The State of the Small Business Surface

The seven shifts

Search is becoming answers

AI summaries now sit on roughly a fifth of US Google searches, and they keep the click. The list of blue doors that local websites were built for is quietly being replaced by a single spoken-for answer.68% of US Google searches ended without any click in early 2026.

The profile is the new front door

Customers meet the Google profile, the map card, and the photo strip before any homepage. In one small-market field read, a keyword tool reported zero for several exact local phrases; the profile still carries a visible first impression.0 estimated monthly searches reported by the tool for “tattoo shop” + town across a three-town BC corridor.

Assistants join the buying journey

ChatGPT crossed 800 million weekly users, and consumers have started asking it for plumbers, groomers, and studios. A business now needs to be legible to a reader that never sees its design.45% used AI tools for local business recommendations in the past year, up from 6%.

Reviews became infrastructure

Nearly every consumer reads reviews for local businesses, and the bar keeps rising: the share who require four stars or better jumped thirteen points in a year. Ratings now work like building codes.68% will only use a business rated four stars or higher, up from 55% a year earlier.

Silence now reads as an answer

Nine in ten consumers expect owners to respond to reviews, most within the week. An unanswered complaint is no longer neutral; it is read as the business’s official position.89% expect owners to respond to reviews; 81% expect it within a week.

The owner is the marketing department

Half of small businesses have nobody dedicated to marketing, and most owners give it under five hours a week on budgets below a thousand dollars a month. Every shift above lands on that desk.50% of small businesses have no marketing staff; 70% of owners spend under 5 hours a week.

The gap between intent and surface

Put the six shifts together and a structural gap opens: what the owner means the business to be, and what profiles, reviews, and answer engines actually show. The gap is wide, and invisible from inside.7 leaks surfaced when byImprint audited its own public surface before selling the method.

Three forces remaking how small businesses get found

These are loud years for quiet businesses. The dog walker, the tattoo studio, the cleaning company, and the wedding florist have not changed what they do. What changed is the machinery between them and their next customer. This briefing reads that machinery the way byImprint reads a single business: from the outside, with sources attached, and with a bias toward what an owner can fix on a Monday.

The first force is the answer engine. Search spent twenty years as a list of doors; it is becoming a concierge that answers directly. By March 2025, roughly one in five US Google searches produced an AI-written summary, and when one appeared, users clicked a result on only 8 percent of those searches, against 15 percent without one. Across early 2026, 68 percent of US Google searches ended without any click. The shelf where a small website used to be found is shrinking.

The second force is quantified trust. Reviews stopped being reading material and became infrastructure: 97 percent of consumers read them when researching local businesses, and the threshold is climbing fast. In successive, independent consumer surveys, the share who said they would only use a business rated four stars or better rose from 55 to 68 percent. Recency rules: three quarters seek reviews written within the last three months.

The third force is the capacity squeeze. Half of small businesses have no one dedicated to marketing; the owner is the department. Seventy percent of owners give marketing less than five hours a week, and only 18 percent of surveyed small-business decision-makers globally say they are very confident their marketing is effective. The work this briefing describes lands on the busiest desk in the building.

What follows. Seven shifts across three parts, each argued from named, dated sources. A fourth part on the week behind the surface: what the owner’s desk already carries, and why so little of it has changed. Then a fifth part on the work itself: what byImprint does about all of this, and what it refuses to promise.

The premise of this briefing

Every small business now has two storefronts: the door on the street and the surface online. Customers reach the second one first, and the owner almost never sees it.

Search is becoming answers

AI summaries moved from experiment to furniture in under two years. By March 2025 they appeared on roughly one in five US Google searches, and tracking across 2025 saw their share of queries swing as high as one in four. When a summary is present, behavior changes sharply: users click any conventional result far less often, and they end the session on the results page far more often.

For a local business this is not an abstract publishing problem. The summary is assembled from whatever the public surface supports: the profile, the reviews, the site’s extractable facts. A weak surface does not just rank lower now; it gets paraphrased thinner.

Local businesses lived on the click: page two was invisible, page one paid the rent. The new economics are harsher. Even a first-position result loses most of its clickthrough when an answer sits above it.

Feed the answer instead of fighting it. The summaries are assembled from what the public record supports: clear extractable facts, consistent naming, plain service descriptions, and live reviews. A surface the answer engine can quote is the closest thing to shelf space the new results page offers.

