Happy Wednesday.
Or, bon mercredi to my industry friends in Cannes. We’ll see if it makes sense for The Aisle to have a presence there this time next year. For now, I’m settling for an Mbappé pronunciation debate with my kids over a wheel of brie. I’m OK with that.
A couple of news items that caught my attention this week…
Amazon is rolling out conversational AI-powered display ads across the web—beyond Amazon’s own digital walls— that include clickable product questions, or prompts, which route shoppers into a chat with Alexa for Shopping (the AI assistant formerly known as Rufus). An Amazon exec said that 20% of people who interact with a prompt continue the conversation, and Amazon says 7 in 10 are new to the brand.
I admit this news gave me a bit of a flashback. Many years ago, I covered digital media companies and ad tech at Ad Age, and remember reporting on former AOL CEO Tim Armstrong’s master plan to reinvent display ads back in the early 2010s. The ad format was called Project Devil (no, I can’t remember a good reason for the name) and featured massive, highly interactive premium display ads. Anyone remember that? Well, I’m sure Amazon ad execs would hope their new format lasts longer than Project Devil did.
Amazon ads also started showing up on ChatGPT in what could either be a short-lived test by the e-commerce giant, or an early chapter of a new long-term advertising relationship replacing the Amazon-Google one.
Lastly, the consumer tech company Glance announced a deal with Samsung to power immersive shopping across Samsung TVs, using a viewer’s selfie as a foundation for AI personalization. I’ve long been skeptical of TV commerce, or T-commerce. Though with video giants TikTok and YouTube investing in commerce initiatives, perhaps this era will be different. You may just have to deal with bigger hair, like my alter ego was sporting in the demo below, from Glance's booth at Shoptalk back in March.

Now on to the good stuff…
The Center Aisle

Wirecutter editor-in-chief Ben Frumin
Disclosure from the top: I first met Ben Frumin almost 20 years ago in graduate school, and I like him a lot. We remain good friends, even though he and his now-wife refused to come to the bar I wanted to visit on my 25th birthday because it was one whole mile away from his apartment. I caved and went to them, and obviously still hold a bit of a grudge to this day. And he still doesn’t feel bad.
But putting all of that aside, Frumin is one of my favorite thinkers and leaders in journalism today. He’s run the newsroom at Wirecutter, the New York Times’ product-review giant, since early 2019. During that time, he’s led a massive expansion of both the staff as well as what the publication covers.
Wirecutter's business model is straightforward: it recommends products, links to them, and earns a commission when readers buy. That affiliate revenue is the heart of the business model. For heavier users, there's also a standalone subscription, as well as access through the Times' premium bundle. The Times doesn't break Wirecutter's financials out separately, but the digital affiliate and licensing segment it describes as consisting primarily of Wirecutter affiliate referral revenue totaled $201 million in 2025, up 8% from the prior year.
That model is under pressure, though. AI shopping tools from OpenAI, Google, Amazon and others increasingly promise to do some of what Wirecutter does—tell you what to buy—without sending readers to Wirecutter first. Prime Day, which is happening as I write, is Wirecutter's single biggest moment of the year. It felt like the perfect time to be asking Frumin about his plans for Wirecutter to survive, and maybe even thrive, during this massive change.
There's another layer Frumin couldn't discuss: the New York Times is currently suing OpenAI and others over copyright infringement, and Wirecutter's affiliate model is at the core of the harm being alleged. His communications team said the litigation was off limits. But its shadow hangs over just about everything we talked about.
Since you became editor in 2019, Wirecutter has grown well beyond its original guide-centric model. How would you describe that evolution?
Ben Frumin: Forgive me, your readers may know all this, but I’ll just say, Wirecutter was founded as an independent entity in 2011. The New York Times bought us in 2016. I have been our editor since, I think, the first week of 2019, so call that seven and a half years. Wirecutter has changed a ton during that time, for a number of reasons, some being changes to the way people search and shop and consume content, but also for many other reasons.
