Original research · 73 studies

Original research, published in the open.

We measure machine readiness at scale, on both sides of the market. On the supply side, we audit how ready real businesses are to be read by the systems that now find and recommend them: six economies, 205,103 businesses, one method. On the demand side, we measure how far buyers have handed those decisions to the machines. The retail and market studies beneath the two capstones trace the same shift, category by category. Every figure is dated, attributed, and tiered for evidence.

First-party audits are measured directly, not modeled. The market studies are built on dated, attributed sources, with every claim tiered established, emerging, or contested. The AI-answer layer has no independent census yet, so readings of it are named as a frontier, never presented as settled.

The Attention Migration Studies

How the market is moving, and what it does to where attention sits

60 studies across 10 series, each answering the same three questions: how the market is evolving in the US and globally, what that does to how buyers discover and choose, and how it redraws the attention terrain. Grounded in dated, attributed evidence.

01

Primary Audits

First-party machine-readiness audits of real small businesses, measured economy by economy.

6

Machine-Readiness Audit: Africa, July 2026

The first Africa edition of a quarterly, multi-region benchmark. We sampled 11,582 independent African small businesses from the map. 41.4% had a real website; the rest reach customers through social and messaging, not a machine-readable page. About 4.5% of all map-listed African businesses are fully machine-readable.

41.4% of 11,582 map-listed independent African businesses had a real website (4,793) Read the study · 20 min

Machine-Readiness Audit: Europe, July 2026

The first Europe edition of a quarterly, multi-region benchmark. We sampled 52,076 independent European small businesses from the map. 76.3% had a real website, and among those, 77.9% had no type an engine can read as a local business. About 12.7% of all map-listed European businesses are fully machine-readable.

76.3% of 52,076 map-listed independent European businesses had a real website (39,722) Read the study · 20 min

Machine-Readiness Audit: India, July 2026

The truer picture of the Indian market. We measured 110,312 independent Indian businesses in all, the majority across the twenty-two trades that actually make up the high street, kirana grocers, tailors, and more. Only 18.4% of those had a real website, and just 1.2% are fully machine-readable, far below what the Western-professional trades suggest.

18.4% of 66,452 map-listed independent Indian businesses had a real website (12,227) Read the study · 20 min

Machine-Readiness Audit: Japan, July 2026

The truer picture of the Japanese market. We measured 60,059 independent Japanese businesses in all, the majority across the eighteen trades that actually make up the high street, izakaya and ramen counters, soba and udon shops, and more. Only 43.4% of those had a real website, and just 2.3% are fully machine-readable, far below what the Western-professional trades suggest.

43.4% of 34,831 map-listed independent Japanese businesses had a real website (15,107) Read the study · 20 min

Machine-Readiness Audit: Southeast Asia, July 2026

The truer picture of the Southeast Asian market. We measured 55,020 independent Southeast Asian businesses in all, the majority across the eighteen trades that actually make up the high street, warung and warteg eateries, sari-sari and toko kelontong grocers, and more. Only 11.4% of those had a real website, and just 0.9% are fully machine-readable, far below what the Western-professional trades suggest.

11.4% of 33,664 map-listed independent Southeast Asian businesses had a real website (3,840) Read the study · 20 min

Machine-Readiness Audit: US MSMEs, July 2026

The first edition of a quarterly benchmark. We measured 43,362 independent US small businesses; one in four was invisible to a machine reader, a majority had no type an engine can read as a local business, and readiness rose steadily with how established the business was. We will repeat it every quarter, using the same method, to track how the machine-readable web evolves.

43,362 independent small-business homepages read, drawn from Google Business Profile listings across 12 trades and 88 US metros Read the study · 24 min
02

The Macro Shift

How discovery itself is changing, in the US and worldwide.

6

Agentic Commerce in 2026: The Rails Are Built, the Trust Is Not

Every major payment network and AI platform has now shipped a way for agents to buy on a shopper's behalf, but the verified data shows trust collapses at the exact moment an agent reaches for a card.

