According to Victoria Griggs, how many hidden Meta AI ad settings does she recommend turning off, and what happens to an ad's social proof when you edit it after it starts running?

Answer at bottom of email

Victoria Griggs spent 17 years inside the corporate consulting world, cutting her teeth at Accenture back when it still operated under the Anderson Consulting name. It was stable work. It was also, by her own account, not remotely what she wanted to be doing with her career for another two decades.

In 2012, when Facebook advertising was still a genuinely new and somewhat chaotic frontier, Victoria made a decision that looked reckless to almost everyone around her. She walked away from corporate consulting entirely and built an agency, Straightline Marketing, with a single-minded focus: Facebook ads, and only

Facebook ads. No diversification across platforms, no hedge into search or display. Just one channel, mastered deeply, while most agencies at the time spread themselves across everything.

That bet paid off because Victoria treated the platform less like an ad buying tool and more like a system to be reverse engineered. While competitors were still thinking in terms of simple audience targeting and basic creative rotation, she was digging into the settings most advertisers never touched, the ones buried three menus deep that quietly shape how an ad performs without ever appearing on a results dashboard.

Years later, that same instinct is what led her to identify a problem almost nobody in the industry was talking about publicly: Meta's own AI-driven ad automation features, turned on by default, were quietly undermining the very campaigns advertisers were trying to run.

Features meant to help were instead eroding social proof, diluting targeting precision, and confusing the platform's own algorithm.

Victoria's core lesson, earned over more than a decade of watching platforms evolve: Read the settings. Every single one.

This week pairs Victoria Griggs' deep dive into the hidden Meta AI settings quietly sabotaging advertiser campaigns with Rich Kahn's sobering breakdown of just how much of global digital ad spend disappears into outright fraud.

Different problems on the surface, but the same underlying warning: modern advertising runs on systems that were not built with your specific interests as the only priority, and the burden falls on you to audit what is actually happening behind the dashboard.

Victoria's research uncovered seven specific Meta AI settings, enabled by default, that actively work against advertiser goals. Rich's research, from a completely different angle, found that global digital ad spend exceeded $700 billion for the first time recently, with an estimated $140 billion of that, somewhere between 20 and 25 percent, effectively stolen through various forms of ad fraud.

Watch These Misfits Episodes to Dive Deeper:

Exclusion audience architecture is a defensive advertising strategy built around a simple reframe: instead of only thinking about who you want to target, build a deliberate, continuously updated list of who you want to actively exclude.

This matters more than most advertisers realize because both ad fraud and simple targeting waste often come from the same overlooked source, audiences that look statistically similar to your ideal customer but never actually convert or engage authentically.

Why this works: platforms like Meta and Google optimize toward whatever signal you feed them, including bad signals. If fraudulent clicks or bot traffic are quietly counted as legitimate engagement, the algorithm learns to chase more of that same low-quality traffic, compounding the problem over time rather than correcting for it.

How to build exclusion audience architecture in practice:

The payoff compounds over time.

A clean signal feeding the algorithm produces better lookalike audiences, more accurate optimization, and ultimately a lower blended cost per acquisition, because you are no longer training the platform's AI on garbage data mixed in with your real customer signal.

Every list ages, and most brands do not notice the slow decline in deliverability until open rates have already cratered. Internet service providers track engagement patterns closely, and a list full of subscribers who never open your emails signals to those providers that your sends might be unwanted, which can push future emails into spam folders even for your genuinely engaged subscribers.

The first step in a deliverability rescue is segmentation, not sending. Separate your list into subscribers who

have engaged (opened or clicked) in the last 90 days, those who have gone quiet for 90 to 180 days, and those who have been completely dark for more than 180 days. Each group needs a different approach, not a single blanket campaign.

For the 90 to 180 day group, run a genuine re-engagement sequence with a clear, low-pressure ask: a simple "are you still interested" email with an easy way to confirm continued interest. For the group dark beyond 180 days, consider a final win-back attempt with your strongest possible offer, followed by removal from your active sending list if there is still no response.

Do not be afraid to actually remove unengaged subscribers permanently. It feels counterintuitive to shrink a list on purpose, but sending fewer emails to a cleaner, more engaged list almost always outperforms sending more emails to a bloated, unresponsive one, both in immediate revenue and in long-term deliverability health.

Monitor your sender reputation directly through tools your email service provider offers, and pay attention to spam complaint rates specifically, since a rising complaint rate is often the earliest warning sign of a deliverability problem before open rates even begin to visibly decline.

