EP 273 – Fragmented Discovery, Google’s €460M DMA Fine, and Apple’s Distribution Bet

BySEO Guru

Local discovery is now a fragmented set of tools no single platform controls — so as Google’s results fill with ads and its regulated modules get pushed around, and as AI assistants ride device defaults into everyday use, the advantage shifts away from optimizing the search result and toward the real-world brand demand and platform distribution that decide who gets surfaced at all.

Takeaways

  • Discovery fragmented, but the human referral didn’t. Across roughly a thousand consumers in several verticals, search sits near 54% and standalone AI near 38% at the discovery stage — yet offline, human-to-human referral has held around 30–35% since Mike’s 2015 survey of lawyers. Digital keeps expanding without eating the personal referral. (02:48–04:16)

  • AI collapses at the bottom of the funnel. People lean on AI heavily to understand a problem and compare options, then largely abandon it to make contact or buy — a trust gap for local information, with Google’s AI Mode a partial exception. Top and mid funnel show real AI usage; the bottom “falls way, way off.” (06:08–07:03)

  • Attribution is a diminishing return. These were screened, real purchase paths recalled from memory across multiple tools, so last-click tells you where someone arrived, not where they decided. “It’s more important to understand customer behavior… than to invest tons and tons of effort into attribution.” (08:00–08:55)

  • Brand is the hidden entry point. Done right in a local market, brand simultaneously feeds AI, Google, and social — and it’s usually invisible to companies staring at their funnel trying to reverse-engineer what customers do. (08:55–09:14)

  • The DMA finding is sweeping — and it’s €460M. The Commission’s Article 6(5) self-preferencing decision carries a €460 million fine (the search portion of a combined €890 million penalty), orders cease-and-desist within 60 days, and reaches across roughly nine services — comparison shopping, Flights, trains, the hotel finder, Places/the local pack, and Maps among them. Noncompliance has run since March 2024. (11:39–13:20)

  • “The local pack is just search” — rejected. Mike sat in DMA meetings 2 years ago where Google argued the local pack was simply search, not an added module; the Commission explicitly rejected that, treating it as a self-preferenced module alongside ads because of its added functionality. (11:39–12:05)

  • AI Overviews are the untouched flank. The decision says nothing about AI Mode or AI Overviews — so if Google routes self-preferencing through AI Overviews with third-party links, it may sidestep the ruling entirely. The remedy also lets Google either elevate third parties or bury its own modules. (13:54–14:10)

  • The hotel SERP is ~40% ads on both sides of the Atlantic. Mike’s mobile audit (Berlin vs. US, first four screens ≈ 99% of use) found sponsored ads occupying about 40% of the real estate in both markets, with the “hotel finder” functioning as just another ad surface dominated by booking.com. Google’s claim that EU results are “the worst consumer experience ever” reads as self-serving — the results are bad everywhere. (14:15–18:35)

  • Siri’s growth is a distribution story, not a quality one. Apple cites 2.5 billion Siri interactions a day (Mike figures half are false positives), an S-curve that inflected in 2024 as Apple surfaced Siri more aggressively — usage climbed despite years of quality criticism, because it’s the default on a rising iPhone base. (20:24–24:07)

  • Wearables become the AI’s input layer. New Apple Watch chips enable on-device conversation transcription for accessibility, and $125 AirPods 5 add audio understanding — positioning the phone as the AI hub with body-worn devices feeding it, and raising Mike’s open question of whether the app survives as the interface. (24:59–26:47)

  • Apple’s privacy moat is the real asset — and it’s under pressure. The counter-example: LG smart TVs caught scanning home networks and logging standby audio for ad revenue. Apple’s privacy halo is what keeps Mike on its hardware — but Maps naming decisions (Gulf of America, then Lake Ontario) and a South Park rebrand mocking Apple and Google show the brand taking hits. (27:11–30:28)

  • For the hosts, Ask Maps beats general search. Greg says AI assistants and Google’s Ask Maps have “almost completely eliminated” his Google search use; once the habit breaks, “the mystique is completely gone.” Google’s durable moat isn’t general search — it’s the local data set behind Maps. (37:25–40:03)

