Artificial intelligence

Text-to-creation: the universal form of creation in the past, present & future

What is text-to-creation? Text-to-creation refers to the category of technologies that takes text as an input, and produces multimedia that span text, audio, images, or video as an output.
– Sandy Diao

Text is the universal form of creation

This statement may not be immediately obvious, but it is perhaps the most logical explanation for the widespread consumer adoption of LLM (large language model)-powered experiences that we are seeing today. As of writing, OpenAI has publicly disclosed that their ChatGPT technology has well over 100 million monthly active users. One third of Gen Z, millennials, and Gen X have tried generative AI technologies (Statistica 2023). The market for AI has grown to $6.6 billion this year, and is expected to triple to $21 billion by the end of the decade (Statistica 2023). It’s important to note that we are only scratching the surface with these numbers when measuring the impact of AI-only tools. When we consider the possibilities of using AI to supercharge industries such as pharmaceuticals ($1.4 trillion) or entertainment ($717 billion), the numbers become truly incomprehensible. So now, the question we all naturally want to understand is what happens from here and what opportunities lie ahead? To answer this question, we must look at how we got here to understand how our behaviors will evolve next.

Text is the primary medium of daily communication

To provide context on how I’ve developed my point of view, I’ve worked behind the scenes to launch several AI products. During my time at Indiegogo, I helped hundreds of founders launch hardware products that incorporated early versions of edge computing on consumer devices, such as security cameras and emotional companion robots. I’ve also worked at Pinterest and Meta, whose advertising platforms allowed end-users to access AI ad delivery through machine-learned audience models for conversion optimization. Currently, I work on growth at Descript, where we’ve built a text-based video editing software powered by generative AI models across transcription, voice synthesis, audio generation, and more.

At first glance, Descript may seem like a surface-level UX change to a video editor. However, when you try to understand why it works, you’ll start to realize that it’s not just a coincidence that Descript is as easily learned and accepted by content creators and businesses as it is. Beyond Descript, the common theme across all types of creation is that we create with text to communicate. Most communication happens asynchronously, and the way we do that is via text – handwritten, emailed, scripted, documented, in blog posts, books, engravings, and the list goes on and on.

When you layer on what’s happening in the landscape of adoption of ChatGPT and consider the incredible versatility and historical importance of the written language, it becomes clear that text is truly the foundation upon which our modern forms of communication are built. When we consider the incredible versatility and historical importance of written language, and layer on what’s happening in the landscape of ChatGPT adoption, it becomes clear that text is truly the foundation upon which our modern world is built. Text-based prompts through a chat interface are the initial form factor that allows us to talk to computers in a way that we’ve never been able to before.

Text can be used to create more than just text

In addition to text-based prompts that generate text-based outcomes, there are many other forms of text-to-creation that are becoming prevalent today. For instance, text-to-speech tools based on transcription models are being utilized in everything from virtual assistants and smart home devices to audiobooks and podcasts.

Text-to-image technology is also on the rise and has the potential to revolutionize everything from advertising and marketing to film and television. This technology can produce realistic, high-quality images from text descriptions, eliminating the need for expensive photo shoots and video productions, making it easier and more affordable for creators to bring their visions to life.

Lastly, text-to-expanded text creation is another area of development that could become even more powerful. As artificial intelligence and machine learning continue to improve, we may soon be able to create prompts with minimal context that are fine-tuned to create results indistinguishable from the work of humans. Imagine being able to write a book in precisely your own style and intent by providing just a log line. This could have far-reaching implications for industries such as journalism, advertising, and content creation, fundamentally altering the way we produce and consume written content.

The emergence of text-to-creation was inevitable

Considering text as the medium of creation throughout history makes a lot of sense. When thinking about the earliest forms of writing, one might envision cave paintings and hieroglyphics from history classes. Fast forward to the computing age, where humans use text to communicate, record, and create through typing. The written word remains the foundation of stories, systems of belief, and education, spanning from religious texts and philosophical treatises to novels and scientific papers.

In recent times, new forms of text-based creativity have emerged. The novel, which rose to prominence in the 18th and 19th centuries, marked a significant departure from traditional forms of storytelling and introduced new techniques for character development and plot construction. In the 20th century, the internet and digital technologies opened up new avenues for text creation, including blogs, social media, e-books, and online journalism.

