Amazon is breaking the product title in two.

Starting July 27, your title splits into a 75 character Product Name and a new 125 character Item Highlights field. Titles cap at 75 characters, down from around 200, in every category except media. The change is already showing up in the wild on mobile.

The Product Name carries identity. What the thing actually is. The Item Highlights carry relevance. The use case, the feature, the material, the reason to buy. Both still feed SEO. This is not a cosmetic trim.

Why Amazon is doing it

Andrew Bell follows where shopping is going and says it makes sense. On a mobile card, 75 characters is enough to recognize a product at a glance. On Alexa for Shopping and Echo, a short name sounds natural read aloud, and the 125 character layer becomes the follow-up Alexa pulls from to answer use case and fit.

Long keyword-stuffed titles index terms but wreck comprehension on a small screen. Amazon also wants product data that travels cleanly across storefronts, assistants, and agentic commerce systems. Product pages are becoming structured inputs, not standalone destinations.

The part that should worry you

Miss the deadline and Amazon rewrites the title for you. Its AI, on its own schedule. A weak machine rewrite can cost you sessions, ranking, click-through, and conversion. You do not want to hand that decision to Amazon on your hero listings.

And no, SEO is not dead. Title dependency is what changed. Identity has to work harder in the 75 characters. Relevance has to work harder in the 125. Everything else, bullets, backend terms, attributes, images, A+, reviews, price, still supports the product graph.

What to do before July 27

Split test your low priority products now. Learn how the new format behaves on stuff that does not matter, so you are ready when you touch the hero ASINs. Do not blindly trust Amazon's listing enhancer to make the call for you.

Move your converting features into Item Highlights. Leak-proof, dishwasher safe, no metallic taste, that kind of language. Keep your money keyword and product type in the Product Name.

A tool built for exactly this: Superfuel

Superfuel is an AI agent built specifically for the title split. It reads your product data and your competitors', keywords, search volume, sessions, CTR, CVR, price, and verified specs, then writes both the title and the Item Highlights inside Amazon's limits.

It picks from six strategies depending on the product, from protecting one dominant keyword to a clean hook built for the click.

Here is the part that separates it from every other title tool. After a change goes live, it measures the real impact and strips out ad spend, price changes, deals, and seasonality, so the lift you see is the title's own.

Then it gives a verdict: keep, monitor, or revert. Anything with a negative impact gets reverted automatically, and every change is logged.

You approve every change before it goes live, auto-publish is optional per product, and you can test it on an ASIN without connecting Seller Central. Results take about two to four weeks, since that is how long it needs to isolate the title's real effect.

One customer, drinkware brand Fellow, credited a single title change with $6,500 in added revenue even while category searches were down.

The clock is ticking. July 27.

Profound tracked around 1,000,000 ChatGPT shopping offers over 30 days. Then they backed it up with an 8-month look back covering 548 million offers.

This is the closest thing we have to a peek inside the black box.

Here is what they found, and what you do about it.

First, how ChatGPT shopping actually works

Someone asks ChatGPT a question. Shopping mode triggers. ChatGPT hands back 3 to 5 products in a carousel, each with its own attributes.

That carousel is the new shelf. Top slot wins. Customers are far more likely to click and buy the top offer, same as page one on Amazon.

ChatGPT fills that carousel from two very different places:

Those two sources give ChatGPT completely different information about your product. That difference is the whole game.

The headline: feeds win the top slot almost every time

Of every product citation that came from a direct product feed, roughly 99.9% landed in the number one position.

Read that again. Feed-sourced products basically don't show up in second or third. They show up first.

Feeds hand ChatGPT clean structured data it can't reliably scrape from a webpage. Brand name, checkout image, merchant subtitle, and the "Best Price" tag showed up on 100% of feed offers. On PDP-only offers, brand and checkout image showed up 0% of the time.

Structured beats scraped. Every time.

The catch: PDPs still do most of the work

Here is the part that keeps PDPs alive.

88.29% of all product offers ChatGPT served still came from web PDPs, not feeds.

Even among the 150 merchants who had already integrated their feeds, 75.81% of their offers still came from PDP crawls.

So feeds own the top slot, but PDPs own the volume. You need both. Ignore either one and you leave money on the table.

Feeds are eating the market in real time

Late last November, feed retrieval was 4.3% of all shopping pulls. In the last six weeks of the study it hit around 20%.

That is roughly 15x growth this year.

The direction is not subtle. ChatGPT is leaning harder on feeds every single month. If you are not integrated, you are slowly going invisible to AI shoppers while your competitors are not.

One more tell: feed offers leaned heavily on Shopify storefronts, thanks to Shopify's OpenAI partnership. PDP offers were spread across the big-box retailers. If you sell on Shopify, the on-ramp is already built for you.

What to actually do

1. Integrate your product feed with ChatGPT

This is the big one. It is the single highest-leverage move in the whole study.

A feed gives you control over what ChatGPT sees instead of praying the crawler reads your page right. And it is your ticket to that near-guaranteed top slot. On Shopify, you are most of the way there already.

2. Fight for the "Best Price" tag

The Best Price tag is machine-assigned to the lowest price ChatGPT can find for that exact product. It showed up on 100% of feed offers and became a clear ranking signal.

Watch competitors selling the same item. Be the lowest price when you can. Then put price and value language right in the PDP copy so the intent is unmistakable.

3. Build trust surfaces on the page

The top-ranked PDPs were loaded with proof. Descriptive, use-based customer reviews. FAQs. Q&As. Videos. Images.