When I started, we had, I think, 80 journalists in our newsroom, and more or less everybody was pointed at doing the one thing that Wirecutter did, which was maintain this editorially exceptional catalog of, let’s say, 1,000 product reviews, or guides, as we call them.
Over the last few years, we just do way more things and way different things. We have way more people doing those things. I think we have nearly 200 journalists, and we’re just doing…all sorts of new things, whether it’s new formats, like we have a podcast, we have all kinds of video offerings, we have a book coming out.
You have a book coming out?
BF: Maybe I’m breaking news, but we do have a book coming out next year. Stay tuned.
To state the obvious, search traffic to a lot of different publications has declined. I’m curious how that has hit you guys, and how you’ve thought about replacing it?
BF: So much of Wirecutter in the very early years was pointed toward high-intent searchers looking for product reviews and advice, and that was very much by design, right? If somebody is searching for the best air purifier, we did the best testing, and we have the best advice, and we want them to find our thing. There’s a little bit of, you want to meet the audience where they are.
In the early years, a lot of people were looking for shopping advice and product reviews by searching for it. But that has changed a ton over the 15 years that Wirecutter has existed. People are looking for advice on what to buy, what to avoid, where to shop. Some people are watching videos, some people are listening to an influencer whose advice they might take.
What we have really focused on is trying to develop vastly more direct relationships with our audience so that we’re not vulnerable to the capricious whims of any number of tech platforms or modes of finding information, search being among them. We have for years invested, just as one example, in a really robust email program, where we have millions upon millions of people who want us in their inbox every day.
I am well aware from my inbox. There are things I did not know I needed or wanted a recommendation on.
BF: I won’t even ask.
I would think that AI summaries and AI search would make a paid subscription offering much tougher today than when you launched it in 2021. Is that a fair assessment?
BF: I’ll say a couple things. We’re real people testing real products in the real world, and whatever slickly synthesized or confidently asserted snippet or summary somebody might see anywhere, whether it’s in an AI summary or anywhere else, it’s the work that gets done before that that matters.
We’ve got a couple hundred people testing thousands and thousands of products a year, and so we believe that is journalism that, one, is worth paying for, but two, really can’t be replicated. Without it, what is there to summarize? What is there to snippet?
No robot at least as far as I can tell is going to be testing lip glosses and air purifiers and washing machines and all this stuff. That’s the core value: the testing and research and reporting rather than the output on the other end.
I assume, though, there are consumers for whom an AI snippet of a review might be enough. Whether or not you think it should.
BF: I am sure there are people out there in the world for whom that is true. And I will say, one of the things that I think is incumbent upon us is, AI can have tremendous value in how we express or personalize our journalism in a way that makes our service feel even more valuable.
You’ve probably seen this, but we have this feature, we call it Wirecutter Finder— basically AI-powered search on our site, which is an early version of this future we imagine where AI is not replacing the journalism but it’s almost personalizing or unlocking the value in it. If you’re like, “Yo, I need a webcam that works in my attic, where my Peloton is being used as a laundry rack,” or whatever, it might be hard to find the right thing if you’re going through several thousand words yourself.
We see AI as a tool this way. To your original question of, is a snippet enough for somebody? Sure. Is a product recommendation devoid of any context enough for some people? Of course. There are plenty of people who just go to Amazon and buy the first thing they see.
But for our discerning, curious, thoughtful audience, of whom there are millions upon millions, we know they value all the work that went into it. To me, it’s how do we learn from some of the ways that they’re consuming information to bring that service on platform, help personalize or tailor the way we express the results of our journalism for them.
Do you know what the breakdown is of people that click on one of those pre-populated questions you surface alongside the Wirecutter Finder AI search box versus asking free form questions?
