Gareth Hates AdTech

@airgups23 Teads!

Given the upfronts (TL;DR, it's becoming more and more about targeting / data / AI, less about premium content), and Q1 earnings (indie ad tech struggling, legacy media doing meh, big tech crushing), here's what we think the current industry pecking order looks like. From this morning's @UofDigital newsletter drop. Subscribe here: uof.digital/newsletters

@EricTilbury_RTB There’s the part of your life when you use an electric toothbrush and then the part of your life when your teeth had fur.

Oral B electric toothbrush is life changing btw

If I had an article for every time I fixed something in an integration and didn’t make more money, I’d have at least 2 articles.

@jamesIII @JudSpencer No, bid density is much higher than that, and if bid density is low dsps tend to know and use it for shading. Second price is superior in almost every way

@HatesAdtech @JudSpencer Agreed on the inefficiency of bid shading, though doesn't the average bid density on most actions contain just one bid? Floor price + $.01 wouldn't be a welcomed return for pubs

Sometimes, when I’m lying down next to my son, I think to myself “man, won’t it be great when the DOJ forces adx to bid into prebid and we can centralize adserving in prebid and go back to vickrey auctions?” It sure will be.

What features make a dsp? While this headline gets people fired up, they often have the wrong takeaway — the list of features that bedrock possesses are now a list of features that index possesses, usable in conjunction with other features and integrations that index possesses. This is a step in modularization — you won’t have “lots of dsps” you will have agents selecting vendors campaign by campaign from a laundry list of supported features inside of the index meta. This is not dissimilar to a trafficker checking boxes in a dsp ui today.

Is the agentic buying future going to be made up of 100s of containerized DSPs on agentic rails where a buyer agent can optimize between all these different DSPs for one campaign? Each DSP basically being its own customized bidder with its own logic… someone tell me I’m wrong

@JudSpencer (Btw I know you don’t think this, I just hear this take constantly when I talk about programmatic)

That’s a pipes problem not an inventory problem. I grow weary of the armchair psychiatrists applying freshman level intuition to “why users browsing Facebook respond to ads more”. It’s A. Midwit and B. A disservice to the literally hundreds of thousands of domains and their diversity of experience. Good thing I have data that can help fix it🤣

@JudSpencer We can but observe Facebook display cpms to surmise where the equilibrium might be. Those cpms are very high.

@HatesAdtech Hmm.

@JudSpencer If our market doesn’t increase scale (ie, increase budgets and increase participants) with better performance from efficient cpms (maybe lower sometimes, maybe not, just the appropriate valuation in the appropriate moment) we don’t have a market we have a racket.

@HatesAdtech So, lower CPMs. Who’s going to agree to this?

The former. I don’t have a problem with bid shading, it’s an expected (almost desirable?) behavior. My problem with first price auctions is the asymptotic inefficiency of bid shading — no matter how good you get at it, it’s still inefficient to some degree. Second price auctions are brilliant in their “perfect bid shading” design.

@HatesAdtech Is it because you think second price auctions are superior or because you really hate bid shading?

@mikeroplastics @daveasprey Privacy is for apes.

@HatesAdtech @daveasprey I love this. Also would love doing all interviews by livestream

The New York Times mentioned me in an article about nicotine. They're skeptical. Here's what they didn't include... Dr. Paul Newhouse at Vanderbilt ran controlled trials showing nicotine improved concentration and cognitive function in people with mild cognitive impairment. Peer-reviewed, published research. There are 30+ epidemiological studies showing an inverse relationship between nicotine and Parkinson's disease. Consistent enough that researchers are actively studying it as a potential protective mechanism. I've invested in clean nicotine and I've been open about that. I've also been pointing to the research on nicotine for years. The Times had a narrative. The studies are worth reading on their own terms. Form your own opinion about nicotine.

@mikeroplastics @daveasprey Would be a boss move to “offer a live interview in the form of a twitter thread”

@daveasprey A good rule when mainstream media reaches out: Email them one sentence that you’d like to see in the article. Otherwise they’ll just selectively quote or misquote to fuel their narrative.

And for an ad tech specific thought, I no longer am sleeping on agencies. I think they have more agency than they have in the past, and they are the primary beneficiaries of the modern advancements in AI. The industry just hasn't realized it yet because they haven't flexed yet...but they will.

I sure am! What spurred the thought was the disruption of seamless -- restaurant licensed delivery services that instead of charging a % of carts charge a licensing fee. And a new Uber competitor that just came online, where instead of charging a % of ride they charge drivers a licensing fee. It's universal. We're on the same page. If you are charging a wildly variable fee completely unrelated to your costs of delivering the service, you will eventually be disrupted. It's a question of when.

Fees that vary independent of delivery cost are inherently fragile, it’s a question of timescale not a question of sustainability.

@JudSpencer Was feeling cryptic, might delete later.

@HatesAdtech I wish you used more words here.

@aripap @JonathanMoffie @AdtechGod Is a vibe generator a vibrator?

@aripap @JonathanMoffie @AdtechGod So much biz is vibe based, buying a vibe generator seems reasonable

Bullish for @aripap and @AdtechGod selling Marketecture to TTD to turn the PR ship around

They’re not going to divest and the product person in me doesn’t feel it’s necessary, but the business person in me knows the practicalities of policing fairness and good faith participation in biddable environments necessitate the divestiture. There are simply way too many plausible “legit” reasons to hide naughty behavior behind.

Betting odds on the Google antitrust remedies by @keithepetri keithpetri.com/2026/03/29/if-…