Concepts

  • Decide / where-to-buy stage — the bottom-funnel step where AI usage collapses and Google functions as a proxy for offline information (phone numbers, store location, hours).
  • Self-preferenced module — a SERP unit (local pack, hotel finder, Flights) the DMA treats as favoring Google’s own service over comparable third parties; the “it’s just search” defense was explicitly rejected.
  • Places sites / aggregators — the EU-specific block of third-party aggregator listings that appears above Google’s own hotel finder in European results.
  • Hotel finder as ad surface — the framing that Google’s hotel module is not a consumer utility but another ad-delivery surface, in practice dominated by booking.com and its subsidiaries.
  • Distribution advantage — the idea that default placement (Siri on every iPhone; Gemini across Search, Android, Chrome, Workspace) drives usage more than product quality does.
  • Body-worn AI inputs — wearables (Watch, AirPods) acting as the capture-and-interface layer that feeds AI running on the phone.
  • Appless environment — a possible future in which AI delivery no longer routes through discrete apps at all.
  • Ask Maps — Google Maps’ Gemini-powered natural-language planning feature, used here to assemble a multi-stop day from one complex query.

Practitioner Notes

  • Survey your own customers before assuming the mix. The research is directional and the channel split compresses or stretches by category — an “immediate 24/7 tow” need behaves nothing like a considered purchase — so measure your own funnel by vertical rather than importing an average.

  • Move budget from attribution to behavior and brand. With multi-tool, from-memory journeys, over-investing in last-click attribution misreads causation; fund the brand and real-world demand that feed every channel at once, which is exactly the throughline of Near Media’s brand-building series.

  • For local, defend the data layer. Google’s real moat is its local data set — the thing feeding both the (now-regulated) local pack and AI answers — so keeping Google Business Profile and listings accurate is what keeps a business surfaceable as answers get generated rather than ranked.

  • Watch the AI Overviews flank in the EU. The DMA remedy doesn’t touch AI Overviews; track how Google reroutes local, hotel, and flights answers through AI surfaces, since that’s where self-preferencing can quietly re-form.

Pick your starting point:

00:10 Intro
00:57 Lastmile Retail consumer-journey research
06:08 AI collapses at the bottom of the funnel
08:00 Why attribution is losing value
08:55 Brand as the entry point
09:50 The DMA Article 6(5) decision
12:05 The €460M fine & what’s next
14:15 Hotel SERP audit: EU vs. US
16:09 ~40% of the results are ads
20:06 Apple event: Siri & 2.5B uses a day
24:59 iPhone 18 Pro, Watch & AirPods 5
27:11 The AI-hardware trust problem
28:30 LG TVs & Apple’s privacy moat
30:28 The $2,000 iPhone Duo
37:25 Ask Maps & breaking the Google habit
41:07 Wrap & Brighton SEO

Full Transcript –> on-line

Editor’s note: This is a cleaned, human-readable version of the episode. Stutters, false starts, and clear transcription errors have been removed; short backchannel (“right,” “yeah,” “mm-hmm”) has been folded into the main speaker’s turn; and wording is otherwise kept faithful to what was said. A few speaker-side factual points were intentionally left as spoken.

Greg (00:10)
Hey everybody, welcome back to the Near Media Podcast with me, Greg Sterling, and as always, Mike Blumenthal. We’re going to be talking about a lot of interesting stuff today. You can see a screen up — this is some new research we’ve done, and we’re going to talk about what it means. We’ll also discuss the DMA’s finding that Google was out of compliance with Article 6(5), and what that means. Mike has done some research on the real estate on Google’s hotel pages, comparing Europe and the US, which is pretty interesting. And finally we’ll talk about Apple and its recent announcements — Siri, and whether people are going to buy the new Duo foldable phone.

Mike B (00:56)
Start with you, Greg.

Greg (00:57)
You were frozen there for a second. So what this slide shows — this is from recent research we did on behalf of Lastmile Retail, which will be unveiled relatively soon at Brighton SEO. Recently we had Damian Rollison from SOCi talking about some of their research, and the bottom line was that the consumer journey is now pretty multifaceted — fragmented, if you want to use that term. Google is certainly the majority use case at different points in the purchase funnel, but it’s not as dominant as it once was. This was a thousand consumers across several verticals, and the slide illustrates a weighted average of what people said they used at the research and discovery phase, then comparing options, then deciding where to buy the product — or in one case, the service. So I’ve broken out search engines, AI, social media, offline (going into or contacting a store), and then the average of the rest, which includes things like YouTube. Did you want to say something, Mike?