BF: The pre-populated questions are in some ways meant to show our audience that this isn’t just a basic keyword search. I think people are so used to, over decades, just typing in almost broken phrases and half sentences or keywords. I think people are still learning, “Oh, I could actually have a conversation. I could actually write a couple sentences asking for something specific.”
It’s really meant as a prompt to show somebody, not that you should ask this specific thing, but these are the kinds of things you can ask or this is the kind of way you could cast things.
But I will say that Finder has been very, very popular. We’re in a situation where we want more people to use it. We’re trying to put more prompts on more pages to direct people to it. When they use it, they really like it.
Certainly, there are some people who will always want to read a long review or view photos. There are other people who, again, it’s early days with this thing, but it’s a really good and personalized experience that we find them seeking out.
What we’re seeing all the retailers talk about, whether you trust all their numbers or not, is if you’re using one of the AI assistants within their retail site, you’re this much more likely to buy, or your basket size might be double the size. Do you have any data you can share about what you see there in terms of interacting with the Finder feature and the likelihood of going on to convert?
BF: Look, I am not a businessman, so I’m never going to think about stuff from the perspective of increasing cart value or orders, blah blah blah. But what I will say is directionally, we do see a very clear link between use of Finder and a much more engaged user experience. We do find people reading more, clicking more, coming back more. It really does help with engagement, or I should say, there’s a clear correlation with engagement.
What have you been thinking as you watch some of the retailers create their own assistants? There’s obviously Amazon’s Rufus [now called Alexa for Shopping], there’s ones at Home Depot, Lowe’s. One could make the argument they’re competitive with you guys. How do you look at that?
BF: Theoretically, any service or organization that is giving people advice about what to buy or not buy, and how to think about it, is competitive with us. It’s not just Consumer Reports and The Strategist and other traditional editorial outlets. The Reddit Buy It for Life subreddit competes with us in ways. Retailers compete with us in some ways.
That said, we see our core differentiator as the couple of hundred journalists who we have independently and rigorously researching, reporting, testing. Not to be a broken record about that, but it is the differentiator for us that we are this human-led, independent journalistic enterprise. That’s just definitionally not true with some of these in-house AI chatbots on a retail website, or any number of other things that might, at the most expansive definition of this, be competitive with us.
That all makes sense. I’m wondering how much more you feel like in this era you need, or your team or marketing needs, to be out there screaming that.
BF: A lot. It’s interesting, I think over our history we haven’t always excelled as much as we should in really foregrounding or illuminating the humans at the heart of Wirecutter.
If people find out what I do for a living, people get really excited. They love Wirecutter. And if I meet a Wirecutter superfan out there, and I’m like, tell me your three favorite Wirecutter journalists, a couple years ago, they couldn’t name a single person. Their relationship was with Wirecutter, the brand, versus our individual journalists. That’s something that is really changing.
Is that because of video or other things?
BF: I think there are any number of things. We are doing vastly more where we’re foregrounding our expert journalists. Early on, we almost wanted our journalists to not quite be invisible, but to recede into the Wirecutter house voice and process. We would tell the epic story, but it was less foregrounding this particular individual.
We’ve really been pushing to change that. We have vastly more first-person, journalist-centered pieces. Whether it’s a podcast, where you have a frankly pretty intimate relationship with a host of that podcast, whether it’s things like video, where again, the person couldn’t be more foregrounded, we really are trying to show this core differentiator: we’re nearly 200 extremely expert and passionate independent journalists who are doing this thing. There’s almost like we can’t do enough of it, but it has been a really big push over the last couple years.

Wirecutter pickle testing. Yes, for real.
Are there product categories that made sense to really target pre-AI engine boom that don’t as much now?
BF: Just generally, we have dramatically expanded the number and types of categories we cover. I think in the early days, to the extent that Wirecutter lived in the popular imagination, people thought of it as largely a consumer technology website. When really the old Sweethome categories, all the softlines home-goods stuff, was tremendously important to Wirecutter from a journalism and audience and business perspective for a really long time.