The profile is the new front door

Checking on local businesses is now a weekly habit for 80 percent of consumers, and the place they do it has moved. The Google profile quietly absorbed the functions a website used to own: hours, prices, photos, proof, booking, directions, questions. Customers decide inside that interface, which means a business’s most-trafficked surface is one the owner may not have opened since the day it was claimed.

In one small-market field read in British Columbia, a keyword tool reported zero estimated monthly searches for a handful of exact trade-plus-town phrases. That result is a caution against treating keyword estimates as the whole market. It does not prove that nobody typed those queries or establish where any unmeasured discovery occurred.

Owners budget for websites and post on social, while the surface that meets the most new customers, the profile, goes unowned. It updates itself: reviews arrive, photos age, hours drift wrong after a holiday, and nobody is assigned to notice.

Treat the profile as the front door: read it monthly as a first-time visitor, keep photos current, answer what arrives, and make hours and booking the most accurate facts the business publishes anywhere.

Assistants join the buying journey

Assistant use crossed from novelty to habit in roughly a year. The instinct to “ask” rather than “search” is spreading into local decisions: recommendations for trades, studios, and services that used to start in the search bar. The assistant answers from the same public record this briefing keeps pointing at: profiles, reviews, structured facts, readable pages.

Two properties make this reader different. It compresses: a business becomes two sentences. And it cannot be charmed: photography and layout do not survive the trip, only facts and reputation do. What it can extract is what it will say.

No one can guarantee placement inside an AI answer, and this document will not pretend otherwise. What can be controlled is the record the answer is drawn from.

1. Consistent name, place, and trade across every public mention
2. Plain-text answers to the questions buyers ask: services, prices, area served, how to book
3. Review text and ratings, which carry both reputation and vocabulary
4. Structured data that machines parse without guessing

Reviews became infrastructure

Reading reviews is now near-universal behavior, and the casual reader is becoming a systematic one: the share who always read reviews before choosing a local business rose by twelve points in a single year. Half of consumers extend reviews the same trust they give personal recommendations.

The threshold is moving with the habit. Four stars used to be comfortable; it is becoming the floor. And the ledger is priced: the classic Harvard study of restaurant revenue puts a single star of rating at 5 to 9 percent of revenue. The number on the map card is a line item.

A 4.3 average with fresh reviews now outperforms a 4.8 earned three years ago and left to age. Recency weighting means reputation is a flow, not a stock: it must keep arriving to keep counting.

Build the ask into the work itself: the follow-up message after the job is the single highest-yield piece of marketing copy most local businesses will ever write. Aim for a steady trickle, not a campaign burst, and never buy what must be earned. Fabricated reviews are the one shortcut consumers actively punish.

Silence now reads as an answer

Expectations crossed a line: responding to reviews moved from courtesy to baseline service, with most consumers expecting an answer inside a week. The audience for each reply is not the reviewer; it is every future customer scrolling past, watching how the business behaves when praised, and when cornered.

This is the cheapest trust lever on the surface. It costs minutes, compounds for years, and most local competitors still do not do it.

Response behavior is public, permanent, and read in aggregate. A profile where every review gets a considered answer performs a quality the business cannot claim about itself: attentiveness, demonstrated rather than promised.

Write for the room, not the reviewer. Thank the praise specifically, answer the complaint factually and once, and let the record show a business that shows up. Fifteen minutes a week covers most local volumes.

The owner is the marketing department

The segment is enormous and structurally under-resourced for this work: 36.2 million small businesses in the US alone, employing almost half the private workforce and creating nine of every ten net new jobs. Most run their public surface on scraps: a few hours a week, a few hundred dollars a month, no specialist in the building. Among surveyed Canadian owners already using digital channels, two-thirds name limited time as a challenge when using or adopting digital tools.

AI arrived inside these businesses faster than any prior technology, but it is pointed at operations: invoices, scheduling, drafting. The public surface, where the customer actually decides, is still mostly nobody’s job.

Adoption is real and accelerating; a parallel QuickBooks survey puts regular AI use at 68 percent of small businesses by spring 2025, with 28 percent using it daily. But the tools point inward, at paperwork and scheduling. More than half of US small businesses still say they struggle just to keep their public content fresh.

The squeeze is not a competence gap. It is a structural mismatch between what the surface now demands and what one busy person can hold.