Today, we continue to tell stories and share information primarily through text, even as we create short and long-form videos, explainer videos on YouTube, and self-created and hosted curricula across over 600 million blogs as of this year (Web Tribunal 2023). In many cases, even videos and movies originate as scripts. And while not all videos require manual text tagging, it’s worth considering the effort required to classify the topics and information in those that do.

The ideas that are expressed in written mediums originate in the minds of writers and creators before they are put onto paper. In order to effectively share and communicate these ideas with others, the text needs to be well-crafted and nearly finalized. Recently, LLMs have given us the ability to communicate sequentially and iteratively, much like natural language conversations. However, just as there are good and bad conversations based on the context provided and the clarity of instructions, there are also more and less effective ways to prompt an LLM. The effectiveness of text-based communication depends on the communicator’s abilities. In this new world of accessible LLMs, it is important to learn how to turn our thoughts into clear communication using natural language.

Now, tools like Descript enable users to create a custom voice model with just a few minutes of training data. In the future, some of these models could be instant and hidden within app experiences. Completing a computer task could be as easy as speaking to it in natural language and asking it to perform a task.

The emerging role of natural language speech

An important development in this area will be enabling speech, which is the most common form of natural language delivery today, to be converted into text so that we can use it for iterative creation. Significant progress has already been made in converting human speech to text through speech-to-text transcription models, such as Whisper, Google Cloud’s Speech-to-text, and Amazon Transcribe. Let’s take a look at the most important moments in the evolution of digitizing natural language speech by decade:

1950s: One of the earliest examples of text-to-speech technology can be traced back to Bell Labs, who introduced the Audrey system. While the technology was still in its infancy, it laid the groundwork for the development of modern text-to-speech systems.
1960s: Physicist John Larry Kelly, Jr used an IBM 704 computer to synthesize speech.
1970s: Itakura developed the line spectral pairs method. MUSA, a stand-alone speech synthesis algorithm used to read Italian out loud, was also released during this year.
1980s: Sun Electronics released Stratovox, a shooting-style arcade game that shook the video game world. DECtalk, the standalone acoustic-mechanical speech machine, was built, and Steve Jobs created NeXT, a system that was developed by Trillium Sound Research.
1990s: Ann Syrdal at AT&T Bell Laboratories developed a female speech synthesizer voice. Engineers worked to make voices more natural-sounding.
2000s: Quality and standards for synthesized speech became a working issue for developers. This coalesced in researchers using a common speech dataset along with deeper research into additional vectors of creation, such as emotions.
2010s: OpenAI released GPT-1, the first iteration of their generative language model. Researchers at the University of North Carolina at Chapel Hill introduced the AttnGAN model, which uses attention mechanisms to generate images from text descriptions. Since then, there have been numerous advancements in this field, with companies like OpenAI and NVIDIA developing their own image-generating models.
2020s: Open AI has continued to refine and develop the technology, culminating in the release of GPT-3 in 2020. With the ability to generate human-like text, GPT-3 has opened up new possibilities for text-to-text creation. In 2022, Open AI released ChatGPT as a public beta, and then released an updated model in 2023 that takes image inputs.

Voice-to-creation is also text-based creation

Before we draw the firm conclusion that text underlies universal creation, we need to consider how voice-based creation fits into the picture. In the past few years, we have used our voices to activate AI assistants on devices powered by Alexa and Siri. In a 2018 interview with Vox, tech entrepreneur and investor Marc Andreessen predicted that “voice is going to be the primary user interface for the majority of computing over the next 10 years.” This prediction may still come true, but for that to happen, we need to understand that the engine powering voice interfaces is still based on text-to-speech models.

Voice-enabled technologies such as virtual assistants, like Amazon’s Alexa or Apple’s Siri, rely on text-to-speech models to convert spoken language into written language. This written language is then processed by natural language processing algorithms that help the technology understand and respond to the user’s request. Even the most advanced virtual assistants on the market, including Google Assistant and Amazon’s Alexa, are powered by text-to-speech models that are trained on vast amounts of written text. This is also true for other forms of media creation – the intermediate step is almost always text-to-creation at this time.