Mike B (02:20)
A couple of things. This distribution varies by category. Sometimes it’s compressed. If you have an immediate toe need, it’s going to be different than if you want a new set of teeth, right?

Greg (02:32)
A hundred percent. Restaurants versus dentists versus primary care physician versus self-storage — it’s all going to be slightly different. These are three verticals represented here in terms of weighted averages, but directionally it’s going to be consistent.

Mike B (02:48)
And to some extent it shows how messy discovery and comparison is, and how many options people use. Obviously they add up to more than a hundred because some people use multiple things.

Greg (03:05)
People are using multiple things.

Mike B (03:08)
And yet even within that — in discovery and research, search engines are 54%, AI is 38%. The question is, how much of that 38% is Google?

Greg (03:20)
Probably not the majority of it. I forget the exact definition we used — I think we used ChatGPT, Gemini, and Claude as examples of AI in the survey. Some of it is Google, but it’s not dominated by Google. And that’s not AI Overviews or AI Mode — that’s standalone AI tools.

Mike B (03:42)
Right. And the other interesting thing: when I did a survey of lawyers in 2015, offline referrals were about 30–35%. And it’s still roughly that across all of these different areas. I find it fascinating that offline would stay at that number over time — that digital isn’t consuming the human-to-human need for a personal referral. It’s still a large player in all of these.

Greg (04:16)
Right. One of the factors in offline is people physically going into stores to see products, or to buy something and take it home the same day. Those things are still compelling. We can talk about vertical X or vertical Y being slightly different, but directionally and overall this reflects what people are doing broadly in the market. We see this in our legal research too: people use lots of different tools depending on the need and the industry — to get information, to vet it, to make decisions, and then to decide where to buy. And down in the decide/where-to-buy category, search engine usage — which is pretty much a stand-in for Google — is really about contact details and store location, sort of offline-oriented stuff. Sometimes it’s e-commerce, but most of the time in this context it’s a proxy for offline information.

Mike B (05:27)
Interesting — and in “decide where to buy,” it doesn’t add up to a hundred.

Greg (05:32)
Well, people could — okay, so —

Mike B (05:35)
It’s about 79% or something.

Greg (05:39)
I’d have to go back and look at why that is. This isn’t a zero-sum thing — these don’t all add up to a hundred because people were allowed to pick multiple choices. We gave them a matrix and asked, in this stage what do you use, in this stage what do you use, and in this stage what do you use — and they could leave it blank if they didn’t use it. So people weren’t required to pick something every time.

Mike B (06:07)
Interesting.

Greg (06:08)
One of the other interesting things is that AI falls way off — so does Google, but not as much — when you get to the bottom of the funnel. People use AI pretty heavily to research and figure out the nature of the problem or product they need, and then to compare things. But when it comes to making a business contact or buying, it’s much less used. That’s really interesting for many reasons — it shows where the trust is. As we know from multiple surveys, people don’t trust AI as much for local information. They still use it sometimes, but it’s not as trusted. Google’s an exception — I think Google AI Mode would be more trusted than something like ChatGPT. So top and mid funnel you have AI usage, and at the bottom of the funnel it falls way off.

Mike B (07:03)
Just to summarize for the people listening rather than watching: discovery and research are very similar. Roughly 50% is Google, roughly 37% is AI, roughly 33% is social — it varies a little but stays in that range. It’s interesting that it stays fairly consistent across both stages. Which essentially says you have to understand how your customers are finding you, and then focus and do well on the channels where that’s happening. Each category, product, and market may be slightly different, so it’s important that you survey your customers and see what they’re doing.

Greg (08:00)
And in fact this raises the whole point about attribution again — because people are using multiple tools, and they’re doing this from memory. These were supposed to be actual purchase paths; the respondents were screened, so these are real purchase paths as opposed to hypothetical use cases. But they’re recalling from memory, so it may not be exact. Given the multiple influences on a purchase, attribution is less and less valuable. You can see where they came from at the last click, but that doesn’t mean that’s where they decided they wanted your product or service. So the point you’re making is really important: it’s more important to understand customer behavior and the channels they use than to invest tons and tons of effort into attribution.

Mike B (08:55)
And to elevate it one more level — referencing our five episodes on brand — when done correctly, brand influences AI, Google, and social. Brand is an entry point into all of these, if done correctly in a local market.

