the other guys?
They’ll call it “consolidation.”
We’ll call it what it really is: a fire sale in slow motion, with venture capital trying to flip the burning building before the insurance lapses.

ADOTAT+ is your seat at the grown-up table:

🧠 Deep Dives with actual reporting, not AI-drafted listicles.
📉 Financial breakdowns on who's about to miss revenue targets and why it matters.
🎙️ Insider leaks from execs, engineers, and sellers who know what's really going on.
🧪 Experimental predictions that actually land.

Oh, and we name names.
You’re not getting “a major DSP.” You’re getting: “Verve is a location-based zombie barely held together by GDPR violations and VC memory loss.”

You're not getting "a potential shakeup in SSP land." You’re getting: “Criteo is retail media cosplay with a 12% revenue drop.”

You’re not getting “transparency issues.” You’re getting: AdLib isn’t a DSP. It’s a Uniqlo hoodie wrapped around an empty Google Sheet.

Subscribe to our premium content at ADOTAT+ to read the rest.

Become a paying subscriber to get access to this post and other subscriber-only content.

Upgrade

Already a paying subscriber? Sign In.

🪦 Demand-Side Implosion: How the Entire DSP Category Became a Case Study in Overfunded Sameness and Underwhelming Innovation

Subscribe to keep reading

Keep Reading

Wireframe