The gap between intent and surface

The gap grows mechanically now. Reviews arrive without permission. Profiles drift stale. Answer engines re-describe the business overnight, in their own words, from whatever the record supports. The owner, standing inside the building, is the one person who never sees any of it the way a buyer does.

byImprint tested this claim on itself before selling it. The audit of its own public surface found seven leaks, including a homepage still selling retired work and profiles sitting dark. The gap is not a failing of bad businesses. It is the default state of any business whose surface nobody reads from outside.

The admin week nobody has automated

Every survey that asks the question finds the same shape: administration takes a large, stable slice of the week, and chasing invoices and payments is the task owners most want gone. The hour figures differ by who asked and how, so this part shows them as a range with each source named, never as one number. Behind the weekly admin sits the paperwork the state requires, and in Canada the smallest firms carry by far the heaviest load per person.

Meanwhile AI use has moved from fringe to common in two years, in Canada and the United States alike. But use mostly means drafting, summarising and answers. Workflow automation, the thing that would give the hours back, remains rare, and the reasons workers give for stopping short are privacy, not knowing where it applies, and skills.

Use is wide and shallow. Three-quarters of the programme owners surveyed use AI and nearly as many say they want help with it, and most owners will not leave a tool unsupervised on even the small tasks. The constraint the sources describe is trust and hands, not the absence of tools.

What the week says

The gap between using AI and getting hours back is trust and hands: owners will not leave a tool unsupervised, and they do not have someone in the building to set one up and stand behind it. That is a different problem from the tools not existing, and it has a different answer.

One rival reading: the tools are not ready. Possible for messy, exception-heavy admin; the surveys do not ask owners why a trial failed, so the sources cannot rule it out.

A second: the admin bill is overstated. The hour figures come from small or vendor-run surveys, and one of the two has 167 respondents. The Canadian regulatory line is a member survey. Every figure carries its sample type for that reason, and none is averaged into a single number.

A third: adoption is real, just elsewhere. AI landed in marketing and drafting first because that is what was sold. This is a timing reading, not a contradiction, and it is the most plausible rival.

What would change the call: a probability-sample survey in which the share automating a workflow rises into the twenties without a matching rise in owners saying they trust the tools.

What it costs a week

About a day a week, every week, on work that nobody has been trusted to take off the desk.

In the owner’s units: hours a week on quoting, invoicing and chasing; days between an enquiry and a quote; evenings spent on paperwork. Those are the units the shifts in parts one to three are paid for in.

byImprint works on exactly this half of the week. A focused project takes one such workflow, understands it, makes it work and makes it yours. An ongoing partnership keeps taking the next one off the desk each month. A focused discovery is the sixty-minute version: one workflow reviewed, a written recommendation, yours either way.

The discipline: rules the work runs under

An owner who commissions research about their own business is buying claims about their own livelihood. A single wrong fact, a competitor miscounted, a defect that does not render on the page a customer actually sees, costs the entire document its authority. So facts are gated before opinion is argued.

The same standard applies inward. Before the methodology was pointed at any paying customer, it was pointed at byImprint itself, and the findings were published and fixed.

1. Every claim must be visitor-real. If a defect or a strength cannot be seen on the rendered page a customer meets, it does not go in the document. Screenshots are the unit of proof.
2. Facts are gated; judgment is argued. Data points pass a verification gate before synthesis. The analysis is opinionated, but never about what the numbers are.
3. Every number ships with its origin. Client documents carry, for each figure, where it came from, when it was checked, and what it can be used to say.
4. Conservative first, transparent math. No projected lift without the arithmetic on the page. Underclaim and let the work overdeliver.
5. Client zero is the practice itself. The first audit ran against byimprint.com. It surfaced seven leaks, and the remediation was done in public before the method was sold.

Eight questions to ask about your own surface

1. Search your trade and your town tonight. Do you appear before your competitors do?
2. Does your homepage sell what you actually sell today, or what you sold when it was built?
3. When did you last read your own Google profile the way a customer does, hours, photos, first three reviews?
4. Do your reviews answer back? What does your silence under the bad one say?
5. Ask an AI assistant what your business does and what it costs. Is the answer true?
6. Which claims on your site could a visitor verify in sixty seconds without calling you?
7. Is your best work visible anywhere, or does it live only in your camera roll?
8. Who owns this in your week? Name the hour it happens in.

The opportunity

Most small businesses are better than they look online. That is the entire opportunity: the work is making the surface tell the truth.