Greg (09:14)
Yeah, the point about brand is very interesting. This gets into a long digression we probably don’t have time for. Brand has a really powerful effect in certain situations and is less powerful in others. For example, if you’re buying a commodity, price starts to enter the mix, and people will often abandon the brand for a cheaper version. Think about grocery stores today with all the inflation — people used to be very brand loyal, and now they’re gravitating toward store brands or white-label products because they’re cheaper. There’s a more nuanced set of factors at play, but what you say is absolutely right: brand is a very powerful influence, and it’s often hidden from companies trying to figure out what their customers are doing.

All right, so now we move on to the DMA and its recent, pretty spectacular — I use that in a neutral sense — finding that Google was out of compliance with Article 6(5), because Google ranked its own services more favorably in search results than third-party services. This has been a long time coming — there have been all these notices and warnings, and Google has been negotiating with them to modify the SERP and doing all kinds of experiments we’ve talked about. This is exclusively in Europe. The finding says noncompliance has been in effect for over two years, since March of 2024. The striking thing is the sweeping nature of the decision. It encompasses comparison shopping, Google Flights, and it says trains — probably any module reflecting transportation — hotels, the hotel finder, Google Places (the local pack), and Google Maps, which gets a privileged position in Google search results. And then some other stuff like translation. But go ahead.

Mike B (11:39)
One comment on this. When we went to the DMA meetings several years ago — David and I — Google was explicit that the local pack was just search, it wasn’t something extra. It’s fascinating to me that obviously the people on the commission are not stupid, and they’ve included it as a self-preferenced module within Google, along with ads and other services.

Greg (12:05)
They explicitly rejected that argument — that these modules (Flights, the local pack, and others) were just part of search — because they have additional functionality. I don’t have the specific language, but they took that argument into consideration and explicitly rejected it. So to finish out what this means: there’s a €460 million fine, which is in excess of 500 million dollars — nothing to Google, but still significant in the absolute sense. And more importantly, Google is required to cease and desist its non-compliant ranking practices within 60 days. They’re supposed to submit a compliance plan for how they’ll remedy this, and then not repeat the conduct in the future. Google can appeal, and they almost certainly will because of its sweeping nature. It’s unclear what will happen at the appellate level — it may be upheld, or overturned in part. The commission doesn’t specify exactly what Google has to do to remediate: Google has the option to elevate third-party services and information, or it can reduce the visibility of its own services — push the local pack way down the page. The other significant thing is that there’s no discussion of AI Mode or AI Overviews, so that’s not included. I suspect that if Google simply presents AI Overviews with third-party links, that might escape this non-compliance — but it hasn’t been dealt with. So, you can take us into the next part.

Mike B (13:54)
So Google, as part of this, released information about the results of the changes they’d made to EU hotel results, and proclaimed that they were the worst consumer results ever — a terrible experience.

Greg (14:10)
It’s a bad experience, right.

Mike B (14:15)
It’s a tremendous decline in user satisfaction, they said. As a result, I was curious what these things actually look like. So I went out and did some mobile searches — this is “hotels near me” in Berlin. The dotted red line is roughly one screen view. It’s one ad; the next scroll is a few more ads. Then we get to the unique European piece, which is “places sites,” which are largely aggregators. Then we get to the hotel finder. Note that the hotel finder isn’t even — it’s like the fourth screen down, pretty far down — and then a whole bunch of Google surfaces: the first organic result is an aggregator, another Google “people also search for,” then two more aggregators, and then, interestingly, in Europe Google is regularly indexing their own search results. People only look at the first four screens of mobile typically —

Greg (15:33)
If that.

Mike B (15:35)
Right — that covers 99% of use. So I analyzed the first four screens and compared them to the United States. In the first four scroll screens — roughly a hundred percent of utilization — both the US and EU had 40% sponsored ads. Forty percent of the real estate was sponsored ads. The finder results were less in the EU than in the US, and third-party directories were less. The search headers were roughly the same. The reasons for those differences are the places sites and a great deal of “people also ask.” But there were virtually no organic results in Europe in the first four pages, and hardly any in the US either. One point I’d make: the hotel finder is really just another ad-delivery surface that, like the sponsored ads, largely helps booking in the hotel industry — in both the US and Europe, but particularly in Europe where booking.com is a total monopoly, so these ads are dominated by booking.com and its subsidiaries. Both of these are abhorrent results — hard to use. In the hotel finder, for example, you’ve got to dig to find a phone number; you can’t find it even when you click through. Clearly Google is trying to drive ads and doesn’t give a rat’s ass about consumer experience. So for them to say the consumer experience is the worst ever — it’s bad across the board, whether you look at the US or Europe. I see that as a passive-aggressive argument where they’re trying to marshal their troops to say the EU’s regulation is making things worse. To some extent you do get more third-party directory stuff in the EU — particularly between the places sites and the directory listings in organic — than you do here. But both results are largely pay-to-play.

