Generative AI is a miracle tech.

 It can produce poems, novels, and even beautiful images or videos in seconds. It’s a profound leap forward in terms of creative development. There’s little doubt that AI will 100X or even 100X our ability to make amazing stuff in the near future. Heck, we made the lead image for this issue in about 10 seconds with ChatGPT.

There are trade-offs in everything though. 

Now, or very soon, almost anyone will be able to spit out a Pixar-style short. Or a Picasso-style painting. Celebrity's faces can be plastered on anyone’s body, or their likeness can be dropped into the middle of a piece of content, without the celeb ever having been involved. 

In short, we’re also facing an explosion of mimicry, replication, duplication, and fakery. 

Art and creativity are valued because they are unique, singular, and noteworthy. Once you strip those things away, it becomes commonplace, uninteresting, and ignored.  

Just a word of caution as we take our first few shaky steps into the brave new AI world. 

In today’s issue:

• TikTok allowing deepfakes

• Here’s how to run ads on TV, but do you even want to? 

• How bad was the iOS14 for advertising? A study

Yay deepfakes (boo, actually)

TikTok is enabling the use of deepfakes. 

Brands will be able to buy ads featuring AI imitations of generic people, OR OF SPECIFIC CREATORS. 

It’s happening, folks. 

• They claim it’s to accelerate content production and help brands access different language bases more easily. Which deepfakes totally do, no debate there, but once the floodgates open, they will be hard close. 

• We don’t know if you’ve seen deepfakes recently but they’re near indistinguishable from reality, especially to the scrolling eye. 

• We’re entering a world where the only thing stopping us from putting unapproved words in someone’s mouth are platform policies and regulations. 

But enough doomsday talk. There are some fixes in the works.

ByteDance (do we have to keep saying they’re TikTok’s parent company?) are banning the use of any synthetic media that contains the likeness of any real private figure, i.e. you. They’re also banning anything that mimics a public figure endorsing a product or violating other policies like hate speech. 

Solid fundamentals. 

Next, they’ll require any deepfake content to be labeled as such. Problem solved? For the general public and people whose likeness may be used, kind of. 

For advertisers, not exactly. 

Let’s talk about the entire purpose of user generated content (UGC) for a second, A.K. A. influencer and creator marketing. 

This content type is now a fundamental pillar in DTC marketing. According to Goldman Sachs, the so-called creator economy and its total addressable market could double from $250 million to half a trillion by 2027. 

The reason it works is because it’s a product endorsement from a third party (i.e. not from the brand), which feels genuine. We go deeper into the topic in our Issue #11: Keeping it real. 

Do you see the problem? 

If you flood the market with fake people mouthing fake reviews and testimonials, it will undermine the value of UGC. If consumers can't tell the difference between real people recommending stuff and fake people recommending stuff, they will tune it out.

Sure, it’s easier for DTC operators or marketers to create and disseminate, but that's what would make it less valuable. UGC/testimonials/social proof is not supposed to be effortless. 

It would be like painting a bunch of rocks with gold paint during a gold rush.

Is there anything you can do about it? Not really, aside from hope that authentic content will become that much more valuable. As long as we can tell it from the deepfakes. 

Takeaway: The “U” in UGC is “User”. If a user isn’t making it, it’s not UGC. It may look and feel like it, but eventually, the audience will catch on and tune it out. 

One of the great paradoxes of marketing is that a rush to leverage a given tactic can eventually cause cynicism or exhaustion in the audience, especially when it is perceived as inauthentic.  As in all things, you should try, test, and analyze, so we’re not saying “don’t do this”. 

Just a note that a tidal wave of fake “UGC” could lead to a complete devaluing of it.  

TV ad(vice)

Have you ever thought about advertising on TV? Like the original thing? 

Might not be worth it for DTC brands nowadays with its high production cost, declining viewership, limited targeting, and elusive attribution. 

But it does make you look legit. More credible. 

It’s also a place where people’s attention spans are longer than when online, so advertisers can go deeper on a subject. 

What about streaming platforms like Netflix, Hulu and Disney+ ? Targeting and measurement capacities are on par with your usual Meta and Google ads, but you face the same challenges of channel saturation. 