What happens next?

So, how do we evolve with text-to-creation from here? There are two key areas that will be innovated in in order to supercharge our ability to create with text:

More context as inputs, such as voice, vision, and motion. Relying solely on text can limit a model’s ability to infer all necessary context. While many models already support image inputs, we also need to incorporate video and other forms of real-world capture to enhance the results of our text-based prompts, and to extend into more creative fields that require understanding the nuances of emotion, intent, and preferences.
Edge computing will allow models to tap into individually unique and private data inputs that fine tune models to generate powerrful outputs that we may have never imagined.

While we can’t predict exactly how text creation will evolve in the future, it is clear that new technologies will inevitably be created to capture more context and compute with more personalized sources of data. There’s a universe of outputs that power use cases that we probably can’t even imagine yet, and I’m excited to see what those are.

Lensa app review: Are Lensa AI avatars worth the cost in 2023?

What is the Lensa app?

Lensa AI is a magic avatar generator app for iOS and Android that lets you create colorful, custom AI portraits of yourself. It takes less than 30 minutes to get a collection of 50+ avatars that are customized based on selfie images that you upload to the app.

How much does the Lensa app cost?

First, my Lensa app review

3 out of 5 stars. The avatars were not bad, but not great. 96% of the images were mediocre at best. However, I did like 2 out of the 50 images that I paid $3.99 for. Considering that it’s a 4K resolution piece of digital art hat I can now use as an avatar, I suppose it could have been worse. That’s about $2 for each digital photo, so I can’t say that I was totally scammed.

There are other cartoon-ify avatar generators that I’ve had better experiences with, such as Wonder. If you’re not interested in spending money on these yet, I recommend taking a look free apps with limited styles, like Snow or Beautycam.

Getting the avatars set up

In all, it took about 20 minutes to get from app installation to a folder of 50 avatars. It was about 5 minutes to set up, configure styles, and select selfies as training inputs. From there, it took approximately 15 minutes to generate the avatars.

22 styles across three categories of realism.

Here’s a step-by-step look at how I created my avatars with the Lensa app:

Sign up for a free 7-day trial to access the Magic Avatars feature.
Choose some styles. You can pick from 22 styles across the categories Essential, Art, and Time Machine.
Upload 10-20 selfies. Online guides suggest avoiding any face coverings, including hands that might be touching the face, since it impacts the training.
Choose to pay $3.99 for 50 generated images (10 styles) or $6.99 for 100 generated images. 50 Sounded like a lot, so I went with 50 (though afterwards seeing how wildly in accurate and gross some of the photos were, I wish I had more options to choose from).
Hit the “generate” button, and it takes between 15-20 minutes for the images to finish.
You’ll see your pack generated, which you can download all at once in 4k or hi-res files. You can also download files individually.

Best and worst styles, according to my results

Before I share the rank ordering, I wanted to share that the biggest problem I observed in the worst style packs was that I think Lensa is probably missing training data for Asian faces. I found the more creative styles, such as Iridescent, Pop, and Comics to be grossly incorrect.

Stylish pack – looks very generic. Don’t know how this was trained using my selfies.

Comics style – doesn’t look like me. feature proportions are off, like the eye shapes.

Anime style – doesn’t look like me. feature proportions are off, especially the eyes.

Most accurate style: Cosmic

This was a bit surprising, since I assumed “Cosmic” would be a bit more abstract. But these avatars looked more like me than any of the other styles.

Most aesthetic: Iridescent

While the first two images don’t look like me at all, they’re still high quality, ethereal styles.

Most creative: Pop

Aside from the bottom right square that is just a cleaned up version of the selfie I submitted, these were unique, art-like renditions.

Most generically bad: Anime

These are so generic that you’d think I just randomly found these pictures while searching online. There are no indications that they were generated from my photos.

Most inaccurate: Focus

None of these look like me. Maybe slightly if I squint, but I look at it and feel terrified at how bizarre these are.