Greg (18:14)
It’s either ads or Google’s own stuff, which in this case is ads. Pay to —

Mike B (18:19)
Or “people also ask,” or in Europe, Google’s own index results. It’s shocking to me that they give up space to their own index results. The argument Google is making is self-serving; it’s all crap.

Greg (18:35)
Sure. Well, they tend to make two arguments against any suggestion that they change anything. One is the consumer argument you just mentioned — it’s going to destroy the consumer experience, or it has, in this case. And the second is that we’re going to be disadvantaging small businesses: “We’re the champion of small business, the local pack is a direct pipeline for small businesses. If you make us elevate third parties, that’s all going to be directories that just do arbitrage and harm small businesses.” There’s some shred of truth to that, but it’s also a cynical, self-serving argument using these proxies. Meta is making a similar argument about being forced to show a chronological feed on Instagram. Adam Mosseri, who essentially runs Instagram and Threads, said — I think it’s Australia that’s compelling them to give people a choice to turn off the algorithm and just get a chronological feed — they’re making similar arguments: the consumer experience is going to be bad, people will be overwhelmed with crap. All these “parade of horribles,” as they say in the law, will happen if you force us to do this — which is originally how these social media sites started out. It’s kind of amazing.

All right, with that, let’s move on to Apple and Siri. Apple had its iPhone event, I think yesterday — it’s amazing how my sense of time is totally distorted.

Mike B (20:06)
Yes, it was yesterday.

Greg (20:08)
Apple had its iPhone event with John Ternus, the new CEO, introducing a couple of new iPhone models and some other stuff. And you captured and historically documented the growth of Siri usage — tell us about this.

Mike B (20:24)
They announced yesterday that Siri gets 2.5 billion interactions a day worldwide. I know they’d announced the number previously, so I went back, looked them up, and charted them over time. Of those 2.5 billion, from personal experience I can tell you 50% are false positives, because I’m trying to get the damn radio station to play —

Greg (20:46)
Right — get it to work.

Mike B (20:50)
Get it to work. But regardless — 2.5 billion. We’re at the beginning of what I see as likely an S-curve. It was somewhat linear for the first ten years, but since 2024 it’s taken off, because Apple has positioned it much more obviously in the interface with search and various other things. It demonstrates the power of distribution. One of the things they did yesterday at the introduction: they said, these are all the things AI can help you do — and they pointed out that all those things are what your phone does now. Your phone is this digital center of the universe, much like they positioned the computer 26 years ago under Steve Jobs. The phone is now the center, and all these peripherals around it are complementing it. So they’re positioning the phone as an AI-first device. With this kind of distribution, clearly, if Siri is half as good as people are saying — I know you’ve played with it a little —

Greg (22:05)
It’s better. It’s not great, but it’s better, for sure.

Mike B (22:07)
Right. We’ll see what the final release is. If you have a powerful enough phone — iPhone 17 or higher — you can experiment with the full range of it. But the distribution is what’s so striking to me. Their market share in phones has gone up in both the US and worldwide, even in China and India. And people who have iPhones use Siri — it’s the default, it’s there, it’s ever-ready. The only time you don’t is when you’re at home and have a Google speaker or an Amazon speaker.

Greg (22:41)
To some degree I’m insulated in the world of technophiles and tech enthusiasts, so I don’t know exactly what the broader view is. But the interesting thing is the growth — and this goes to your argument about distribution. The growth of Siri usage has come at a time when there’s a lot of criticism about its quality, especially relative to some of the AI assistants over the last two years. I mean, OpenAI emerged in 2022, and yet Siri has grown. Maybe Siri has benefited to some degree from its association with AI assistants. So what I’m trying to say is: if they get it right, this should accelerate further.