48% of online shoppers used AI to research their most recent purchase Read the study · 12 min

The AI-Answer Adoption Curve

Buyers picked up ChatGPT, Gemini, and Copilot faster than they picked up the PC or the internet, but the data shows a wide, still-unresolved gap between how many people open the tool and how many actually let it decide.

44% of US adults report using ChatGPT in 2026, up from 18% in 2023 Read the study · 13 min

The Global Divergence

AI-answer reliance is splitting the world in two: flat in the US and major Western European markets, accelerating across parts of Asia, Southern Europe, and Latin America, and still invisible to any independent, non-vendor measurement.

15% vs. 8% Click rate on a traditional search result, no AI summary vs. AI summary present Read the study · 10 min

The Great Unbundling

Since Google's AI Overviews launched in May 2024, a rising share of search sessions now end inside a synthesized answer instead of a click to your website, and the strongest data available says that shift is real, accelerating, and still only partly measured.

2 billion monthly users, 200+ countries Reach of Google's AI Overviews as of Q2 2025, up from 1.5 billion in May 2025 Read the study · 15 min

The MSME Visibility Gap: Why AI Search May Be Widening the Small Business Divide

Search usage is falling as AI answers rise, and the early evidence says structured, corroborated data, not brand size alone, decides who gets cited, which is a harder bar for small businesses to clear than for enterprises.

8% vs 15% Click-through rate to a web link when Google shows an AI summary vs. when it does not Read the study · 11 min

Zero-Click and the Vanishing Referral

Independent measurements disagree on how much smaller the number is, but they all agree it is smaller: the click from search to your website is disappearing, and nobody has yet measured what that costs a single local business.

8% vs 15% Click-through rate to a traditional search result when an AI summary appears vs. when it does not (tracked browsing, 900 US adults, March 2025) Read the study · 13 min
03

The Human-Capital Shift

How AI is reshaping jobs, degrees, and the classroom, from the value of a credential to the entry-level ladder.

5

Skills Over Degrees: The Hiring Shift, Measured Against Reality

Employers are stripping degree language from job postings and states are dropping degree mandates outright. But the best independent outcome study found that removing the requirement changed hiring for fewer than 1 in 700 new hires.

51% → 44% share of US job postings requiring a bachelor's degree, 2017 to 2024; 63% of the underlying occupational resets happened before the pandemic Read the study · 18 min

The Broken Bottom Rung: The Entry-Level Job Market Under AI

Recent college graduates in the United States now face a higher unemployment rate than the country as a whole, the first time that has happened in roughly 45 years of data, while entry-level postings fall and senior postings climb.

5.6-5.7% unemployment rate for recent college graduates aged 22-27 (late 2025 to mid-2026), above the national rate of about 4.2% for the first time in roughly 45 years of data Read the study · 18 min

The Classroom Meets the Machine: AI and the University

Undergraduate AI use for assessed work climbed from 53% to 94% in two years, take-home essays lost their evidentiary value, and computer-science enrollment fell for the first time in a decade, reversing a decade of "learn to code" advice.

94% of UK undergraduates used AI for assessed work in 2026, up from 53% in 2024 and 88% in 2025 Read the study · 18 min

The Credential Boom: Certificates, Bootcamps, and the Reskilling Economy

Coursera alone reached 197 million learners and $757.5 million in revenue in 2025, but a machine-learning read of 37.7 million resumes finds only a job-relevant certificate actually raises wages. A pile of unrelated badges can pay nothing, or less than nothing.

197 m / $757.5 m Coursera's registered learners and full-year revenue in 2025 (revenue up 9%, growth decelerating from prior years) Read the study · 18 min

The Degree in Question: What a College Credential Is Worth in the AI Era

The average college wage premium sits near a record 80 percent and is still rising, and at the same moment 28 percent of bachelor's programs run a negative financial return. The degree is winning on average and losing at the margins, and generative AI is biting hardest at those margins.

~80% college wage premium (bachelor's-or-higher vs. high-school-only), 2024, and still rising: college earnings grew 6.3% vs. 3.2% for high-school-only, 2004 to 2024 Read the study · 18 min
05

Buyer Behavior Science

How people decide now, from the shortlist to the click.