Don’t forget to check the privacy policy and the terms of use

Before you rush to download the app and start generating avatars, keep in mind that as consumers, we’re entering a whole new era of privacy and data rights management around the images that we entrust with generative AI companies. If this is something that you care about, make sure you understand the details in their privacy policy and terms of use to understand what the app does with your selfies, data, and generated images.

Lensa app review: Are Lensa AI avatars worth the cost in 2023?

What is the Lensa app?

First, my Lensa app review

3 out of 5 stars. The avatars weren’t bad, but also not great. 96% of the images were mediocre at best. However, I did like 2 out of the 50 images that I paid $3.99 for. Considering that it’s a 4K resolution piece of digital art hat I can now use as an avatar, I suppose it could have been worse. That’s about $2 for each digital photo, so I can’t say that I was totally scammed.

Getting the avatars set up

Here’s a step-by-step look at how I created my avatars with the Lensa app:

Sign up for a free 7-day trial to access the Magic Avatars feature.
Choose some styles. You can pick from 22 styles across the categories Essential, Art, and Time Machine.
Upload 10-20 selfies. Online guides suggest avoiding any face coverings, including hands that might be touching the face, since it impacts the training.
Choose to pay $3.99 for 50 generated images (10 styles) or $6.99 for 100 generated images. 50 sounded like a lot, so I went with 50 (though afterwards seeing how wildly in accurate and gross some of the photos were, I wish I had more options to choose from).
Hit the “generate” button, and it takes between 15-20 minutes for the images to finish.
You’ll see your pack generated, which you can download all at once in 4k or hi-res files. You can also download files individually.

Best and worst styles, according to my results

Most accurate style: Cosmic

Most aesthetic: Iridescent

Most creative: Pop

Most generically bad: Anime

Most inaccurate: Focus

Don’t forget to check the privacy policy and the terms of use

How to use AI writing tools to your advantage

Artificial intelligence (AI) has revolutionized the way we live and work, and writing is no exception. With the help of AI-powered writing tools, writers can generate ideas, improve their writing skills, and automate time-consuming writing tasks. Even prior to the advent of ChatGPT and other conversational AI bots that now act as personal writing assistants, several companies have for months and even year, developed some of the leading AI writing tools available today such as Writesonic, Jasper (formerly known as Jarvis.ai), Grammarly, ProWritingAid, Copysmith, Rytr, Articoolo, and more. These tools use machine learning algorithms to analyze large amounts of data, learn from it, and generate human-like content as the output. In this article, we’ll explore how you can use AI writing tools to your advantage, and which tools are the best fit for practical parts of the writing process.

What is AI Writing?

AI writing is a technology that uses natural language processing (NLP) algorithms and machine learning (ML) models to generate content without human intervention. AI writing software are usually designed to help with specific writing tasks, such as generating blog articles, email newsletters, social media posts, and even creative writing.

AI writing tools have proven to be a game-changer for many creators and businesses. For instance, the team at HubSpot has used AI writing tools like Writesonic to create dozens of social media posts in just a few minutes. Content creators have also creatively applied AI writing tools – in 2020, a group of digital marketers used GPT-3 to generate several blog posts that received thousands of views and shares. The tool’s ability to generate content quickly and efficiently helped them reach a broader audience, resulting in increased traffic and engagement. Even traditional authors have used AI writing tools to generate outlines and ideas for their books. A popular tool for this is Articoolo, which allows writers to input a topic or keyword and generates a full-length article on that topic in just a few seconds.

Can an AI Write My Essay?

Yes, in theory and practice, AI can write your entire essay (and I don’t even want to think about what teachers and professors are doing now to grade students’ work with this as a much more accessible technology). Still, while AI writing tools can be useful in generating content, they’re not yet advanced enough to write an entire essay without any human input. Students and other essay writers will still need to guide the writing as much as possible to ensure that it addresses the objectives of the essay.

Where essay writers can use AI tools most effectively today will be in assisting with idea generation, improving writing, and providing suggestions for revisions. For example, tools like Grammarly may provide suggestions for essay structure or offer ways to improve the essay’s clarity, style, or grammar. Another creative application could also be to use AI to search for research sources to support an essay’s arguments.

Can an AI Write a Script?