Thin pages lost. Rich pages won. Give the crawler something to chew on.

4. Fix your product titles

Clear and descriptive beat clever and niche. The winning titles spelled out real use cases and features in plain language.

Name the product for how people actually search for it, not for how cute it sounds in a brand deck.

Discovery is moving. The carousel inside ChatGPT is becoming a shelf that sits before Amazon and before Google, and the rules for winning it are already being written in public.

Feeds win rank. PDPs win reach. The Best Price tag, real reviews, and clear titles decide the rest.

The sellers who set this up now get a running start while everyone else is still arguing about whether AI shopping is real.

It is real. The data is a million offers deep.

Years ago, Sam Hinkie ran the Philadelphia 76ers.

If you follow basketball, that name means something. Hinkie led the Sixers through what people later called "The Process." The plan was to lose enough games to build the team through the draft, stack the roster with young talent, and win big down the road. Basketball's version of Moneyball.

He's not with the Sixers anymore. But a lot of people still consider him one of the sharpest sports minds of the last 20 years. And he once said something that stuck with me:

"Get you some players, and then let them play."

Here's what that means for you.

You Can't Sub Yourself In

As the GM, Hinkie couldn't walk onto the court and put himself in the game when the team was down at halftime. It's against the rules. It would crush the players. And it wouldn't work anyway.

Same goes for you.

If you want your Amazon business to grow, it cannot run on you alone. You can't write every listing, answer every customer message, manage every PPC campaign, and chase every supplier email. You hired people. You have VAs, a PPC manager, an ops lead, maybe a brand manager. They can do the job without you hovering over their shoulder.

The number one job of a CEO is to hire great people and develop them. That's it. That gets you out of the weeds and into an actual management seat.

And if your team can't deliver, you have two choices:

Notice what's not on that list. "Jump in and do it yourself" is not an option. That's the trap that keeps sellers stuck at seven figures forever.

The Founder's Journey

Every founder goes through a change as the company grows. Most fight it.

You start as Michael Jordan. Running the floor, playing every position, scoring 50 a night. You source the product, you launch it, you run the ads, you handle the returns. And for a while, that works.

It does not scale.

At some point you have to become Phil Jackson. You stop scoring and you start coaching. You motivate the team, you get people working together, you draw up the plays that beat your competition. Your job is no longer to do the work. It's to make sure the work gets done well.

Then, if you build it right, you move to the owner's box. You hire executives who can coach for you, so the whole thing isn't riding on you standing on the sideline every night. Jerry Reinsdorf made Phil Jackson head coach in 1989, then watched from up top while the Bulls won six titles.

That's the path. Player to coach to owner. Most sellers get stuck on the first rung because letting go feels like losing control.

Let Them Cut Down the Nets

One more thing from Hinkie's talk.

When a team wins a championship, the players cut down the nets and lift the trophy. Not the GM (unless you are Jerry Jones).

If you build a business that's winning, give the credit to the people who did the work. Let your team celebrate. Let them feel the reward. Don't run out and grab the scissors yourself.

You're the CEO. Wanting recognition for the company you built is natural. Everybody feels it. But unless you're a true one-person shop, watch what happens when you let your people cut down the nets. They work harder. They stay longer. They start acting like owners.

A Note on Hiring

Since we're on the subject of getting the right players, one tool worth your time in the hiring process is a personality assessment.

Most sellers hire on gut and a resume. Then they're shocked when the new VA can't handle ambiguity or the ops hire hates talking to suppliers. A good assessment tells you how someone communicates, how they make decisions, and how they work with others before you sign them on.

The big companies figured this out a long time ago. Google uses assessments to find people who fit the culture, not just the skill test. Southwest hires for the service mindset. Marriott uses the results to build training around each person's strengths.

You don't need to be a giant to do this. For remote teams especially, where you're hiring people you may never meet in person, knowing how someone is wired is worth a lot. And the same tests tell you plenty about yourself, which helps you figure out what kind of people you actually need around you.

Get you some players. Then let them play.

Visibility Labs ran 20,000 ChatGPT responses to see how much product recommendations shift when web search is on versus off. The gap is bigger than anyone expected.

Turn search on, and 80.2% of the recommended products change. Only 19.8% of what ChatGPT suggests from memory alone survives once it starts pulling from the web.

Here's the part that should stop you cold. Products that got recommended 100% of the time with search off? Only 15.8% of them still showed up with search on. The "sure things" were the least safe.

A few more numbers worth knowing:

ChatGPT names about 5.2 products per answer with search on, 6.2 with search off. Run the same prompt 10 times and you get around 19 unique products with search, 21.8 without. So search actually narrows the field a little, and it reshuffles almost the entire deck.

Why does this matter? Because search is the default now. When a real shopper asks "what's the best [your category]," ChatGPT is reading live web pages, not what it memorized during training.

And the study found the thing that moves the needle. The more often your product shows up in the sources ChatGPT cites, the more often it gets recommended. Products cited zero times barely registered. Products cited around 3 times per answer were the ones getting recommended every single time.

The seller takeaway:

Get your product into the articles, roundups, and pages that ChatGPT is already pulling for the queries you care about. That is AEO in one sentence. Be in the sources it reads, and you are in the answer it gives.

DragonFish can do this for you with its proprietary Bone Digger tool. We find what the sources getting cited are, and tell you what you need to do to instantly piggy back on that.

✌🏼 See you again Thursday …

The answer to today’s STUMP BEZOS is
17.3% of those who click on a live selling event buy