6

Reputation Physics

Buyers now demand fresher, more numerous, and imperfectly rated reviews than they did a year ago, and the businesses AI engines choose to recommend look different from the ones winning on search rank alone.

6% to 45% Share of US consumers using AI tools like ChatGPT to find local business recommendations, in one year Read the study · 9 min

The Consideration-Set Collapse

Buyers were already comparing a handful of options before any engine got involved. What a single synthesized answer removes is the visible list itself, and no one has yet measured what replaces it.

7, plus or minus 2 Human working-memory capacity for holding discrete items in mind at once, the cognitive ceiling behind why shortlists were always short Read the study · 9 min

The Credence-Good Problem in an AI World

When a buyer can't judge your work until it's too late, AI answers aren't closing that old trust gap, they're moving it, and the habit of checking a second source is fading right when it matters most.

49% of US adults now use AI chatbots, up from 33% a year earlier; ChatGPT use specifically rose from 18% to 44% Read the study · 11 min

The Local-Intent Shift

Near-me searches no longer settle on one ranked page: they scatter across classic search, AI answers, and reviews, and the evidence shows that winning one surface is no guarantee you win the others.

8% vs 15% Click-through rate on a traditional search result when an AI summary appears, versus when it does not Read the study · 12 min

The Speed-to-Answer Standard: Four Decades of Delay Data and the AI-Answer Frontier Nobody Has Measured Yet

Every documented increment of delay carries a measurable cost, and the newest instant-answer layer, generative AI, is now the fastest-growing part of that story with the least independent measurement behind it.

0.1s / 1.0s / 10s Nielsen's three response-time thresholds: instant, flow-preserving, and attention-holding Read the study · 13 min

Trust in the Machine

Chatbot use jumped from a third of US adults to about half in two years, but the moment a synthesized answer appears on a results page, buyers click and verify less, not more.

49% (up from 33% in 2024) Share of US adults who use AI chatbots Read the study · 14 min
07

Vertical Market Evolution

How each kind of business gets found and chosen, on its own terms.

19

Accounting and Tax: The Credence Gap AI Hasn't Closed

Tax and accounting clients cannot verify the work they are buying, so they lean on reputation and credentials instead of price, and even fast-growing AI adoption for filing has not closed that trust gap.

~2.3% Share of eligible taxpayers who registered for the IRS's free Direct File pilot in the 2025 season; most of those who did register never actually filed through it Read the study · 12 min

Fitness and Wellness: How Members Actually Find a Gym in a Record Year

US gym and studio membership just hit a record high, but the decision that gets someone through the door still runs through star ratings, short clips, and word of mouth, with the AI-answer layer still too new for anyone to measure.

81 million members, 26.1% penetration US gym, studio, and fitness-facility members in 2025, a record and a quarter of the population age 6+ Read the study · 9 min

Home Services and Contractors: Market Evolution

Homeowners still call a local number to book a job, but the road that gets them there now runs through AI-produced summaries and answer engines long before anyone compares a bid.

68.01% (up from 60.45% in 2024) US Google searches that end with no click at all, Jan-Apr 2026 Read the study · 13 min

Hospitality and Lodging: Where Guest Discovery Is Headed

OTAs still take a growing share of independent-hotel bookings, but the bigger shift is in where travelers start looking, and no one has mapped that terrain for hospitality yet.

63.4% Share of independent-hotel bookings that went through an OTA globally in 2025, up from 61.3% in 2024 Read the study · 11 min

How Diners Actually Decide: Reviews Still Pay, AI Answers Remain Unproven

The restaurant-discovery stack is splitting five ways, but only one channel, the star review, has a proven link to revenue; everything after it, including AI-mediated answers, is still an open bet.

5 to 9% revenue increase Causal effect of a one-star Yelp rating increase on independent-restaurant revenue Read the study · 9 min

How People Find a Lawyer in 2026: Search Is Slipping, AI Answers Are Rising, and the Middle Is Still Unmapped

Google's share of attorney research dropped nearly 15 points in a single year and willingness to ask an AI chatbot the same question has climbed steadily since 2023, but no verified data yet shows how often a legal question actually lands inside the answer an AI engine gives.

