OpenAI's marketing masterclass (maybe)?! 🤔

Hello marketers. Welcome to AI Marketing School, where we dish out the latest and greatest in AI-powered marketing. In this week’s issue:

1. AI Marketing Update: Inside OpenAI’s mad marketing maneuvers

2. The Stack: ElevenLabs drops Audio Native

3. Consultant’s Corner: Small batch programmatic SEO

Onwards!

AI MARKETING UPDATE

OpenAI masters the hype cycle

OpenAI's launch of GPT-4o has once again showcased the company's mastery of the hype cycle. 

Here are four key marketing observations from another week of crazy events in AI land:

1. Time major launch with your competitor’s conference, and then pretend it was a coincidence

OpenAI almost certainly timed GPT-4o’s release to coincide with Google's flagship I/O event.

Some even suggested that the “o” for “omni” in GPT-4o felt like a cheeky nod to Google.

Altman drew attention to the timing when he posted oh so innocently, "I try not to think about competitors too much, but I cannot stop thinking about the aesthetic difference between OpenAI and Google."

2. Hype up your products as ‘magical’ even though they’re minor upgrades

When the big reveal finally came, GPT-4o's most hyped feature – its ability to process audio input and output in real-time – was absent, still stuck in pre-alpha development. 

Then, early evaluations of GPT-4o's performance show that despite being fast and cheap, it struggles in some tasks.

AI researcher Bindu Reddy' showed that GPT-4o succeeded in only 79/96 ‘hard tasks’ compared to GPT-4's 90/96, which raised eyebrows.

One frustrated developer said on the OpenAI forum, "GPT-4o is very bad compared to GPT-4 and even GPT-4-turbo for our uses, but we switched to GPT-4o anyway because of the price and have our scripts filter out the terrible outputs we receive sometimes…" 

This is OpenAI’s new ‘flagship’ model, remember?

3. Get in the papers by flying close to the wind with a Hollywood actress

Scarlett Johansson personally attacked OpenAI and Altman for impersonating her voice from the film “Her,” in which Joaquin Phoenix falls in love with a flirty AI that sounded near-identical to the voice OpenAI demoed for GPT-4o. 

Altman is oddly fascinated by “Her,” having mentioned it multiple times in interviews.

He then dropped the word “her” in the hours leading up to GPT-4o’s launch, incurring the wrath of Johannson herself for red-letter media coverage.

OpenAI soon removed the Johannson-like voice in ‘good faith.’

4. Convert users with a freemium masterstroke

GPT-4o is almost definitely a ‘freemium’ stepping stone to bigger things coming.

It will collect tons of free users (and their data) and then aim to convert them to paid subs once GPT-5 and/or Sora drops.

Also, OpenAI was shrewd enough to lock GPT-4o on mobile while making it free for desktop users.

Desktop users got a taste of the new product, while mobile users were incentivized to go pro – even though the much-hyped voice mode hasn’t rolled out yet.

It's a clever game that leverages GPT-4o's hype to drive adoption and revenue in the short term while setting the stage for a more transformative offering down the line. 

Whether GPT-5 will live up to these lofty expectations remains to be seen, but one thing is clear: OpenAI knows how to keep us all on the edge of our seats.

THE STACK

ElevenLabs drops audio native: AI voices that don't suck

If you've ever dabbled with text-to-speech, you know most AI voices sound like they came straight out of a '90s sci-fi flick.

ElevenLabs, the leader in text-to-speech tech, is changing that with Audio Native.

Audio Native is an embeddable player that automatically generates human-like voiceovers for your blog posts, articles, or newsletters. Crucially, these voices are fluent and engaging, with plenty of emotion and nuance.

I’m not sure if you can use clones of your own voice for Audio Native yet, but that would be a great addition.

Setting up Audio Native is very easy.

Log in to your ElevenLabs account, whitelist your domain, pick a voice, customize the player, and embed the code. Boom, done.

But why should you care about giving your content a voice?

Two words: accessibility and engagement.

By offering narration, you're opening up your content to a wider audience.

People who prefer listening, folks with visual impairments, and multitaskers who want to consume your brilliant ideas while doing the dishes.

You could also increase engagement and dwell time, so there are plenty of SEO benefits to be had.

CONSULTANT’S CORNER

As Google's March Core Update takes aim at "scaled content abuse," many in the SEO community are wondering if programmatic SEO (pSEO) is in the crosshairs.

After all, pSEO is all about generating content at scale using templates and databases.

So, is this the end of the road for this technique?

Not necessarily. While Google is indeed cracking down on low-quality, automated content, there's still a place for programmatic SEO – if it's done right.

The key is to focus on "small batch" pSEO: generating content at a smaller scale, but strongly emphasizing quality and humanization.

The process starts with keyword research. But instead of targeting broad, high-volume keywords, the focus is on identifying niche, long-tail keyword pools that can be served by a single high-quality template.

These pools might have a lower combined search volume, but they offer less competition and more opportunity to provide genuine value.

For example, let's say you're in the fitness niche. A broad keyword like "weight loss" is highly competitive and would require a wide range of content to serve effectively.

But a more niche pool like "weight loss tips for new moms" could be served by a single in-depth, high-quality template.

To find these keyword pools, you can use tools like Ahrefs or SEMrush. Look for clusters of keywords that share a common theme or intent.

Pay attention to the search volume and difficulty scores to find pools that offer a good balance of opportunity and competition.

Creating your pSEO template 🗎

Once you've identified your keyword pools, it's time to craft your content template. This is where you define the structure and elements of your pages.

A strong template should include:

• An engaging, keyword-rich title

• A compelling introduction

• Subheadings for key sections

• Spots for unique insights and data points

• Placeholders for quotes from experts or users

• A conclusion with a strong call-to-action

• Optimized meta tags and headers

Your template should be designed to provide genuine value to the user. It should answer their questions, provide actionable insights, and engage them with a mix of information and personality.

With your template created, you can start filling in your database. This is where you compile all the data points, insights, quotes, and other variable elements that will be used to generate your individual pages.

For small batch pSEO, it's crucial that this database includes real, high-quality information.

Tools like Bing in ChatGPT can help you find up-to-date, relevant statistics and quotes. The goal is to humanize each page with authentic data and insights.

Use small batch pSEO to take on a more editorial role for your content, accelerating the production of routine pages that are valuable but, quite frankly, boring to write from scratch.

Humanizing them and juicing them up with topical info, stats, etc, is a much more palatable endeavor (and this can be outsourced).

After all that, it's time to generate your pages.

If you're using WordPress, plugins like WP All Import or Multi Page Generator can automate this process. These tools allow you to map fields from your database to template elements and then dynamically generate pages or posts.

Find a great tutorial for that here.

Hope you enjoyed this week’s issue. If you missed it last week, you can read it here.

Until next time. Happy marketing.

—The AI Marketer