Mike B (23:28)
I agree with you a hundred percent. I see this as a starting point, and I see the numbers going up exponentially from here. I see people’s habits becoming “Siri” — or “Apple told me” — if it’s effective enough.

Greg (23:48)
They should be able to. With all their resources and talent — even though they’ve lost people to other companies in the AI category — they should get it right. Siri doesn’t have to do everything; it has to do a range of things well, consistently and reliably, and then usage will explode. If it does, it directly impacts Google, because a lot of the basic knowledge questions you’d ask your phone — weather or whatever — wouldn’t go to Google. Google is trying to improve the Gemini app and do all kinds of things to retain that usage. But to the point about distribution: I’d argue that but for Google’s monopolistic distribution across all these platforms — search, Android, Chrome, Google Workspace including Gmail — Gemini would not have the market share it has today. It’s exclusively this distribution thing, because the Gemini experience, although certain things are very well done, is mostly inferior to Claude and ChatGPT, in my opinion.

Mike B (24:59)
So yesterday, in addition to the Duo — which we’re going to talk about — they introduced the iPhone 18 Pro, which has a new microprocessor very focused on AI, on its capability of doing AI directly on the phone. They also introduced new watches with new chips, and new AirPods 5 with new chips, both of which have significant AI capacity. The watch, for example, in a disability-assistance mode can be turned on to record a conversation locally, and you can then look at your watch to see what was said. So if you’re having trouble understanding what’s going on, you can get a written reference to it in nearly real time, in a very private way — it doesn’t leave your phone or your watch, and it’s almost instantly available. It’s kind of like translation, but done through the watch. There are similar capabilities — improved audio understanding and Siri interactions — with the new AirPods, which are only $125. So as body-worn interfaces to AI, Apple is upping the ante — using the phone as a digital center, and these other body-worn devices as inputs to the AI, which I think is a very powerful combination. The long-haul question I have is whether the app is still the best way to deliver that information, or whether Apple is too dependent on the app, or whether AI will engender an appless environment at some point — although it’ll be years before we get there.

Greg (26:47)
Meta’s trying to do a version of that with its glasses, and OpenAI is supposed to be coming out with their device. To some degree, the argument Ternus was making — that the phone is still the best AI device — is an argument against these forthcoming devices from competitors.

Mike B (27:11)
Right, at least for the foreseeable future. That doesn’t mean interfaces can’t change or devices can’t be developed. But I have trouble seeing how ChatGPT — particularly given the lawsuit against them, having stolen all this information from Apple — is going to get a hardware device out the door that people accept. And with Facebook, there’s a huge trust issue we’ve talked about before.

Greg (27:32)
Right.

Mike B (27:33)
The question is whether Apple can retain that high level of trust they’ve built, given the increased use of AI and their increased use of ads. Just as an example of this trust: I was reading about LG TVs the other day — I think I shared that link with you — how even when they’re shut off, they’re sending audio conversations back to the cloud. They’re scanning your whole network, sending device types, and associating them with your profile for their advertising unit. One of the reasons I use Apple — I have an Apple TV box — is to isolate my TV for privacy reasons. If the TV is stealing that much data — every device on your Wi-Fi network — profiling it and associating it with you, LG’s making more money from ads than from TVs.

Greg (28:30)
That’s why TVs are cheap.

Mike B (28:32)
That’s why TVs are cheap, but it’s also why they spy on you. They spy on everything you watch, listen to everything you say, look at everything on your network. It’s incredibly invasive. And that’s why, so far, I’ve chosen Apple — because Apple doesn’t make its money from spying on you.

Greg (28:53)
To that point, I think Apple is the best. Microsoft is somewhere in there — the current incarnation of Microsoft is somewhat more ethical than the Bill Gates era. But Apple has carefully cultivated this reputation for privacy and consumer protection: “we’re not going to make you the product, we’re not going to exploit your data.” We talked a while ago about ads and whether that undermines the message — and Ternus, in one of his first messages, said everything they do with AI is private: it’s either on the device or in the private cloud, and they can’t access it. But I do think Apple has made a big mistake with the renaming of the Gulf of America, and then most recently Lake Ontario. I’ve seen a lot of people on social media really roasting Apple over the announcements, based on a diminished respect for the company in the wake of that. That’s not necessarily the mainstream consumer, but their brand has taken a bigger hit than I think they realize. South Park is making fun of them, as well as Google. So they have to be careful about not turning into just another tech company that’s going to surveil you and exploit your data — because if they lose that privileged position, even though their products are often better, losing the brand halo will show up in sales. So we’ll see.