AI writing tools can assist in scriptwriting by providing suggestions for character dialogue, plot development, and even story arcs. While an AI writing tool probably can’t write an entire Oscar award-winning script on its own, they can be useful in providing ideas and improving the writing process.

The entertainment industry has already started using AI writing tools to generate scripts for movies, TV shows, and video games. In 2016, a short film titled Sunspring was created using an AI writing tool called Benjamin. Benjamin, which was developed by filmmaker Ross Goodwin, used a recurrent neural network to analyze hundreds of science fiction screenplays and generate a script for the film. Even media giants like Netflix has been using AI writing tools like LDA (Latent Dirichlet Allocation) to analyze viewer data and generate script ideas for its original shows. The streaming giant also used an AI writing tool called AIVA to generate a musical score for one of its shows.

How Do I Create Content for My Blog Using AI?

Creating content for a blog can be a time-consuming process, especially if you’re looking to produce high-quality, engaging content consistently. Fortunately, AI writing tools have made it easier to create blog content quickly and efficiently.

One popular AI writing tool for blogs is Jasper. Jasper offers a variety of features, including content ideation, writing, and editing. For example, if you’re struggling to come up with a blog post idea, you can use their content ideation feature to generate multiple ideas based on a keyword or topic. The tool uses machine learning to analyze thousands of blog posts and suggest ideas that will perform well.

AI writing tools can also help with writing SEO-optimized blog articles. AI tools can help improve the quality of your blog content by suggesting relevant keywords, optimizing meta descriptions and titles, and analyzing the content’s readability. SurferSEO is an AI-powered SEO writing tool that uses machine learning algorithms to analyze search engine results and suggest relevant keywords and phrases for your blog posts. It can also provide feedback on the length of the content, its structure, and readability. Clearscope is another tool that uses AI to analyze your content, and the app will suggest keywords and topics that can help improve its relevance and ranking on search engines. It’s like a real-time grader for SEO.

Practical Ways to Start Using AI Writing Tools

Now that we’ve covered the basics of AI writing tools, let’s discuss how you can use them to your advantage. Here are some tips for using AI writing tools effectively:

Use AI writing tools like Writesonic or Jasper to generate ideas quickly. These AI-powered tools can analyze your keyword or topic and provide you with content ideas that you can build upon.
Refine your writing style and grammar using Grammarly or ProWritingAid. These tools use AI to scan your writing for errors and provide feedback to help you improve your writing.
Automate your writing process using tools like Copysmith or Rytr. These tools can generate content for you quickly, freeing up time to focus on other tasks. Copysmith can create high-quality product descriptions, blog intros and outros, and more, while Rytr can help you create blog posts, social media posts, or even ad copy.
Experiment with new writing styles and genres using tools like Articoolo or Writesonic. Articoolo can create high-quality articles based on your keywords, while Writesonic can generate content for social media, ads, blogs, and more.

While AI writing tools like the ones mentioned above are not capable of writing an entire book or essay without any human input, they can certainly assist writers in generating ideas and creating content quickly and efficiently. By using these tools to your advantage, you can improve your writing, save time, and even have fun exploring new creative ideas.

The dark side of AI-powered marketing

Artificial intelligence (AI) has become a major tool in the growth marketing arsenal, helping businesses drive growth and increase revenue. In fact, as the world has been introduced to consumer versions of AI through tools like ChatGPT recently, businesses have been using AI-powered campaigns and marketing channels for a long time. From TikTok video feeds to Facebook ad delivery, machine-learned targeting models have already seeped into our daily interactions and experiences on our phones and devices. Even Siri and Alexa (as primitive as they seem) voice assistants are deep learning TTS (text-to-speech) model rooted in AI technology. Businesses use AI-powered marketing rooted in machine learning algorithms to analyze user behavior and personalize marketing campaigns in real-time. This allows businesses to reach consumers more effectively and efficiently than ever before. However, as AI becomes increasingly prevalent in marketing, there are emerging pockets of ethical risks that potentially pose harm to you and me.