But we have really expanded in the last couple years into what we call more lifestyle categories. Whether that’s beauty and personal care, fashion and style basics, home decor, gifts. These have been tremendously successful and engaging with our audience.
I’ll tell you, I’m going to protect the innocent and not name this person, but we did our first skincare package a year and a half ago. We published a handful of guides to skincare products, vitamin C serum, moisturizer, etc.
You use all those, I can tell.
BF: I don’t even use soap. But I had a colleague at the Times, this is probably four or five days after we launched this package, and she walked into not one but two department stores in Tokyo, and two separate salespeople walked up to her and were like, “Are you looking for the Wirecutter-recommended skincare products?”
It was extraordinary that that’s not what most people historically have thought of with Wirecutter. They think about laptops or phones or smartphone devices. We have seen just tremendous audience growth and engagement and diversification in all of these new categories. And look, many of them are pretty personal things, and also not necessarily the sorts of things that you have a high-intent precision search for. If I’m buying you a birthday gift, it’s not the same as googling the best air purifier. I’m looking for that advice and recommendation in a different way. So many of these new categories are just more personal, where you trust our experts’ taste, and you’re seeking that out more generally.
In the example you gave of the store in Tokyo, is that retailer paying you guys something?
BF: No.
I thought you had some sort of partnerships that expand into the physical world. I could see a world where more Wirecutter badging in the physical world could make a lot of sense, with some people sort of going back to the physical world as a means to get away from AI taking over everything online.
BF: Totally. We want Wirecutter to be everywhere. Our traditional way people found us and the way our business model worked was almost entirely online transactions in this very high-intent way. But we want to be everywhere that people shop. We want to be everywhere that people are thinking about what to buy.
You mentioned badging. We do have a licensing or badging business like that, but we also have been exploring various partnerships. What does it look like if you’re going to have Wirecutter badging in store? What if there was an end cap of Wirecutter picks?
I’ll share this example because it’s public, but we have a partnership with Sam’s Club that we just launched this year, which is exactly that: highlighting independently tested and anointed Wirecutter picks that are available at Sam’s Club, with various badging and signage and other things in store. We think there’s a big opportunity there for sure.
On some of the AI engines, you can now create your own app experience within the AI app. I assume that would not happen while litigation is going on, but at a high level, hypothetically, could there be a world where a Wirecutter experience directly within an AI service could make sense from your editorial point of view?
BF: Again, I am not a businessman involved in all these various discussions with any potential partners. I will just say editorially, I believe that the work that we’re doing, that all of our expert journalists are doing, is the best in the business. There’s so much noise out there. I sometimes describe this as the blizzard of bullshit out there, where there are so many, frankly, cynical or bad actors looking to bamboozle and hoodwink people.
I know that our advice is independent and excellent and trustworthy, and will actually make people’s lives better. In that sense, just speaking editorially, I want that advice to reach as many people as possible, and in as many ways as possible. To me, just editorially, I want Wirecutter’s journalism to be anywhere that people are asking the sorts of questions that we can answer or raising the sorts of problems that we can solve.
Do you have any go-tos for shopping besides Wirecutter, whether it be Reddit or anything else, or a new AI-native recommendation engine?
BF: Obviously, I always go to Wirecutter first. But yes, I regularly read all of our competitors, and I’m curious about their advice, or how they’re casting stuff, or sometimes if they cover categories that we don’t, how they’re doing it, or what their advice might be. We cover a lot. We don’t cover everything.
I do find that something like Wirecutter Finder is the thing that I find myself the most hungry for when I’m having trouble finding the thing that’s right for me. For instance, the fridge in my kitchen, I thought I was gonna have to replace it, this is probably a year or so ago, but the dimensions of it are weird. There’s an overhang cabinet. It’s just weird dimensions. I found the experience of using, in this case, Wirecutter Finder, but I will just say a personalized agent like that, to be helpful in figuring out, okay, I know this is the fridge that’s great, but will it work for me?
