Howdy hey ,

I wrote this newsletter before Google announced its AI strategy at I/O 2024 on Tuesday, specifically that AI-generated summaries would be the default and live at the top of search pages. I need more time to dig into the implications of this change, including top-of-funnel content and distribution.

Regardless, Google's sudden shift toward AI search shows the urgent need to address what we'll cover in this issue. I welcome your feedback and ideas!

In 2006, UX research firm Nielsen Norman Group (NN/g) announced its discovery of the “F-shaped pattern” which digital marketers will recognize. The F-shaped pattern is how most web users review online content:

• They read the first few lines.
• They read some of the next few lines.
• They scroll and skim the rest.

Thus, an “F” shape.

While NN/g’s research uncovered other patterns, managing the F-shaped pattern quickly became de rigueur for optimizing web experiences. The firm’s conclusions reflected the reality of engagement online:

“The vast majority of the web users would rather finish their tasks as fast as possible with the minimum amount of effort; they visit a page because they want to find a quick answer rather than read a dissertation on the topic and educate themselves.”

They say speed kills, and it appears end users agree. This is especially true in B2B marketing, where people average 37 seconds “reading” (aka skimming) online content.

I bet that number shrinks as genAI products spread throughout online interactions. For instance, LinkedIn has offered AI-generated summaries of posts to its Premium users since November 2023. Google’s “Search Generative Experience” (SGE) compiles summaries from top-ranking articles to answer queries and marks early moves in Google’s AI pivot. ChatGPT can summarize webpages in seconds — something Apple users will probably see more of soon.

This AI use case aligns closely with NN/g’s assertion of web users maximizing return on minimum effort. Plus, it just makes sense: If you’re tasked with researching something at work, you’re eager to get ‘er done ASAP. Why not have AI compile and analyze bunches of information? Even skimming seems like too much work.

In this potential timeline, where everyone asks AI for the answer, why does the rest of the content matter? Specifically, what do we do about the more human-facing elements like your brand's voice?

Much Ado About Brand Voice

Let’s define “brand voice” as how you present and promote your experiences and expertise to your audiences. As a primarily long-form writer, I consider voice as the narrator guiding your reader through your story (like a good novel). A strong narrator hooks you into the plot, endears you to the characters, and leaves a memorable impression (crucial in making compelling content).

How will AI affect this vital part of storytelling? I’ve struggled to find good discourse on this question. So far, I’ve compiled a long list of genAI products capable of infusing their output with your company’s brand voice. And at marketing events, I’ve heard personal brands will matter more as buyers get wary of company brands. Users will want to engage with people, so content marketing should focus on staging compelling personal brands.

But that doesn’t answer my question. If my job is to find answers, I probably don’t care if Salesforce, the brand, or Marc Benioff, the personal brand, delivers what I need.

Then, I can have ChatGPT summarize Salesforce’s blog or LinkedIn Premium turn Benioff’s long post into five bullet points. Who cares how said information is presented if I get my answer?

I’ll set aside the larger dialogue about zero-click search results and how SGE could widen the moat the biggest ad spenders enjoy. But in short, distributing and positioning content in front of users is just damn hard these days, and — at least for longer-form content — I don’t see brand voice majorly affecting that.

Instead, voice’s role is to push audiences past informative AI summaries and toward persuasive, change-your-mind content — the kind that breaks the F-shape.

A Return to Rhetorical Persuasion

GenAI’s domain is managing informational content: It can locate, analyze, and summarize product features and benefits, competitive sets, and the how-to or “What is XYZ?” content better than people can.

It’s a start, but buyers will still ask: Why should I buy from you instead of the other guy?

You need more to effectively persuade them.

The shift from informational to persuasive writing reflects the increased opportunity and transaction costs your user incurs. At the informational stage, a quick answer is low-stakes; if they make a mistake, it’s usually reversible. At the persuasive stage, however, business users are making larger-stakes decisions: 

• Advocating for a strategic direction
• Managing business, team, and process changes
• Recommending solutions for purchase

Mistakes here might lead to larger, more painful consequences, so end users want more details about you, your business and offers, and ways to validate your expertise.

