We ran Liquid Death's ad library through an MCP. Their brand voice barely shows up

Liquid Death is the brand every marketer points to when they talk about voice. Gothic can labels, unhinged social captions, 7.75 million Instagram followers built almost entirely on organic chaos. So we pulled their live Meta ad data through adextract's MCP to see how that voice actually shows up in paid media. It doesn't.

What we pulled and how

adextract connects to the Meta Ad Library through a single MCP endpoint. Two calls were enough for this teardown: one to search every active ad running under Liquid Death's Facebook page, and one to run creative analysis on the ad that looked most like their actual brand voice.

No manual scrolling through the ad library. No screenshots. Just two calls:

await search_meta_ads({ page_id: "liquid death", country: "US" })

await ad_reader({ url: creative_url })

The first call returned 229 active US ads. We analyzed the top 30, sorted by impressions, since that sample carries the most weight in what people actually see.

Only 13 percent of their ads point to their own site

Of the 30 ads analyzed, 4 linked to liquiddeath.com. The other 26 sent traffic straight to Amazon, Walmart, or Target. For a brand with a fully built ecommerce presence and a following in the millions, that split is unusual. It suggests the paid budget is optimized for retail sales velocity, not for building a direct relationship with the customer.

This matters for anyone running a similar model. If your paid spend routes through a marketplace, you're renting the customer relationship from Amazon or Walmart every time an ad clicks. The brand equity might be yours. The transaction data usually isn't.

97 percent of the ads use the same button

29 out of 30 ads used the exact same call to action: "Shop now." One ad used "See details." That is it. No jokes, no gothic flourishes, no reference to the brand's entire public identity. The CTA that shows up most on a brand famous for never doing anything ordinary is the single most ordinary CTA in advertising.

This isn't a criticism of Liquid Death's marketing team. It's a reminder that CTA testing tends to converge on "Shop now" because it performs, regardless of how loud the brand voice is everywhere else. Voice and conversion copy often optimize for different things.

93 percent of the ads have no custom creative

Digging into the ad payloads, 28 of the 30 ads had no custom creative attached at all. The headline and body fields were literally template variables, product.name and product.brand, filled in dynamically from a retailer's catalog feed. There's no copywriter behind these. There's a feed and a rules engine.

This is dynamic product ads at work, and it's extremely common for DTC brands running performance campaigns through retail partners. The takeaway for anyone auditing a competitor: don't assume a high ad count means a high creative production budget. Most of what you'll find in an active ad set is templated.

What their one real creative ad actually says

Among the 30 ads, exactly one had custom photography and copy: a flat-lay image promoting Liquid Death Sparkling Energy alongside a pink cardigan and jewelry, styled as a lifestyle shot. We ran it through adextract's creative analysis.

• Primary emotion: anticipation
• Marketing goal: consideration
• Message archetype: aspirational, benefit and feature

None of that reads as shock, chaos, or gothic humor. The one ad in the set that got real creative attention is calm, aspirational, and built around a clean label pitch: vitamins, no sugar, clean energy. It's the opposite of the brand's public persona. Which raises the real question: is the media buying team optimizing for a different customer than the one following them on Instagram?

The gap between brand voice and paid media isn't unique to Liquid Death

Organic and paid are usually run by different teams, on different timelines, against different KPIs. The organic team is rewarded for engagement and cultural relevance. The paid team is rewarded for return on ad spend. Liquid Death is just an unusually visible example, because their organic voice is loud enough that the contrast is obvious the moment you look at the actual ad library.

If you want to see how this shows up across a full account, not just one brand, our breakdown of how AI agents surface a competitor's best performing ads covers the workflow in more detail.

How to run this teardown on any brand

The same two calls work on any advertiser with an active Meta page, and the same MCP server covers Google, TikTok, LinkedIn, and Reddit ad libraries too. There's no manual ad library scrolling, no screenshot folders, no spreadsheet of CTAs typed out by hand.

• search_meta_ads(page_id) returns every active ad for a page, sorted by impressions
• ad_reader(url) returns structured creative analysis: emotion, marketing goal, archetype, target audience
• The same pattern works across Meta, Google, TikTok, LinkedIn, and Reddit ad libraries

For the mechanics of how the scanning itself works under the hood, see how adextract's agents scan millions of ads across platforms.

If you're studying a brand for inspiration, the ad library tells you more than their social feed does. Before assuming a competitor's public voice is their whole strategy, check:

• What percentage of their ads point to their own site versus a retail partner
• Whether their CTAs are actually distinctive or converge on the same one or two options
• How much of their active ad set is custom creative versus dynamic catalog templates
• What emotion and archetype their real creative work is actually optimized for

Fifty free searches on adextract.co is enough to run this exact teardown on any brand you're studying.

Frequently asked questions