Will AI finally kill UA?
According to Twitter, AI will code the next billion dollar SaaS product from scratch. AI will also write a Drake song better than Drake himself. Fuck it – AI might be writing this entire article (at least one of these are true…the Drake one).
AI is so hyped right now that it almost feels like if a CEO just says the phrase “generative AI,” their stocks will reach an all-time high. But here on my blog, we talk no-bullshit. In this article I want to go over four key things about AI in mobile app marketing:
- What is happening in AI right now?
- What AI tools are out there right now?
- How effective is AI in the world of UA?
- How will AI affect UA managers in the future?
1. What is happening in AI right now?
Every tech company is putting AI front and center of every earnings call, conference, and company strategy meeting. We’re entering the ‘wild wild west’ of AI, where everyone is throwing shit on the wall and seeing what sticks.
Here are a few examples:
- Snapchat star Caryn Marjorie launched a chatbot and over 1,000 simps signed up to be her virtual boyfriend for $71,610 in a week. She charges $60 an hour to talk to a bot.
A fake AI song (with 15m TikTok views) purporting to be Drake and the Weeknd got pulled from streaming services by an anonymous TikTok user that goes by Ghostwriter977.
- An AI voice addon was created for Classic Wow, revitalizing the game. Here’s a streamer’s reaction.
- AI-generated photo of the Pentagon exploding went viral and briefly fucked up US markets.
- NVIDIA showed a demo of a generative AI NPC, which they call ACE (Avatar Cloud Engine)on Unreal Engine 5 to have scarily off-script conversations.
- AI-generated audio of Obama, Biden, andTrump trashtalking while gaming is taking over TikTok
- Photoshop’s AI generative fill feature is so good, it’s creating full sized images of popular memes.
- An entire anime was generated with AI.
Okay but more relevant stuff to UA:
Meta announced that they’re shifting investment away from the Metaverse to AI, so get ready for more AI-powered creation and optimization tools.
While Meta’s targeting capabilities dwindle with post-IDFA changes, this may mean a bigger black box in terms of forcing advertisers to use automated campaign optimization. However on the creative side, Meta announced the AI Sandbox to duplicate ads, change backgrounds, generate images, adjust brightness, and crop photos.
Google’s I/O conference was all about AI – every single Google tool has an AI feature (regardless if it’s necessary or not). The big news is that Google will offer generative AI tools like images, videos, and text options to generate ads on Google Ads. Google search will be powered by Bard, and competition for organic results will be worse than ever – SEO will see a painful fall.
ChatGPT will launch plugins soon, so UA managers can input campaign data, creatives, and reports to build powerful acquisition strategies in real-time. Airtable’s generative AI features will make it easier than ever to write and build out campaign briefs and automate workflows. We’re finally moving on from static templates.
But most of this is theoretical – how good these tools will be is unclear. But for now, it sounds pretty promising.
2. What AI tools are out there right now?
There are a lot of empty promises and Twitter threads claiming that 99% of people are using AI wrong. But what are the actual tools that can help UA managers today?
AI improves speed of production, not quality. AI isn’t going to replace your expensive ad agency that writes and films your Nike ads but AI has essentially wiped out the Fiverr marketplace. Cheap and quick production is now automated – and actually may be even better than a $5 job abroad.
Look, Coca-Cola just pulled off one of the coolest ads of the year, but it was the world’s best creatives who masterfully used the tool to make it happen.
AI images and copy are almost there, but they have a very distinctive tone and style in the uncanny valley territory. But nothing will replace industry expertise – having a pulse on what’s working in the industry and how to strategically test new creative directions is a skill. Now there’s a valuable new skill – competency with AI tools.
Here are some of the tools that are actually being used by UA managers right now:
- AI avatars that look like humans where you can script what they want to say: Synethesia, Deepbrain, Synthesis
- Video editing software that can automatically cut and organize your videos: Invideo, Pictory, Veed.io, Adobe Premiere
- Multivariate and creative ad testing: AdCreative.ai, Pencil, Marpipe
- Copywriting: ChatGPT, Jasper, copy.ai
- Generative design: Dall-e, Photoshop, Canva, Memorable AI
- Audio engineering: Adobe Podcast, Adobe Premiere, Voice.AI, Uberduck
- Website builder: Headlime, Dora AI
- Epic intro for 2.5 gamers no-bullshit gaming show
AI-generated ad examples
We’re still in the early stages of seeing AI being implemented into UA creatives, but here are a few that I’ve found in the wild.