How do you convince your audience? Persuasive content includes the three Aristotlean artistic proofs: ethos, logos, and pathos. The mixture of credibility, logic, and emotion shapes the outcomes of persuasive writing.

Brand voice (company and personal) reflects our chosen rhetorical methods for presenting information and speaking with our audiences. That voice can engender or erode trust. More than anything, the voice behind our content gives us the credibility and trustworthiness (the ethos) needed to persuade successfully.

Thought Leadership Takes Center Stage

Ideally, shaping ethos is the domain of high-quality thought leadership: FT Longitude reports that nearly 60% of B2B business decision-makers say thought leadership directly led to them awarding business to a supplier. 

Great thought leadership doesn’t merely inform; it persuades, argues, hooks a reader and leads them on a journey. It’s where you spend Thinking Time developing unique viewpoints or conducting bespoke research. And it’s how you show the world who you are, raw and unsummarized.

GenAI tools struggle (and will likely always struggle) to create voice-y persuasive thought leadership copy. Even when trained on a company’s voice-related datasets, AI follows a formula dictated by training data and code — a formula that led to the F-shaped pattern.

Voice isn’t just a dataset of preferred diction and formatting; it’s your unique angle and sound. It’s having something to say and engaging people with your storytelling. That’s what breaks readers from the F-shaped pattern. 

Ditch the formulaic and be yourself. Readers will want to push past AI-generated bullet points to get to know (and trust) you.

Break the F-Shape with Brand Voice

I’m a writer, so naturally, I agonize over the narrator. It shapes your view of me as a competent business owner and content expert. Ethos matters to me — and I believe it matters to all company and personal brands.

I’ll keep researching how AI searches and summaries affect top-of-funnel discovery and brand awareness. Voice-y top-of-funnel content should matter for forming strong first impressions. Yet, if search goes AI’s way, I question whether voice will significantly influence the early steps in the buying process.

For now, agency leaders should help clients establish and grow their credibility with audiences by:

• Doubling down on defining and managing clients’ brand voices in thought leadership content.
• Conducting and recording more interviews with client experts to capture the nuances of their voices.
• Practicing rhetorical writing exercises with staff to master persuasive writing.

Freeing users from the F-shaped pattern will matter even more as skimming transforms into AI summaries. But that requires giving readers a reason to want to read further (which isn’t how awesome your product is and where to buy it). 

Brand voice can do that — let it shine through.

§

Stellar content about content

Microsoft Deploys Generative AI for US Spies

I've talked about how smaller, more efficient AI models living in your company's intranet are AI's path to actually helping businesses grow. But walling off your AI model from the big, wide internet and its grubby data-stealing mitts might hamper your system's performance, right?

Well, Microsoft recently launched an AI model for US spies to analyze intelligence data. The model is air-gapped, meaning it's never connected to the internet, to limit the risk of leaking sensitive information. While companies' data may not warrant top-secret clearance, this demonstrates that models can run independently of large cloud models (e.g. ChatGPT).

I'll be very curious about how this walled-garden model performs and its implications for using AI to work more efficiently with owned internal company data.

§

How to Write a Funding Announcement Press Release

by Ashley Twombly, York IE

I like what Kate Campbell's team at investment firm York IE is doing regarding content in the SaaS/tech space. They've produced solid long-form content, including a good funding announcement press release breakdown. I've written my fair share of these documents, and this article reflects what I saw on the ground.

Give them a follow and try their templates.

Content from my pocket of the galaxy

🎥 AI is for Admin

As agencies struggle to determine generative AI's place in their workflow, much attention has been paid to copy and image creation.

However, AI and its ability to parse context from data offer even more advantages elsewhere.

Let's discuss administrative tasks and how AI helps agencies better manage owned and internal data.

Click here to watch the video.

§



Sell Me a Story: Apply Storytelling to Your Marketing

I enjoyed presenting to Ball State students and marketing professionals at BSU's Digital Marketing Summit. I've linked a PDF of the slides (again) so you can see what we discussed.

Any day I can introduce people to the Complaint Tablet to Ea-Nasir is a good day.

§

See you soon,

Alex