Craftpine

Corner · Store study · July 2026 · Stoney Creek, Hamilton

520 Barton

Corner planned a store of its own here. Could an unattended store with no staff pay for itself on this corner? The study tested eight concepts, and none survived. Its kill criteria were set before any money was spent.

Read the full study
On the web · 9 chapters, 4 appendicesDownload the PDF
78 pages · 11 MB
High aerial view of the Barton Street and Dewitt Road junction, with No. 520 and its lot at centre among houses and townhouses
Barton Street and Dewitt Road from the north-east, April 2025. No. 520 is the building at centre.

The verdict

The verdict was kill. Measured demand, the payback arithmetic and the category's own record each ruled the store out. The evidence was public and free the whole time.

Demandch. 036.8cups a day, the bestpublished resultThe plan needed 12Paybackch. 0727–39months for the two realbuilds at 12 cups a dayThe kill line was 24Precedentch. 03 · app. D0published street-facingstores living on cup marginsRicher operators lostVerdictKillThe cost to prove this model exceeds what itreturns at any demand anyone has measured.Demandch. 036.8cups a day, the bestpublished resultThe plan needed 12Paybackch. 0727–39months for the two realbuilds at 12 cups a dayThe kill line was 24Precedentch. 03 · app. D0published street-facingstores living on cup marginsRicher operators lostVerdictKillThe cost to prove this model exceeds what itreturns at any demand anyone has measured.
The three findings that closed the question, each charted in full further down. From the study's chapters 03, 07 and 09 and Appendix D.

The site

A 90 by 160 foot lot on a signalized corner, zoned C2, Neighbourhood Commercial. The study asked what the small structure at the back could carry.

Barton St EDewitt RdNo. 520: two storeys, three baysNo. 5202 storeysRear structure, 12 by 20 feetbillboard160 ft90 ftNrear structureRear structure, 12 × 20 ft.The subject of the study.No. 520: two storeys, three vehicle bays.Lot 90 × 160 ft, zoned C2.Signalized junction, 21,000+ vehiclesa day, as published.Billboard across the junction, 21,034impressions a day, as listed by its vendor.HSR bus routes 1, 2, 3, 4 and 10.Barton St EDewitt RdNo. 520: two storeys, three baysNo. 520Rear structure, 12 by 20 feetbillboard160 ft90 ftNrearstructureRear structure, 12 × 20 ft: the subjectNo. 520: two storeys, three vehicle baysLot 90 × 160 ft, zoned C2Signalized junction, 21,000+ vehicles aday, as publishedBillboard: 21,034 impressions a day,as listed by its vendorHSR bus routes 1, 2, 3, 4 and 10
The lot, the building and the rear structure the study was about, north up. Schematic, with geometry from the 2007 Ontario Land Surveyor site plan of the parcel.

The traffic

The site was not what killed the plan. The junction's published count is 21,000 or more vehicles a day, and third parties put a price on that traffic. The digital billboard across the junction is listed at 21,034 impressions a day. Five HSR bus routes run the corridor.

Aerial view along Barton Street toward the Dewitt Road junction, a digital billboard across the intersection and the escarpment behind
Barton Street running east toward the corner, aerial, April 2025. The lot is on the right and the escarpment is in the distance.

The structure

The structure is 12 by 20 feet, behind the main building. Any store in it had to run with nobody working there. Every walking route to it from either sidewalk crosses ground that cars drive and park on. The study treats that as a bound on this one structure's uses, and one that could be engineered around. It moved none of the arithmetic.

The small rear structure with its dark roll-up door, across the rear lot at dusk
The rear lot at dusk, April 2025. The structure is the shed at centre.

Eight concepts, and what ended each

Eight concepts entered and none survived. Every exit landed on July 13 or 14, when the evidence arrived together.

  1. CPG micro-retailApr 30 → Jul 13Parked. Payment processing takes about 15% of a $2.50 basket, at any volume.
  2. Staffed bubble teaJul 13Rejected the same day it was proposed. It set the rule of no staff in store operations.
  3. Unattended bubble teaJul 13 → 14Pivoted. Tapioca needs a heated hold, the exact burden an unattended format cannot carry lightly.
  4. Unattended smoothieJul 14Killed. The one measured figure in the category is 6.8 cups a day, against a 12-cup floor. Capital of $26K to $37.6K pays back in 2.2 to 3.3 years, past the 24-month line. Better-funded operators ran it first and lost.
  5. Prepackaged cold drinksJul 13 → 14Passed every format test and was never falsified. It closed with the program and is the most revivable line.
  6. Venue-placed unitsJul 14Closed. Where these machines survive, venue service fees carry them, not cup margins.
  7. Private training venueJul 14Killed on paper, on the change-of-use permit and assembly-occupancy obligations.
  8. Micro officeundated → Jul 14Considered, then closed with the venue lines. The record holds no dated entry for it, and none is invented.
The concept record, as in the study's Figure 2.1. The line in blue passed its own tests and closed with the program.

