AI Ad Creative: Revolution or Risk?

Ad structure and ad creative. 

If you’re marketing on Meta, this stuff takes up a huge amount of rent in your mind space these days. 

Finding “the right” bid strategy and account structure is an ongoing debate amongst the best and brightest in DTC. Cost Caps vs lowest cost continues to be a contentious debate, and unlikely to be resolved any time soon. 

That said, if you’re curious about testing cost caps in your ads manager, we share a framework on how to do it properly. 

Then there’s AI creative vs traditional - the newest battleground in the Meta ads arena.

The debate flared up this past week with the introduction of impressive AI-developed assets, grounded in the idea that DTC brands will need to find ways to continually drop the cost of asset development if the goal is to scale profitably.

Not everyone is convinced.  

Finally, we look at what could be a landmark FTC decision that could deeply impact the way brands manage UGC and “creator” advertising. 

This week:

• How to switch to cost controls

• Is the AI revolution here?

• The FTC ruling on testimonials and reviews in advertising

• Quick hits ← Find 15 great bits of advice from industry professionals

How to switch to cost controls

Andrew Faris, a cost-control advocate, shared how to switch from Auto-bids to his preferred method in 4 steps. 

But first, why does he think cost controls are better? 

Lowest cost = Meta spends my entire budget, everyday

Cost caps = Spend moderated by a profit goal 

Here’s Andrew’s 4-step plan to switch to cost controls.

1. Re-launch your old auto-bid campaigns at the same exact budget. 

This is your control group so you can compare it to the cost-controlled campaign you’ll launch. 

2. Launch the cost-controlled campaign with a 50% lower ROAS/CAC (your efficiency metric) than desired. 

This essentially creates a duplicate of the first campaign, which isn’t meant to be efficient. 

3. Tick your efficiency metric 5-10% closer to the desired level every day. 

You might see a reduction in spend as you go, but that’s ok. You’re spending more efficiently. 

4. Once you hit the target efficiency, increase your budget to the point where you never fully spend it again. 

Awesome. Thanks Andrew. 

How to know what efficiency to aim for -

It depends on the ad’s contribution margin. 

• The “unit variable cost” is the average cost per ad you produce. 

• Subtract this from the average revenue generated per ad, and you get the contribution margin, or in other words, the ad’s contribution to your net profit. 

• If it’s negative, set a higher efficiency target. 

• If you’re getting low to no spend, rethink your campaign. 

Takeaway: Despite the conflicting opinions on cost caps vs. lowest value, it’s important to remember – we have never settled on the one, clear way to win on Meta.

Every brand’s situation is unique, and what might work for another may not for you.

That’s why it’s wise to test all options. If you haven’t given cost caps a try yet, Andrew’s 4-step plan is a good way to get started. 

And remember…

Which brings us to…

Is the AI revolution here?

Dara Denney thinks the answer is…complicated.

Taylor Holiday is bullish.

You might remember his reasoning from our last issue: he’s seen no hard evidence that more expensive ads = better results.

• If a $1 static can do just as well as a $5,000 video, why choose the latter?

• AI makes great statics, which means you can use it to scale production and drive cost-per-unit down even further. 

• AI is not the only option for producing high-converting statics at scale. If you missed it, check out our feature on the Creative OS platform and how it helps marketers produce the best static assets, fast.

• There’s risk in high-cost creative, especially in an environment where volume and velocity of creative matter.

• If you need to make a lot of ads to test at scale, an expensive creative development process (both in time and money) can become a major hindrance.

Now for the counter-argument: go to this link and watch both videos. 

Here’s what Fraser Cottrell had to say about it:

The results are different, but which one performs? That’s what will truly matter.

Here’s an example of an AI vid crushing in an ad account:

(If you don’t turn the sound on, you don’t notice it’s AI.)

A truly compelling use case for generative AI is animated backgrounds.

You don’t risk the uncanny valley that comes with creating fake human models, and the results can be as stunning and eye-catching as a professional photo shoot.

Takeaway:

Creative development and testing have never been more important in marketing. You need to feed the algo as much information and variation as possible if you’re looking to find new audiences and new winners.

The implication is you need to find better, faster, and cheaper ways to do this in order to combat CAC-bloat and the costs associated with asset creation.

AI seems like it is opening doors for marketers to make more ads, faster, and at a better price point. The question is whether they will actually perform (or be legal…see our next section)

If you are considering investing more into ads, make sure you have a really good reason to. Preferably a data-backed one.

New FTC UGC Rules

Speaking of AI creatives, the FTC rolled out a new set of rules that make AI UGC basically illegal. 

Credit to Rob Freund for reporting on this.

 The FTC prohibits fake reviews and testimonials.

TL;DR —> you can’t make AI-generated UGC testimonials of a real creator if they haven’t actually tried your product and expressed the views shown in the creative. 

This is a major blow considering ad production spend is shifting toward AI and UGC: 

Other things you can’t do anymore according to the FTC:

• Buy reviews.

  • “leave us a review and get 20% off your next order”.

• Insider reviews

  • If someone with a material connection to the business leaves a review, that connection must be made explicit. 

  • If you own a review site and post reviews of your other company on it, this must be made explicit. 

• Review suppression.

  • A more common practice than you might think. Rob shares an example of a plastic surgery provider who included the following clauses in a client NDA: 

• Many companies also filter out negative reviews from their website. 

• Fake followers, i.e. buying follows from bot accounts. 

Check out this thread for a deeper dive on the FTC’s new rules.

Takeaway: Let’s call this “enforced authenticity”. With the popularity and effectiveness of UGC ads blossoming in DTC over the years, it was inevitable that brands and marketers would find ways to make them cheaper and at scale (ie; make facsimiles of legit testimonials). 

This ruling might save DTC marketing from itself, because eventually, the explosion of inauthentic content was going to backfire, with consumer trust inevitably dropping in the face of reviews that are not genuine. 

Quick Hits

• PostPilot, aka Meta for mail, is launching live on September 9th. 

• Advice from the pros:

  • 7 random bits of advice from @theisaacmed, founder of 7-figure @minikatana and @kanpaifoods

  • 3 bits of advice from @graceclarke on profitable marketing.

  • 4 bits from @RyanBartlett, co-founder of @trueclassictees on better funnels.

  • And a Meta ads tip from Marin Istvanic: organize your dashboard into a funnel. It’ll help you identify and solve blockers.

• Meta bug of the week - If you tried to launch Labor Day weekend ads on Wednesday of this week, there’s a solid chance your ad set just…didn’t serve.