Regardless of which you choose and whether you’re in it yet, Sarah Carusona has some advice for you. Who, you ask? In the words of Sean Frank from Ridge, she “has built and ran every aspect of ecom brands from 2-200m”. 

According to Sarah, here are the questions she says you should ask your agency/buyer:

1️⃣ What is the attribution window for traffic and for conversion?

The window for traffic can be anywhere from 5 minutes to 30 days. The window for conversion often starts after the initial traffic was attributed, and can be another 7 to 30 days. 

2️⃣ What is your overall monthly frequency and reach? 

Frequency refers to how often your ad is shown to the same audience, while reach is the amount of unique viewers. If your frequency starts to increase, you should look to expand on other networks, shows, or times of day.

3️⃣ If running on streaming, what percentage of your budget is going to display or mobile-only? (Be careful here with some agencies that like to hide this.)

Here’s why you should care about this. 

4️⃣ What time is most of your traditional tv budget going into? 

If it’s all at 2am, you’ve got a problem. 

5️⃣ What categories perform best for your brand? 

Your buyer should be able to explain this from a high level. They should also know which creative works best for each category. 

It isn’t standard practice, but good buyers/agencies will bucket the networks and shows you advertise on into categories like:

• News

• Information

• Entertainment

• Sports

• Comedy

• Etc

Then they’ll look at how different creatives performed within each bucket. E.g. an emotive 30-second ad might have done well in the home improvement category. 

6️⃣ Ask for brand lift survey opportunities. 

It’s always good to run your own tests because of the unique time, place, and position you are in. 

Some agencies provide these free of charge. 

7️⃣ Do they measure Amazon impact? This is a huge plus if they do. Sarah knows a few agencies that are doing this or planning to (try to DM her if you want recommendations.)

There you have it. Your run-DTC-ads-on-TV starting kit. 

One more thing before you leave. 

If you’re thinking about switching to TV because of the attention span thing, you should read this -  

Craig Clemens claims that either you or someone you know has bought something from him. He recently went viral on Twitter for his controversial take that “Short attention span is a lie!”. 

While most of the industry believes that TV is the only place to run long ads, one of his most successful ecomm ads was 1 hour and 15 minutes long. 

His point is, if viewers believe you can help them with their problem, they will stick around. 

“And this is true for short ads, too.”

Takeaway: Sarah Carusona teaches you which questions to ask if you decide to run TV ads, whether linear or digital. However, if you’re considering TV because it's a better place for longer-form ads, just know this: long-form can work on digital too. 

If you know what keeps your customer up at night and can create engaging creative around it., that is.

How bad was iOS14 really? 

A commonly held belief in the advertising industry is that Apple’s iOS14 privacy measures completely crippled Meta ad effectiveness. 

Now, we have some actual evidence. 

Daniel McCarthy, Guy Aridor, Yeon-Koo Che, Brett Hollenbeck, and Max Kaiser studied the impact of Apple’s App Tracking Transparency (ATT) privacy measures on advertising effectiveness and published the results in February this year. 

iOS 14 updates severely limited the ability to track user activity across apps and websites, thereby making it hard to effectively target users and attribute conversions. 

The study looked at revenue and ad spend data for 4,243 companies, before and after ATT. 

Their observations: 

1. Ad effectiveness significantly lessened. Of note, a 37.1% reduction in click-through rates because of poor targeting. 

2. Meta’s spending share fell by around 7% over the year following ATT

3. The more brands relied on Meta ads and spend from iOS devices, the more revenue they lost. 

Takeaway: As we noted last week, Meta continues to evolve its ad offerings so advertisers can better understand not just broad sales conversions, but also incremental lift. 

It’s easier said than done in the post-iOS14 landscape, however. This year has been perhaps one of the least stable in terms of Meta performance for many brands on the platform, suggesting the battle is ongoing.

Quick hits

• Alex Garcia reports that Poorboy Coffee has added ~40,000 followers in 4 months, with an engagement rate 10x that of competing coffee brands. The secret? Founder POV creative. 

• Chris Long, VP of Marketing of SEO firm @GoFishDigital, shares his top five takeaways from the Google Search API leaks. 

• The eight-year rollercoaster ride of Care/of:

• A how to guide for auditing Meta accounts: 

• Four mistakes to avoid when expanding to retail.