How the Squamish report was made

The August records are dated 14 August 2026; the September intake census ran on 8 September, about three and a half weeks later. The August map contains 889 candidates: 12 marked closed or gone and 877 marked open. Of the open businesses, 716 carry an audit record. Mapping a business did not mean auditing it.

The September scan used the August frame. It attempted 755 websites and read 642: 62 were blocked, 43 unreachable and eight excluded by robots rules. It looked for booking, quote, order and enquiry paths on at most six pages per site. It did not submit a booking, order or message.

The reads have different populations and methods. August browser findings use the audited businesses as their denominator; September path shares use the 642 readable sites. Neither denominator represents all businesses operating today.

The discovery frame combined 82 Google Maps category searches, Chamber listings and Downtown BIA directories across Squamish, Garibaldi Highlands, Brackendale, Valleycliffe and Britannia. After deduplication it contained 889 candidates. There is no coverage estimate against a complete business register.

Open means the map did not mark the business permanently closed. It is a dated source classification, not a certification of current trading or a census of every business in the area.

The map was originally built for commercial prospecting, including repair opportunity and fit scores. Of 161 open businesses without audit records, 113 were marked as non-targets, including chains, organisations and brokerages; 48 remained commercial prospects. This selection limits any inference from the audited rate to the wider business population.

The August audits looked at public customer paths, mobile presentation, phone links and domain reachability. The records combine manual browser work and automated browser batches. A severity verdict describes an observed public path; it does not measure bookings, revenue, service quality or what happens inside the business after a customer makes contact. The surviving records do not establish that every failed path was checked again later.

Of 716 businesses with audit records, 399 have a CLEAR verdict, 61 a severity 3 verdict, 68 severity 4 and 44 severity 5. A further 96 have a NO-WEBSITE verdict and 48 have an audit status without a verdict. The 112 severity 4 or 5 verdicts are 15.6% of the 716 audited businesses, rounded to 16%. This is a rate within a selected audit set, not an estimate for the whole town.

CLEAR means that the audit recorded no finding under its customer-path checks. It does not certify a claimed or maintained Google profile, replies to reviews, or every aspect of the business.

No business names or individual records are published. August category rows require at least ten open businesses in the map, but show audited coverage separately. A zero-audit category is not assessed, not clear. The August table contains small aggregate counts; the category threshold does not guarantee anonymity.

For the September category table, the count is withheld whenever it or its complement is below five. The current public aggregate files withhold those exact splits too. Every percentage names its population: the selected August audit set, businesses with sampled reviews, or the readable September websites.

The original 160 of 877 figure is the rescue or repair commercial segment, not a browser-verified failure count. Of those 160, 107 have severity 4 or 5 browser verdicts, two severity 3, 28 an audit status without a verdict and 23 no audit. The 595 originally labelled clear are the other commercial segments; 149 have no browser verdict and 56 have severity 3 verdicts.

The 122 without a website of their own comprises 109 marked as having no website and 13 using only a social or ordering platform. It is distinct from the 96 NO-WEBSITE browser verdicts.

Review enrichment requested up to the three newest reviews per business. Of 775 businesses with response data, 468 had no owner response in that sample, rounded to 60%. Of 393 businesses with at least thirty total reviews, 216 had no response in the sampled reviews. Neither measure establishes that a business has never replied.

The 4.69 rating is the unweighted average of rated profiles. It is not an average weighted by the number of reviews.

The report shows what a customer would have run into on the day of the read, nothing more. It does not say why a given category struggles, and it does not attach a dollar figure to any finding: a severity 5 result means a path is closed, not a number lost, because a read like this cannot see a business’s actual bookings or sales.

These are dated records, not a running measurement. Businesses may have changed since August or September 2026. Missing observations and differences in coverage limit what can be inferred beyond the assessed businesses and readable websites.

Every section and every exhibit in the report carries its own cite line, so a reference points to a specific place in the report rather than to the whole document. The format is: "Cite: byImprint Research, The State of the Squamish Surface, 2nd edition (September 2026), Exhibit C. byimprint.com/squamish#exhibit-c", naming the exhibit and the anchor used.

The eight exhibits in the report run from #exhibit-a to #exhibit-h, in the order they appear. Link directly to the anchor for the figure you are citing, not just to the report’s front page, so a reader lands on the same table or chart you are quoting from.

The business-level records are held privately. Derived counts can be checked against those records, but the original report also included hand-set figures. A preserved archive is not proof that every historical claim can be reconstructed or that the original collection can be repeated today.

The original totals of 51 mobile failures, 56 conflicting phone numbers and 17 expired, parked or unreachable domains have no recovered reproducible classification rule. They are withdrawn from the corrected exhibits pending verification. Historical PDFs preserve what was originally published.

Only aggregates are published. The private records are not offered as a bulk dataset. This is a handling boundary, not a claim that every collection or reuse method is permitted by a platform’s terms. Any future collection needs its own source-rights and action-boundary review.

The first edition was issued in August 2026. The second edition added the September intake census. The 15 September revision changed publisher presentation. The 20 September 2026 correction revises claims, denominators and captions using the existing records; it is not a new data collection.

Quarterly editions are planned for December 2026, March 2027 and June 2027. Before another collection, the frame, method, source rights and action boundaries need to be confirmed. A repeat of a fixed August cohort would describe that cohort; a new town-wide claim would require an updated frame and coverage assessment.