One of the leading concerns about AI-powered marketing is a tale that we’re all familiar with: the spreading of misinformation and fake news through AI-powered social media campaigns. AI algorithms can be easily manipulated to create fake news and spread false information to millions of people. According to a study by MIT Technology Review, false information spreads faster and more easily on social media than true information, which can have serious consequences for public opinion, trust, and democracy (MIT Technology Review, 2018). For example, the 2016 US Presidential election was influenced by AI-powered media campaigns on social media, which helped shape public opinion and ultimately influenced the outcome of the election (The Guardian, 2018).

Another area of concern is the manipulation of consumer behavior through AI-powered personalization. AI algorithms can track and analyze consumer behavior, including their browsing history, search history, and social media activity, to personalize marketing campaigns. While this can be a powerful tool for businesses to reach consumers more effectively, it also has the potential to exploit vulnerabilities and sway consumer behavior in potentially unintended ways. According to a report by the World Economic Forum, as AI algorithms become more advanced, they will have the potential to manipulate human behavior to a much greater extent, creating new risks for consumer welfare (World Economic Forum, 2021). For example, brands or organizations can use AI algorithms can be used to target vulnerable populations and exploit data to drive adoption or education, leading to negative consequences.

Big data could become another driving factor behind unethical practices in AI-powered growth marketing. AI algorithms are fed massive amounts of data to analyze consumer behavior and personalize marketing campaigns. According to a study by McKinsey & Company, big data analytics has the potential to generate significant value for businesses, but it also presents significant risks to privacy and security (McKinsey & Company, 2016). Vast quantities of data collected by algorithms can be easily manipulated to create false patterns and trends, leading to skewed results and unethical practices. Having worked in technology companies with access to large data sets, I know that there are several levels upon which data can be biased, both in how we determine which signals to track, how we define them, and ultimately how we interpret what the data is telling us. The scary thing is that we don’t fully realize that we’re making biased decisions, because we think the data is objective.

One could even see a world in which perfectly fine-tuned AI, which is based on mathematical models and algorithms that are designed to analyze and interpret patterns in data, can lead to a worse or more mediocre customer experience overall. These algorithms struggle to understand the complexity of human behavior and the context in which these behaviors occur. According to a report by Harvard Business Review, AI algorithms are still limited in their ability to understand human behavior and the context in which it occurs, which can lead to homogenous and generic experiences for consumers (Harvest Business Review, 2020). As a result, AI-powered growth marketing can create homogenous and generic experiences for consumers, ignoring important cultural and contextual factors that can influence consumer behavior. Imagine a world where every experience on the Internet is a 3.5-star Yelp restaurant one – homogenous, not too good nor too bad. Optimized right for the crowd.

To address these ethical concerns and limitations in AI-powered marketing, I believe that businesses must take a responsible and ethical approach. This might come in the from of more businesses putting ethical guidelines front and center, in addition to their terms of services. We need a cultural system (or even formal regulation) that can ensure the right incentives for transparency and accountability in AI-driven marketing campaigns, protect consumer privacy and data security, and create meaningful and respectful experiences for consumers. Our best bet will be to find the sweet spot in combining AI with human intuition and creativity, and then putting together the charter upfront for how these interact.

What the future of performance marketing holds

There is a lot of buzz around AI-driven marketing today. Tools like Jasper and ChatGPT have caused many go-to-market professionals to question whether their jobs will be replaced by robots in the near future. It can be unsettling to think about, but I’m also excited about the potential of these technologies to drastically improve our work and productivity. In my daily work, I’ve been able to use Jasper to quickly generate strategy statements and ad taglines with minimal input. It’s even helped me proofread parts of this write-up.

While this is all novel and exciting, AI-driven marketing is already making an impact on performance marketing channels. In fact, I’ve been using an ML-based performance ad channel over the last year in my role running growth at Descript, and these campaigns have broken our performance ceilings month after month. Let’s explore what’s happening now and how it affects our digital marketing and go-to-market decisions.