Mike B (30:28)
And speaking of better products — you wanted to talk about the new Duo.

Greg (30:32)
Yeah. So the Duo is a very interesting product. Samsung has had a foldable phone, and I think there’s a Google-branded foldable too. There are a few foldables out there, and they’re expensive, and they get very mixed reviews. The Duo is a slightly wider phone — it fits in your hand, but it’s slightly wider than the normal iPhone — and then it opens to essentially a mini iPad. It looks very nice. I haven’t held one, but they’re getting good anecdotal reviews from people who’ve seen and held them.

Mike B (31:06)
And from people I trust — Horace Dediu, Carolina Milanesi, Darren Shaw. He hadn’t touched it, he just drooled on it. He wants to buy one.

Greg (31:12)
He wants to buy one, yeah.

Mike B (31:14)
It’s fascinating. When folded, it’s the size of a passport. The front screen is effectively a full-functioning smartphone, and it uses thumb or fingerprint ID rather than facial ID. All the aspect ratios have been changed to match the form factor, which none of the other folks did — so you end up with a lot of wasted space when you run traditional apps on a Samsung. Apple made APIs available so it’s fairly trivial: they’ve made all their apps compatible with the new aspect ratio, moved the menus to the side on both the front view and the landscape view, and made APIs available so every other program can do it fairly easily. It appears very elegant. When you open it, an image on the home screen automatically transfers to the inside home screen —

Greg (32:12)
It’s a transition, yeah.

Mike B (32:13)
A transition that’s very beautiful. The screen doesn’t have a crease in it, and the hinge —

Greg (32:19)
Doesn’t have a visible crease.

Mike B (32:22)
No visible crease, and the hinge is very well designed. So it’s a beautiful piece of equipment. It’s two thousand dollars. I don’t know —

Greg (32:29)
That’s the problem for a lot of people, but many people will spend that.

Mike B (32:34)
For a road warrior who has an iPad, a phone, and a computer, it may suffice. And in about six months — in iOS 27.2, I think they said — you’ll be able to use the pencil with it as well.

Greg (32:45)
Yeah, you can also use the pencil. I could certainly see business customers getting this because of the dual functionality — the phone and the iPad-mini-style screen. It’ll be interesting to see how many people buy it. Prices have come up because of tariffs and other things — everything is more expensive — which represents an opportunity for Apple to increase its average revenue, but it’s also a deterrent for many people who don’t want to lay out that kind of money. We’ll see.

Mike B (33:17)
To give the opposite side equal time: David Mihm put up a puking emoji when he saw the announcement. He’s obviously not going to buy one. But it’s one of those things — I’d bet David a beer that it’ll be successful, well received, and useful to the crowd that needs it. I’m not in that crowd; I don’t travel that much anymore, and I don’t have an iPad.

Greg (33:41)
It makes the entertainment experience on the phone better because you have a bigger screen, so some people will be attracted to it for that. “I’m watching this movie on my tiny iPhone screen, and now I can see it on a bigger screen.”

Mike B (33:55)
Right. They showed a video of somebody watching a YouTube video on the upper part of the screen with email in the lower part. My setup is: I watch the video on the TV and hold the phone in my hand, so I have that same effective outcome. Big TV, little phone — I don’t need a Duo to do that, as my wife pointed out to me.

Greg (34:16)
I think this was a product long foreshadowed by Mark Gurman at Bloomberg and others — everybody knew it was coming. The surprise is that it’s as nicely done as it is, and it addresses many of the complaints about these phones. To say Apple’s been coasting isn’t entirely fair, but they kind of have been — making incremental improvements and relying on a product set largely established under Steve Jobs. I don’t know if anything is totally new under Tim Cook. Tim Cook obviously took the company to a five-trillion-dollar valuation, whatever it is today — he’s done very well for investors, and consumers have continued to buy — but there’s a sense that things are a little stale. They’re going to come out with these glasses, and with the negative halo from Meta and the accusations of surveillance — “these are pervert glasses” — that may doom the product, or really diminish demand. But they need to come up with some new stuff, genuinely innovative new products they can bring to market. Otherwise, at a certain point, there’s a lack of —

Mike B (35:38)
My sense of this — and I follow it fairly closely; I could be accused of being an Apple fanboy, though I think my reasons are legitimate: privacy and build quality — is that their build schedule was predicated on more rapid development of their AI skills. Their AI capacity was delayed; they went down a couple of wrong paths.