79% (2025), up from 57% three years earlier Legal clients who research attorneys online before making contact Read the study · 11 min

Med-Spa and Aesthetics: Market Evolution

Patients are moving to search, short video, and AI chat faster than the aesthetics industry is moving credentials onto those same surfaces, and 2026's prosecutions show what that gap costs.

$17B+ US med spa industry annual revenue, confirmed live on the trade association's own statistics page Read the study · 14 min

Medical and Healthcare: How Patients Discover and Choose Providers Now

AI chatbots have caught up to social media as a health information source, but Google keeps them out of "near me" searches, so the local pack and online reviews still decide which provider gets the call.

32% of US adults used an AI chatbot for health information in the past year, now roughly matching social media use for the same purpose Read the study · 11 min

Record Reliance, New Rules: How Buyers Find a Real Estate Agent Now

Home buyers and sellers use an agent at record or near-record rates, but the settlement that pulled commissions off listings and the fast rise of AI chatbots as a discovery layer have redrawn how they find, vet, and choose one.

88% of home buyers purchased through a real estate agent or broker (July 2024 to June 2025) Read the study · 9 min

The Agentic Shopper Premium: Measuring the AI-Order Effect in Ecommerce

Shoppers who arrive at your store from an AI answer convert better and spend more than shoppers who arrive from a search result, across every measurement stack that has looked, but the channel is still small and every number so far comes from a vendor grading its own traffic.

~13x YoY order growth, +14% AOV AI-referred order growth and average-order-value lift on Shopify's merchant network, Q1 2026 Read the study · 11 min

The Comparison Paradox: Insurance Shoppers and Recommendation Behavior in the Answer Engine

Insurance shoppers are comparing more than ever and turning to AI to help, but the AI engines they consult often disagree with each other and, on at least one measured question, get the answer wrong more often than right.

3.5 quotes Average number of quotes pulled by US auto insurance shoppers, the highest recorded in J.D. Power's history Read the study · 11 min

The Five-to-One Rule: How AI Research Now Outweighs Sales Contact in Professional-Services Selection

Buyers now spend far more of their time on independent, AI-assisted research than in conversation with any single vendor's sales rep, and agencies and service providers top the list of what they research with AI.

27% of buying time on independent research vs. roughly 5 to 6% with any one vendor's sales rep Independent research vs. time with a single vendor rep Read the study · 12 min

The Local Retail Discovery Split: Map Pack, Marketplace, and the Rise of the AI Answer

Most local retail dollars still cross a physical counter, but the search that decides which counter is increasingly settled by an AI system before a human ever sees a list of choices.

16.9% Share of total US retail sales that happened online in Q1 2026 ($326.7B of $1,929.0B), with e-commerce growing 9.8% year over year against 3.9% for total retail Read the study · 11 min

The Offline Resolve: Why Chiropractic and Physical-Therapy Patients Research Like Buyers but Book Like Patients

Patients now research chiropractic and physical-therapy care like a considered purchase, but the evidence shows they still book it like a phone call to someone they trust.

84% Patients who check online reviews before choosing a healthcare provider Read the study · 10 min

The Parent Funnel Fork: How AI Homework Tools Are Splitting Tutoring Demand

Teens are adopting AI chatbots for homework faster than their parents notice, while the paid, human side of tutoring keeps growing right alongside it.

54% US teens (ages 13-17) who have used an AI chatbot for schoolwork help Read the study · 12 min

The Six-Screen Search for a Dentist

Patients now spread their trust across an average of six discovery sources, including AI answers, while most dental practices still can't see what those sources are saying about them.

$189 billion U.S. national dental expenditure in 2024, up 4% year over year after inflation, driven mainly by a 9% rise in government dental program spending Read the study · 11 min

The Verification Funnel: How AI Search Is Rewiring Auto Repair Discovery

AI answer engines cite businesses far less often than Google's map pack surfaces them, in every category researchers have measured so far. No one has measured the gap for auto repair yet, but the category's math points the same direction.