ML-based performance marketing is already here

Channels are all converging toward ML-based conversion optimization for targeting. Example: Google’s Performance Max campaigns; Facebook/Instagram’s broad targeting with conversion optimization campaigns
- For those who aren’t familiar: Google Performance Max is a paid search campaign type that uses machine learning to optimize ad delivery and targeting in real-time, which results in achieving better performance and ROI outcomes. Instead of setting up targeting, you set up ‘”anti-targeting’” by telling the system what’s not considered a valid conversion.
ML-based targeting is ideal from the lens of ROI – the machine-learned system will most certainly be more precise in targeting and use each ad dollar in more effective way than human allocation could. But, it’s not ideal from the lens of gaining customer insights.
Perhaps in 6-12 months time, all paid search and paid social will utilize blackbox targeting where we don’t define or know who the platforms are targeting, except understanding broad filters geography, demographics, or platform, to name a few examples.
This creates a reliance on ads, where you can get a ton of results but you don’t know what drove its success and you can only count on the blackbox targeting system to continue working well.
- A good parallel might be to understand what Amazon does with brands/sellers today. You pay Amazon to sell on your behalf, and the units sell or they don’t, but you don’t know who the customer is at the end of the day.
- Overall, the Amazon-esque system can work incredibly well, but it rules out a certain type of seller that isn’t savvy enough to learn and exploit the system.

Skills for the future are technical and strategic

Performance marketing becomes a technical sport

Less focus on hiring those with ads operating experience, which tends to be the background of most junior performance marketers or agency hires. The things that performance marketers spend time on today, such as building the post, creating targeting parameters like frequency caps and daily budgets, setting up an a/b experiment, or optimizing based on results, will all become automated. Instead, people will spend more time defining clear optimization signals and structuring campaigns across these signals. Think of it almost as prompt creation.
Performance marketing as a whole will become a technically performant function where the operators behind the scenes will need to be data-minded in order to understand how utilize targeting and ranking mechanisms.
Instrumenting marketing data systems isn’t a common skillset for data engineers. That’s why the performance marketer will need to become an acting product manager to help ensure accurate instrumentation, and define a data taxonomy that becomes useful for marketing teams.

Content marketing and product marketing work requires more precise customer segmentation

It will be imperative to develop strengths across multiple niches in order to gain scale. Otherwise, we’ll hit the ceiling within the constraints of our economics – i.e. willingness to pay for a conversion – and time horizon.
More micro-targeted content positioned to the customer that will be likely to convert. This could mean developing more landing pages with focused content that gives greater chance of fitting into the blackbox targeting match.
Having a multi-faceted product that spans customer segments should play to our advantage in that it allows us to compete in multiple audience segments at one time, which increases our chances of exposure and conversion.
This increases our need to have a clear data-driven signal customer segments that can be exposed externally somehow.

The chasm between digital performance and brand marketing tactics will widen, yet the function of these are intertwined

Today, performance marketers look toward increasing reach as a mechanism to improve the likelihood of finding a converting audience. Tactics include bidding on a CPM basis to achieve a broader set of conversion outcomes. However, in the world of ML targeting, the system will reach as few users as possible to reach conversion objectives, which minimizes total reach and frequency of reach.
Brand and performance assets are thought of as separate entities right now, but these are becoming inseparable parts of the customer’s experience. For example, a user that is abruptly shown a use case landing page will most certainly want to explore the brand’s larger offering on the homepage or other branded pages (in my experience, this happens 20-30% of the time). Both types of pages will be relevant to cold prospects who are looking to learn about the brand while trying to get more information about their use cases. It’s about identifying the ideal navigation path and understanding what happens after a user interacts with a brand for the first time, or vice versa.
The bridge from brand marketing to performance marketing can be built via a multi-touch attribution model that companies either develop themselves by centralizing all data within a CDP like Segment, or they can be outsourced and imported to a 3rd party tool that does this for you, like Branch.

Preparing for the future means hiring for technical skills & building for channel-market fit

In a world of ML-driven customer acquisition, digital marketing teams must create clear signals and high-quality user experiences to be successful. This means that the better you perform today, the better you’ll perform tomorrow – it’s a performance flywheel. This wasn’t the case in the past, when performance marketing was limited by target audience ceilings and manual optimization decisions. The more signals you have, the better these platforms will be able to acquire new customers for you.

To accomplish this, you should hire ad managers from diverse backgrounds, with an emphasis on those with technical expertise. Content quality is still paramount for delivering persuasive messages that convert, but product marketing, content marketing, growth marketing, and product teams must work together to create a well-managed journey from first touch to product entry.