- Survivor.io
The extremely popular 2D battle royale game is testing out an interesting AI model in this ad. Already talked about it here and here!
- Sssnaker
Similarly, SSSnaker is copying the strategy. We’ll let you decide who did it first, of course… Well, its the same company after all! Right?
- Horselife
AI-generated voiceovers are super popular in the ads world, and this ad by Horselife is a prime example. It seems to be a more popular strategy on TikTok, though.
- Messi visits Saudi Arabia
The World Cup Champion’s face is superimposed on some models in an ad about visiting Saudi Arabia. This raises some questions about the future of image rights.
- Charity Right
The UK-based charity is experimenting with some AI-generated creatives to help feed hungry and neglected communities around the world. Using AI-generated starving children though brings up some interesting ethical questions.
We could go on and on and on with the use of AI in creatives. I am pretty sure you understand the use & the importance!
3. How effective is AI in the world of UA?
The most advanced use case for AI is copywriting. ChatGPT’s explosive popularity is the reason why this AI movement started in the first place. So naturally, generative AI focused on language is strongest.
While Chat-GPT4 is more powerful than ever, it still falls short on long-form content unless you feed it tons of your own research. But it’s super easy for short-form copywriting for ads. Ads can also be a lot less awkward from international markets, as it just takes a few seconds to correct the grammar using a generative AI tool or even an AI translation tool like DeepL.
Most mobile UA ad creatives aren’t extraordinary in the first place, so AI does the job for now.
4. How will AI affect UA Managers in the future?
While some are claiming AI is having an ‘iPhone moment,’ changing the course of technology forever, I’ll try to be more grounded. On the creative side, there is a lot of upside for UA managers. If you can dream it, you can create it. Teams can be lean and be more productive.
But for now, we’re still in the ‘uncanny valley’ territory where you can tell whether or not something is AI-generated (they can’t seem to generate hands right). We’re so close though.
Paid Advertising
AI testing and optimization tools embedded within each channel (Meta, TikTok, Google, etc) will prove to be more and more worthwhile, especially if they can manipulate and optimize the ad creatives themselves. The best part of it all is that reporting, forecasting, and optimization will continue to be more and more sophisticated, so you have data to make important campaign decisions without a Master’s degree in data science.
More data means you’ll have a better understanding of your market. And in a world where targeting is becoming more and more difficult, getting the creatives right is more important than ever. I recently covered some UA tips on my substack. Something I drill down frequently is iterating a winning ad within six dimensions:
- Attract attention
- Branding & gameplay
- Connect with viewer
- Design to be seen
- Ending & direct CTA
- Formats & networks
AI will eventually get to the point that can extract each of these elements and algorithmically tweak them to perform better, giving UA managers a superpower.
My predictions
In the same newsletter I make some other predictions that I’d like to reiterate here. I think gaming companies are targeting shorter payback periods during times of economic uncertainty. They may have targeted D365 profitability but will expect D180 or even D90.
In turn, ad-monetized games will be favored for their short payback periods and lower CPIs. Although AdROAS campaigns may be cheaper, they still attract high-quality players that consume ads en masse.
Facebook will die out for the better. I’ve added Applovin and Ironsource to my UA mix already, and we’ll continue to see new players as privacy changes continue to fade Zuckerberg out. Did somebody say fingerprinting? Oh well, no no no … right, nobody!
So what?
And the hope is that this all leads to…well…better ads. There is a legitimate fear that everything might start to look and feel the same (since when did mobile marketers care?), but that presents a fresh opportunity for those actually skilled at this thing to stand out even more.
Skilled UA managers – who know how to strategize and operate AI tools will be successful in this field. And if there’s one thing I know about good UA managers, they’re resilient as shit.