For help with email marketing for your brand, contact dragon.fish

The Brand: Athletic Brewing built a genuinely new category, non-alcoholic craft beer, in a market that had every reason to be skeptical that such a product could actually taste good.

The Strategy: Rather than leaning on traditional beer advertising tropes, the brand built its early growth through a community-first approach targeting athletes, parents, and people moderating alcohol consumption for health reasons, groups that traditional beer marketing had almost entirely ignored.

The strategy worked because the brand treated its underlying audience insight as the actual product differentiator, not just a marketing angle.

Rather than positioning the beer as a compromise or substitute, Athletic Brewing built messaging around performance and lifestyle enhancement, appealing directly to a demographic that wanted to stay social without the tradeoffs of alcohol.

The Result: Athletic Brewing became the top-selling craft brewery in the United

States by dollar sales within a relatively short window after launch, an extraordinary outcome in a beer market that had shown little previous appetite for non-alcoholic options at that price point and quality tier.

The Lesson: The steal-worthy takeaway: identifying an underserved audience that existing category leaders have overlooked, then building a product and message specifically for that audience, can outperform trying to out-market established players on their own turf.

Not every whiskey barrel gets used just once, and the difference between a first-fill and a refill barrel changes flavor more dramatically than most drinkers assume. American bourbon regulations require new, charred oak barrels for every batch, which means a huge supply of barely-used American oak barrels becomes available to other whiskey-producing regions once bourbon distillers are finished with them.

Scotch and Irish producers frequently purchase these used bourbon barrels specifically because the first fill already softened the harshest wood tannins, allowing a gentler, more balanced extraction of vanilla and caramel notes on the second use.

A barrel can realistically be refilled and reused for several decades before the wood is considered fully spent of flavor-contributing compounds, and some producers specifically seek out third or fourth fill barrels for spirits that need extended aging without picking up overwhelming wood character.

This global barrel trade, bourbon barrels flowing outward to Scotland, Ireland, and increasingly Japan, means a bottle sitting on a shelf thousands of miles from Kentucky may have aged inside wood that once held an entirely different spirit.

The next Collective Minds Society Cigar & Whiskey trip is Feb 18-22, 2027

Use AI to help structure a disciplined creative testing process before you ever touch Meta's automated creative tools, following the five-document prep approach Victoria Griggs uses with clients.

1. Gather your sales page transcript, a sample of your brand's spoken tone-of-voice content, your customer avatar documentation, a short promo description, and any original research or data you have collected about your audience.

2. Feed all five documents to an AI model in a single session so it has full context before generating anything.

3. Ask the model to draft three distinct ad angles, each rooted in a different piece of the source material rather than generating generic variations of the same idea.

4. Request that each angle include a specific hook, a body that addresses one real objection from your avatar documentation, and a call to action matched to your actual offer structure.

5. Manually review each draft against your original sales page transcript to confirm the AI has not introduced any claims your business cannot actually support.

6. Test the strongest two or three angles in small-budget campaigns before committing significant spend to any single creative direction.

Sample Prompt:

"Using this sales transcript, tone-of-voice sample, customer avatar, and research summary, draft three distinct ad angles that each address a different real objection from the avatar documentation."

Feeding AI real source material beats asking it to invent your brand voice from a single-line prompt, every time.

Q: What happens to an ad's performance data if I edit it after it has been running?

According to Victoria Griggs on Meta's AI is Ruining Your Ads, editing a live ad wipes out all of its accumulated social proof, including likes, comments, and shares, effectively forcing the ad to rebuild momentum from zero.

Q: How large is the global ad fraud problem?

Rich Kahn estimates roughly $140 billion, or 20 to 25 percent, of the more than $700 billion in global digital ad spend is effectively stolen through fraud.

Q: What is the single fastest free fix for reducing ad fraud exposure on Google?

Unchecking the "Partner Networks" setting in Google Ads is the quickest, no-cost step advertisers can take, since Display Network fraud rates run significantly higher than search-only campaigns, per Rich Kahn.

Q: How much can a webinar replay option improve attendance and ROAS?

Victoria Griggs notes that offering a just-in-time replay can lift effective attendance from roughly 22 to 25 percent up toward 75 percent, moving return on ad spend from around 50 percent toward 200 percent in her client data.

Victoria Griggs recommends turning off seven hidden Meta AI ad settings, and editing a live ad resets all of its accumulated social proof back to zero.

She warned advertisers that small edits to a performing ad can undo weeks of built-up engagement momentum.

Watch the full conversation