Liquid Death dropped LiveRamp over one AI fear

Happy Tuesday.

Somewhere in your stack there's a vendor that knows which of your ads actually sell. Have you checked who owns it this week?

Benoit Vatere has. He's chief media officer at Liquid Death, and in August, weeks after Publicis Groupe's plan to buy LiveRamp went public, he stopped using LiveRamp's identity graph, Digiday reports. The deal has since closed.

Names and email addresses don't bother him, since Publicis won't see personal information. Buying patterns do. Publicis also works for brands Liquid Death competes with, like Evian and PepsiCo, and he worries what LiveRamp learned could end up in the data Publicis uses to train its AI tools.

"I don't want to train or have them see that super valuable information [for] anybody that I'm trying to get market share from," he told Digiday. The deal's terms say that can't happen. He expects terms to change, and the FTC's watch over the deal lasts only five years.

He isn't singling out Publicis. "I'm not mad at Publicis. I just don't want any of the holdcos to own it," he said, which rules out WPP and Omnicom too.

Independents don't get a free pass, because most are built to be sold. So whoever he picks has to sign a clause that ends the relationship and deletes Liquid Death's data the moment control changes hands.

Not everyone's with him. Ebiquity CEO Ruben Schreurs called the worry "a little overplayed." Forrester's Nikhil Lai went the other way: "privacy-safe and structurally neutral are no longer synonymous." And Digiday says competitors and non-Publicis businesses are roughly 5% of LiveRamp's $800 million annual revenue, so a few defections won't sink anything.

He still needs a replacement: the contract ends this year and he can't build his own graph. Digiday says rivals are already in his DMs.

Even if you think he's paranoid, the homework is cheap. Pull every contract with a vendor that sees your customer or performance data. Find the line on what happens if they're acquired and the line on whether your data can train anyone's AI. I'd bet at least one of those lines is missing.

If your identity vendor gets bought tomorrow, whose AI ends up learning from your customers?

Today’s sponsor:

Peep Laja, CEO of Wynter, on where AI research ends:

Wynter sells buyer research, and 200 is a small sample. I'd still ask someone on your team to read your homepage as a buyer who already knows your pricing and your rivals. What's left on it to convince them?

Up to 250%. How far above auction results Amazon allegedly pushed ad prices on days like Prime Day and Black Friday, in the unsealed parts of the FTC's complaint with 22 state attorneys general, SEJ reports. Up to 50% on ordinary days.

These are allegations, still being litigated, with examples from 2021 to 2023. Amazon has denied deceiving advertisers and says they never pay more than their maximum bid. If your benchmarks lean on those peak weeks, check them before Black Friday.

Find out which vendors hold your data and what you'd lose if one got bought.

Paste in: a spreadsheet of your tools with cost, contract end date, owner and what data each one holds. You get back: a list of overlaps, ownership gaps and exposed data, plus five ranked fixes. The full Marketing Tech Debt Audit adds friction points, revenue impact and a prioritized remediation plan.

Every time a vendor gets bought, the same scramble starts. Legal rereads the contract, ops tries to work out what can be exported, someone emails the account rep. Then it goes quiet until the next deal.

That's a design problem sitting in your procurement process, and Liquid Death shows why it's getting more expensive. The data you left with a vendor may be part of what got bought.

So fix it once. Have your ops lead and legal add two lines to every renewal: what happens to our data on a change of control, and whether it can train anyone's models. Put one person's name on the list of vendors that hold customer data.

Then test it. Twice a year, pick one vendor and actually export everything. You'll find out how locked in you are on a day you chose.

The AI Vendor Selection Framework for Marketing Leaders is our method for evaluating AI vendors across five dimensions: functional capability, integration requirements, implementation timeline, cost structure and vendor viability. It comes with a weighted scoring model and a red-flag checklist. Use it before your next renewal or new AI contract, when the sales pitch is doing most of the talking.

First step this week: build the matrix before you talk to anyone. List your actual stack, set your budget ceiling and decide how much weight vendor viability deserves after Liquid Death's exit.

→ We publish workshops, courses, guides, manuals, and more on AI marketing for free. Browse our learning resources for free.

You’re reading this, and so are nearly 20,000 of your peers. Some of them are your next customers.

If you sell to marketing teams, a single placement is €500 and a full week is €2,000.

If you want to place an ad, simply reply to this email or check our media kit.

Want to see case studies of AI-powered campaigns? We publish a daily edition on the most creative use cases of AI in advertising.

Interested only in AI-powered revenue and sales? Our new daily publication focuses only on GTM strategies powered by AI.