68.01% of US Google searches ended without a click in the first four months of 2026, up from 60.45% in 2024 Read the study · 11 min

Validation, Not Discovery: How Beauty and Salon Bookings Really Move Through Search and Social

The "78% of bookings trace to local search" statistic you've probably seen has no real source behind it, but the verified evidence still points somewhere useful: toward reviews and local search as the checkpoint nearly every booking passes through, no matter where discovery started.

97% of consumers read online reviews for local businesses Read the study · 9 min

Veterinary and Pet Care: Market Evolution

Pet spending keeps climbing while what the average household spends on its own dog or cat has already softened once, and a growing, still-uncounted share of the clinics collecting that spending answer to owners whose names never appear on the sign.

$41.0B → $42.4B (2026 proj.) US veterinary care and product sales, 2024 actual to 2026 projected, within a total pet-industry spend of $158B to $165B Read the study · 11 min
09

Global & Regional Commerce

How discovery works in the markets that do not run on Google.

6

China's Parallel Answer Engines: Why DeepSeek, Doubao, and Baidu ERNIE Are a Separate Visibility Game

China's generative-AI users now number in the hundreds of millions, running on apps, under rules, and drawing on sources that the Western AI-visibility playbook was never built for.

602 million generative-AI users in China by end of December 2025, up 141.7% year over year (42.8% national adoption) Read the study · 11 min

Sold in the DM: What the Evidence Actually Shows About Latin America's Conversational Commerce Edge

The "45 to 60 percent, twelve times a normal store" conversion claim behind Latin America's conversational commerce hype cannot be traced to any real study, but the verified numbers underneath it, a multibillion-dollar WhatsApp business and the world's fastest-growing ecommerce region, tell a real story on their own.

$2B+ WhatsApp business-messaging annual run rate, Q4 2025, up 54% year over year across the wider Family of Apps other revenue line Read the study · 12 min

Sovereign Search: The Markets Where a Local Engine, Not Google, Answers First

In Russia and South Korea, a home-grown search engine holds more of the market than Google does, and the timing of when each one shipped its own AI answer looks like part of the reason, though it is not the whole story.

71.03% Yandex's share of Russian search, versus Google's 26.53% Read the study · 10 min

The Data-Cost Ceiling: How Mobile Pricing Splits Africa's Attention Map

Mobile data pricing does not gate the whole continent evenly: it draws a hard line between a small set of extreme-cost, text-and-voice-locked markets and a much larger cheap-data majority where video-first platforms are already winning.

$43.75/GB Zimbabwe's mobile data price, the single most expensive market on Earth Read the study · 10 min

The Livestream Loop: How Southeast Asia Collapsed Discovery, Consideration, and Checkout Into One Screen

TikTok Shop's Southeast Asia sales roughly doubled to $45.6 billion in 2025, and a government forced its discovery layer and its checkout layer into two separate companies: the clearest sign yet that browsing and buying have become one screen, even though the exact size of the conversion advantage is still not settled.

$45.6B TikTok Shop Southeast Asia GMV in 2025, roughly double the prior year Read the study · 11 min

The WhatsApp Storefront: How India's MSMEs Are Discovered Without a Search Bar

In India, chat has quietly become the discovery engine: a majority of online adults message businesses every week, and the surface where that happens sits almost entirely outside the web an AI engine can read.

91% of online adults in India chat with a business weekly Read the study · 11 min

The Retail Visibility Study

Who the machine names when a shopper looks to buy

Ten primary-research reports across retail categories and five US metros, measuring who local search, AI Overviews, and the standalone chat engines actually name, and where independent local shops stand.

The headline, in one chart

Independents own the local pack, in every category

Apparel boutiques
100%
Florists
100%
Garden centers and plant nurseries
99%
Independent bookstores
97%
Jewelry
93%
Bike shops
92%
Furniture
90%
Toys and games
87%
Hardware stores
82%
Pet supplies
73%
Independent local businesses as a share of the Google map three-pack, by category. Captured 2026-07-22 across five US metros. The contest is not the pack; it is the organic layer and the answer box above and below it.