Sources and method

Prepared by byImprint, June 2026; second edition with Part four, September 2026. Every external figure in this document is tied to a named, dated source listed on the closing pages; review and consumer statistics were checked against their primary publishers during preparation.

byImprint field observations are marked as such and come from reads of real owner-operator businesses in British Columbia, with details anonymized.

1. Pew Research Center, “Google users are less likely to click on links when an AI summary appears,” July 2025; browsing data from 900 US adults, covering 68,879 unique Google searches.
2. Semrush, AI Overviews study, 10M+ keyword panel, 2025.
3. Seer Interactive, AIO impact on Google CTR, September 2025 update.
4. SparkToro / Similarweb, US zero-click search analysis, January-April 2026.
5. OpenAI, weekly-user figure at DevDay, as reported by TechCrunch, October 2025.
6. Chatterji, Cunningham, Deming et al., “How People Use ChatGPT,” NBER working paper 34255, 2025.
7. Adobe Analytics, generative-AI referral traffic to US retail, March 2025; Q1 2026 update as reported by TechCrunch, April 2026.
8. Salesforce, Connected Shoppers Report, sixth edition, 2025; 8,350 shoppers.
9. BrightLocal, Local Consumer Review Survey 2026; representative panel of 1,002 US adults. The 2025 comparison uses a separate representative panel of 1,026 US adults, not the same tracked respondents.
10. BrightLocal, Consumer Search Behavior study, 2025.
11. SOCi, Consumer Behavior Index, 2024 and 2025 editions.
12. Whitespark, Local Search Ranking Factors survey, 2023.
13. Google / Purchased, mobile location-search diary study, Think with Google, 2016. Cited as the canonical near-me conversion figure; age noted.
14. ReviewTrackers, Online Reviews Survey, 2022.
15. PowerReviews, How UGC Impacts Conversion, 2023 edition. E-commerce panel; used as directional context only.
16. Michael Luca, “Reviews, Reputation, and Revenue: The Case of Yelp.com,” Harvard Business School working paper 12-016, revised 2016.
17. SBA Office of Advocacy, 2025 Small Business Profile, June 2025.
18. CFIB, Small Business Digital Presence, Your Voice September 2025 survey; 2,355 responses to the digital-tools challenges question. The 66% result describes owners already using digital channels.
19. Fiverr, Small Business Month survey, released May 2025; nearly 6,000 active Fiverr users across 25 countries responded from July 2024 through March 2025.
20. LocaliQ (Gannett), Small Business Marketing Trends Report 2026; 300+ owners, 87% US or Canada.
21. U.S. Chamber of Commerce / Teneo Research, Empowering Small Business, 2025.
22. Intuit QuickBooks, small business survey, April 2025.
23. Verizon Business, State of Small Business Survey, 2025.
24. Constant Contact, The State of Small Business Marketing, September 2025; 18% were very confident in marketing effectiveness, among 2,500 decision-makers surveyed in June 2025.
25. byImprint field data, search-volume pulls, May 2026; anonymized engagement, Sea-to-Sky corridor, BC.
26. byImprint, client-zero audit of byimprint.com and engagement reads, 2026.
27. U.S. Chamber of Commerce Foundation with Ipsos, “Main Street AI Monitor,” 2026; probability sample of 1,070 US small-business workers, fielded May 2026.
28. Statistics Canada, “Analysis on artificial intelligence use by businesses in Canada, second quarter of 2026,” 2026; official statistics, Canadian Survey on Business Conditions.
29. Goldman Sachs 10,000 Small Businesses Voices, “Small businesses embrace AI but need training and support,” 2026; self-selected sample of 1,256 programme participants, fielded February 2026.
30. Bluevine with Centiment, “2026 Small Business AI Trends Report,” 2026; online panel of 942 owners, fielded April 2026.
31. HeyBRB, “UK Admin Drain Report 2026,” 2026; self-run online survey of 167 UK owners, March 2026. American Express with Small Business Saturday UK, “SME Business Barometer,” 2026; 1,000 UK owners, as reported by Retail Times, July 2026.
32. Canadian Federation of Independent Business (CFIB), “Canada’s Red Tape Report: The Cost of Regulation to Small Business (7th edition),” 2025; 2,891 respondents surveyed in 2024, with 2,230 responses used for estimates after filtering and excluding outliers.
33. Business Development Bank of Canada, “The Digital Transformation of SMEs in the Age of Artificial Intelligence,” 2026; online panel of 1,500 owners and decision makers, fielded February 2026.