Send us Fan Mail

On October 22, Amazon renamed Kindle Unlimited to Kindle+ and put audiobooks inside the subscription. The headline was the rebrand. The real change is the default: an AI-narrated version of any enrolled ebook now goes onto the shelf automatically, with no decision and no studio.

Meanwhile the industry's own numbers say US audiobook sales hit $2.43 billion last year, 58% of American adults have listened, and only 16% of listeners have ever tried an AI-voiced book, with willingness to try one falling from 70% to 61% in a year.

In this episode, Jeff lays out why a #1 bestseller badge and six hours of the author's own voice do two different jobs, why the buyer who hires you is more likely to be in a car than at a desk with your paperback, and how he has chosen 210 audiobooks since 2009 with one rule: is the author reading it?

Amazon built the shelf. It did not build your relationship.

The Move: read it yourself, on purpose.

Send us Fan Mail

Two years ago, Instagram told businesses to keep video under three minutes. This week it announced a test of videos up to ninety minutes, in the main feed, with a path to your television. Same company, same customers, opposite instruction.

In this episode, Jeff walks through the two reversals, why the stated reason has shifted from how people behave to where people are, and what it means when a platform asks businesses for their most expensive asset, long-form video, in exchange for distribution it can re-price at will and no share of the revenue.

The heart of the episode is a distinction every CEO can use: two videos can look identical on the day they post. One is built for the feed and belongs to the platform. The other is built to answer a question your buyer must have answered before they’ll trust you, and that one is yours wherever it lives.

The move: the next time someone says, "We should be doing video," don’t answer with a platform. Answer with a question.

Send us Fan Mail

On October 22, Amazon renamed Kindle Unlimited to Kindle+ and put audiobooks inside the subscription. The headline was the rebrand. The real change is the default: an AI-narrated version of any enrolled ebook now goes onto the shelf automatically, with no decision and no studio.

Meanwhile, the industry’s own numbers say US audiobook sales hit $2.43 billion last year, 58% of American adults have listened, and only 16% of listeners have ever tried an AI-voiced book, with willingness to try one falling from 70% to 61% in a year.

In this episode, Jeff lays out why a #1 bestseller badge and six hours of the author’s own voice do two different jobs, why the buyer who hires you is more likely to be in a car than at a desk with your paperback, and how he has chosen 210 audiobooks since 2009 with one rule: is the author reading it? Amazon built the shelf. It did not build your relationship.

The Move: read it yourself, on purpose.

Send us Fan Mail

For about a decade, your website and Google had a deal: describe your business in a machine-readable format, and Google decorates your listing—stars, question boxes, extra height on the page. Since 2023, Google has been taking those decorations back, one at a time, and nobody got penalized. The reward just stopped showing up, and the line item didn’t.

In this episode, Jeff answers the question, "Why are we still doing this?" and how it is the wrong question. The audience for that work changed — from a human scanning a results page to a machine composing an answer — and most companies never noticed, because decorating and describing looked identical on a monthly report.

He offers one test to tell them apart, one question to ask whoever runs your marketing, and two self-audit questions for the week.

Send us Fan Mail

Most AI visibility work chases citations. This episode shows why that is only half the answer. New research compared the same questions to ChatGPT with search on and off, then looked at brands that were actually on the pages the AI cited.

The model named brands it rarely mentioned from memory only about 11% of the time, even when cited. It named brands it already knew 99% of the time. Same pages, same citations. The difference was what the model believed before it searched.

Jeff explains why a citation is a ticket, not a seat — and why the two-to-five-month lag between a model's training and its release means what gets written about you this fall is what next spring's answer believes.

Proof: Daniel Goodwin of Provident 1031, whose #1 Amazon book, standing Kiplinger byline, and Masterclass series built a reputation the search only had to find.

Send us Fan Mail

Over the last twelve months, the sentence Jeff hears most from business owners is "AI killed our search traffic." Here's a second sentence from the same twelve months: Google's paid clicks grew 13%, and Google search advertising brought in $63 billion in one quarter. One of those sentences is the headline. The other is the explanation — and most companies have them backward.

In this episode, Jeff slows down the monthly review where "AI did it" shows up on a slide and nobody checks. A new three-year analysis of roughly 10,000 websites shows organic losses are concentrated in publishers and large sites; the typical small business site was roughly flat. Meanwhile, Google moved the goalposts: a sticky "Sponsored results" header, ad blocks roughly doubled in size, the first organic result below the fold, and aggressive algorithm updates. And the suspect everyone names — AI referral traffic — is about one visit in 200.

Three suspects. Two of them can be answered. The one you can't do much about is exactly why it's the convenient one to name.

Send us Fan Mail

This fall, the internet's most trusted reference sites lost Google visibility at the same time. Wikipedia hit its lowest reading since June. Every dictionary in a 2,600-company index fell in the same five weeks. Meanwhile, video, reviews, listings, and discussions gained — YouTube alone added more than the next four gainers combined.

Authority didn’t save the dictionary. What separates the winners from the losers is whether a sentence can replace you. An AI Overview absorbs a definition, and the click never happens. A customer story, a video, a review — the AI can only point at those, and the click survives.

In this episode, Jeff explains why being the answer is now the way to lose if the answer isn’t yours to give, shares how a Portland periodontist built a hundred things no AI could absorb, and gives CEOs a one-hour audit to sort their site into definition and evidence.

Plus: Why Google’s longest spam update in years didn’t move the index at all.

Send us Fan Mail

In 1975, Sylvester Stallone was broke, and a studio kept raising its offer for the Rocky script — reportedly approaching $350,000 — on one condition: a bankable star plays the lead. He said no every time, until they let him play the part himself on a budget of about $1 million.

Most people stop the story there. Jeff doesn't. The price of keeping the role was the budget: no crew, no permits, no safety net, and the famous museum steps run shot at dawn on a prototype rig because they couldn't afford to close a street. The constraint became the movie.

Every business owner has a version of that table — a buyer, a partner, a big client, a term sheet with one line that sounds like a detail. This episode is about what conviction actually costs, why you can't keep control and comfort at the same time, and how to know in advance which part of your business is the role and which part is just the script.

Jeff also tells his own version: the clients he's let go, kindly, because they needed to lead.