I think that is a huge opportunity for us, and something we can learn from, the ways that people are shopping and asking questions and getting answers to them. How do we make sure that we’re not just giving the best recommendation for most people in a way that requires folks to search for it, but helping people find the thing that’s right just for them?
I’ve thought a lot about how there’s all this talk about who will build the everything personal shopping agent. I think some folks assume, if it does happen, that that company might have a retail background or a combination of a retail and AI background. But it would be interesting to think about whether someone with a deep editorial corpus could play that type of role beyond your own walls.
BF: I think that business or experience that you’re positing, I would say there’s an empty core of it unless it is powered by independent recommendations that a person can trust. Everything else is a little bit of a slick overlay, or the patina of a great user experience. But the thing at the core to me that sets Wirecutter apart, and that is critical to making something like that work, is actually the things that it’s recommending are good, and the recommendations themselves are trustworthy and aren’t corrupted by any business rationales that have anything at their heart other than what’s best for the shopper, what’s best for the user, what’s best for the reader.
I know you run editorial, not the business, but obviously affiliate is core to how you guys monetize. I’m curious how you’re thinking about that business model changing over the next few years.
BF: Again, with the caveat that I run editorial and I am not the businessman, affiliate revenue is a huge part of our business. It historically has been, and it is today. I believe it will be in the future.
Just as we’re trying to reach people in all kinds of different ways, where we’re not primarily looking for high-intent searchers, we’re looking to develop direct relationships with all these people, whether it’s in their inboxes, on YouTube, in the podcast app that they use, on their coffee table. We’re trying to develop all of these relationships with people.
So too will our business model surely evolve and grow. Again, it’s less of an instead-of and more of an addition. I mentioned licensing and badging and in-store partnerships. I’m sure you’ve seen we have advertisements on our site. That’s a growing business for us. There are all kinds of ways that our business is growing and diversifying, but again, not in a replacement way, but in an additive diversification and growth sort of way.
What’s one thing that someone who visits Wirecutter maybe only once a month or less would be surprised to know that Wirecutter is now doing?
BF: That’s a great question. I’m going to plug our YouTube. We just soft-launched a new long-form video operation earlier this year. We’ve been piloting the show called Wirecutter Taste Test, which is really fun. We have a couple of our journalists, and maybe an outside guest, taste-testing sardines and butter and olive oil and various things.
I think some casual Wirecutter audience members might not even know that we cover grocery staples like that. When they come across this coverage, it’s just tremendously popular. We did this package for Memorial Day of taste-tested hot dogs, ketchup, mustard, pickles, all kinds of fun stuff, and people love this coverage. It does gangbusters numbers. It is super engaging. They should check that out.
OK, last one. Should an Amazon shopper use Wirecutter instead of Amazon’s Alexa for Shopping AI assistant on Prime Day? If so, why?
BF: Wirecutter should be any discerning shopper's first stop on Prime Day—and any day.
Most so-called "deals" that most retailers push on shoppers are overhyped discounts on so-so products. Especially during big sales events, it can be overwhelming for any shopper to figure out what's a quality product they can trust at a truly great price.
That's why our newsroom of nearly 200 expert journalists spends all year independently and rigorously testing thousands and thousands of products, while our deals experts spend all year meticulously tracking prices on the great gadgets and gear that Wirecutter recommends. So that during the confusing cacophony of a big sale like Prime Day, discerning shoppers can visit Wirecutter to be sure that anything they buy is a truly great deal on a truly great product.
Thoughts or tips on the topics in this newsletter or others? Just reply to this email or message me at [email protected] or confidentially on Signal.
Interested in sponsoring The Aisle or otherwise partnering on events? Send a note to [email protected].
And feel free to forward this email to someone who might find it useful. If someone forwarded this email to you, please consider subscribing to The Aisle here.