What the ten studies found together

Four things that held across every category

01

Independents own the local pack

Across every category we measured, independent local shops hold the great majority of the Google map three-pack, from roughly three quarters to a clean sweep. On local intent, the big-box and marketplace fear does not hold.

02

But they lose the layer above it

In the organic results beneath the pack, a directory and marketplace tax takes over. Yelp, Reddit, and city listicles outrank the shops themselves, and the independent share falls sharply. The pack is won; the page is not.

03

The AI answer splits by engine

Google rarely serves an AI Overview for these local queries, and where it does it cites directories. The standalone chat engines behave differently: ChatGPT, Perplexity, and Gemini all name real local independents, but only ChatGPT cites the shops’ own sites, and Gemini masks its sources entirely.

04

Demand is seasonal, so timing is a lever

Every category has a real demand curve, from a spring peak for nurseries to December for toys and jewelry. Visibility has to be earned before the season, not during it.

Choose a category

The ten studies

try-on, experiential, style and fit drivenApparel boutiques

When a Shopper Looks for a Boutique, Who Does the Machine Name? Apparel Boutique Visibility Across Five US Metros

100%of the Google local pack is independent boutiques
Read the study
Specialty durable with recurring service attachBike shops

Who the Machine Names for the Local Bike Shop: Share of Answer, the Map Pack, and the Service Moat

92%of Google local map pack slots were independent shops
Read the study
time-critical, same-day, occasion-drivenFlorists

Who the Machine Names When Flowers Cannot Wait: Order Gatherers, Local Florists, and the Time-Critical Visibility Terrain

100%of the Google local pack was independent local florists
Read the study
Durable, showroom, high-considerationFurniture

Who the Machine Names When a Shopper Looks for a Furniture Store Near Them

90.1%of the Google local pack is independent shops
Read the study
Seasonal, perishable, local-onlyGarden centers and plant nurseries

When the Machine Recommends a Nursery: Local Visibility for Garden Centers and Plant Nurseries

98.9%of local-pack slots went to independent nurseries, not chains (88 of 89)
Read the study
Urgent replenishment and project buyingHardware stores

The Aisle the Machine Skips: Independent Hardware Store Visibility Across Search and AI Answers

81.8%of Google local-pack slots go to independents and co-op stores
Read the study
Specialty, discovery and serendipity, community-anchoredIndependent bookstores

Who the Machine Names When a Reader Looks for a Bookstore

97%of the Google local pack was independent bookstores
Read the study
High-ticket, occasion-driven, trust-led (a credence good)Jewelry

Who the Machine Names When a Shopper Looks for a Jeweler: Independent Visibility Across Five US Metros

93.3%of the Google map pack went to independent local jewelers
Read the study
Recurring replenishment, subscription-pronePet supplies

The Replenishment Blind Spot: Who the Machine Names When Shoppers Look for a Local Pet Store

73%of Google local pack slots went to independent local shops
Read the study
gift-driven, seasonal, occasionToys and games

Who the Machine Names When a Shopper Wants a Toy Store: Local Visibility in a Gift-Driven, Seasonal Category

87%of the Google local pack is independent shops
Read the study

How each study is built

Four evidence layers, on every report

01

Share of answer

A dated read of who each surface names, classified independent, chain, marketplace, or directory, across five metros.

02

Intra-day stability

A k=5 recapture testing how much each surface reshuffles within a single day.

03

Demand history

Twelve months of real Google search volume, quarter by quarter, with the seasonal peak and trough.

04

The AI answer engines

ChatGPT, Perplexity, and Gemini captured with web search, read for who they name and where they cite from.

The program

Measured on a quarterly cadence

The audits repeat each quarter and the market studies extend as the shift does, so every figure here reflects where the market stands now. The instruments that produced these numbers read a single business as readily as a whole economy. For your own market and your own site, that read is a free Machine-Readiness Score.

Want this read for your own market?

A specialist-reviewed read of how machine-ready you are across search and AI answers, for your business and your market. No obligation to continue.