Greg (36:02)
And they lost a bunch of key people, which undoubtedly affected them.

Mike B (36:06)
But I do believe there’s a whole range — they didn’t announce an iPhone 18 yesterday, they announced the iPhone 18 Pro. And the iPhone 17 has been one of their absolute best-selling phones ever, garnering increases in market share while everybody else is losing it. I believe a large number of products were predicated on AI as an integral part, and because of the delays in AI, these other, mostly home-based products were delayed. I think you’ll see over the next twelve months a ramping of product releases — products that are largely done and were waiting on decent AI, which is now theoretically in place. Whether those products are any good, I haven’t seen them, but there’s apparently a whole range of home and minor robotic products — kitchen-based products, and so on.

Greg (36:59)
Apple does a good job on the products it releases — they tend to be better than competing products. The whole smart-screen thing was a big deal and then sort of isn’t anymore, but that’s an area of opportunity for Apple, and I think some version of that is in their queue.

Mike B (37:25)
On a related note: I traveled last weekend to Chicago, and one of the things Apple hasn’t done yet — which hopefully will happen in iOS 27 — is bringing AI into Maps. I used Google Maps regularly during my excursion to do these complicated searches using Ask Maps. Regular maps, who needs it? Ask Maps let me plan a half day with another couple in Chicago — a walk along the river walk, a good place to find a hot Italian beef sandwich (which, by the way, is delicious; I’d never had one before), then the museum, and back to our apartment before the wedding started. It helped me plan that complicated day. And driving back — it’s a twelve-hour drive, so we wanted to split it into two — I said, find me a quaint town on the lake with a good meal, and so on. It pointed me to Erie Bluffs on Lake Erie. It was spectacular. I never would have found these places otherwise.

Greg (38:37)
Hands down, in my opinion, Ask Maps is the best thing Google’s doing in AI.

Mike B (38:43)
I’d like a mapping product from Google that just has that — none of the other stuff. Clear the interface out and give me Ask Maps. Hopefully Apple will do something similar. We’ll see.

Greg (38:51)
They’re supposed to bring AI into Maps, and we’ll see what that’s like. Google’s advantage, as we’ve discussed many times, is its data set — it has a better local data set than anybody. You could argue shopping is the same, with their merchant and product database, but local is hands down their single competitive advantage. They have brand trust, which showed up in the slide I showed at the beginning, where search still retains the pole position in each of these customer-journey phases. But it’s really local that is the anchor now for Google. Google search — who cares? What you showed is evidence that it’s just polluted by ads and Google modules and crap. We’ve talked about this before, and you like to rip me that I still use Gmail and Chrome periodically. But in terms of search, I don’t use — yes, yes, all of that — but search —

Mike B (39:54)
And YouTube, and Office, and Docs and Sheets, and so on. We all do.

Greg (40:03)
That’s the monopoly thing. But search is the bread-and-butter product, and I’ve almost completely eliminated that usage, with the exception of Ask Maps. Claude, ChatGPT, Apple, Siri — all these other things have replaced it for me. Very occasionally I’ll use it. The interesting thing is, once you break out of that habitual behavior, the mystique is completely gone. It’s over. Google doesn’t offer better outcomes or better results than I can get from ChatGPT or Apple, for the most part. In some cases with local it does, but I’m like a heroin addict who’s now broken the habit. That’s what Google relies on — habitual behavior. It’s not a better product in most cases; in some it is, but general search is not a better product. It’s the default, good-enough product because it has more — yeah. Anyway.

Mike B (41:07)
On that note, we should find something positive to say and wrap it.

Greg (41:13)
Yes. I could go on more rants. The positive thing is that Brighton SEO is coming up next week, and David and I will both be there. Hopefully we’ll see some of you there, and we appreciate all of you.

Mike B (41:24)
And if you’d like to be on the show, give Greg a topic while you’re there and pitch yourself. We love to have guests who are exploring interesting topics.

Greg (41:35)
Who are interesting and have interesting things to say. So thank you very much for your loyalty and listening, and as always, we’ll see you next week.

Similar Posts

Leave a Reply