Demand

The best throughput anyone has published for an unattended fresh-drink machine is 6.8 cups a day, from one kiosk inside a members-only innovation hub. The smallest version of this plan needed 12. That is 75 percent more, from a cold start.

036912Best published resultone kiosk, captive venue, 20236.8 cups a day, best published6.8 cups a daySmallest version of this planthe floor case it needed12 cups a day, needed12 cups a day75% higher04812Best published resultone kiosk, captive venue, 20236.8 cups a daySmallest version of this planthe floor case it needed12 cups a day75% higher
The plan's floor sits 75 percent above the best measured rate. Zenblen's founder reported 2,500 smoothies from one kiosk in 2023, and the floor case is from the study's unit model.

Payback

The program's kill line was 24 months. At 12 cups a day, a $15K machine on its own would pay back in 15.6 months, inside that line. But the machine needs a structure around it, and building that out costs $17,638 to $23,720. With the structure included, the two real builds take 27 and 39 months at the same volume.

0122436months to pay back at 12 cups a day$15K machinethe machine alone, no structure$15K machine: 15.6 months15.6 months$26K buildsmallest real setup$26K build: 27 months, 2.2 years$26K build: 27 months, 2.2 years27 months, 2.2 years$37.6K buildsame, parts at the high end$37.6K build: 39 months, 3.3 years$37.6K build: 39 months, 3.3 years39 months, 3.3 years24-month kill line0122436$15K machine15.6 months$15K machine: 15.6 months$26K build27 months, 2.2 years$26K build: 27 months, 2.2 years$26K build: 27 months, 2.2 years$37.6K build39 months, 3.3 years$37.6K build: 39 months, 3.3 years$37.6K build: 39 months, 3.3 years24-month kill linemonths to pay back at 12 cups a day
Only the machine on its own pays back inside the 24-month kill line. Contribution of $3.87 a cup and fixed costs of $450 a month, from the study's unit model.

Who tried first

Better-funded operators ran this experiment first. Blendid, with about US$36 million paid in, now reports going-concern doubt. Briggo, with US$26 million of stock and note sales on file, was bought by Costa Coffee. Where these machines survive, they run on venue service fees rather than cup margins.

Capital documented, US$ millionsWhere it stands010203040BlendidBlendid: US$36.0 million36.0Going-concern doubt, FY2025BriggoBriggo: US$26.0 million26.0Acquired by Costa Coffee, 2020ArtlyArtly: US$10.3 million10.3Going-concern doubt, FY2025Cafe XCafe X: US$9.7 million9.7Closed city stores, runs airport barsZenblenZenblen: US$4.6 million4.6Active, earns venue service feesBobacinoBobacino: US$3.5 million3.5Closed before it shippedTea Industriesnot disclosedOpened 2024, closed 2025Capital documented, US$ millionsBlendidUS$36.0MGoing-concern doubt, FY2025BriggoUS$26.0MAcquired by Costa Coffee, 2020ArtlyUS$10.3MGoing-concern doubt, FY2025Cafe XUS$9.7MClosed city stores, runs airport barsZenblenUS$4.6MActive, earns venue service feesBobacinoUS$3.5MClosed before it shippedTea Industriesnot disclosedOpened 2024, closed 2025
Earlier operators by documented capital, and where each stood in October 2026. Capital from SEC filings, as cited in the study's Appendix D.

What it cost

The program ran 76 days and spent about zero on the venue. Nothing at the site was built, bought or permitted before the numbers closed it. The spend went on bench hardware and domain renewals. The software the store would have run on is on Corner's page.

$0$10K$20K$30K$40KSpent at the sitethe 76-day programVenue capital spent: about zero≈ $0nothing built, bought or permittedSmallest real buildtotal capital it neededSmallest real build: about $26,000Up to $37,600 with the component build at the top of its range$26,000 to $37,600$0$10K$20K$30K$40KSpent at the sitethe 76-day program≈ $0nothing built, bought or permittedSmallest real buildtotal capital it needed$26,000 to $37,600
Venue capital spent, against the smallest real build the plan needed. From the study's chapters 07 and 09.
Low aerial view of No. 520 from the north, its two bays facing Barton Street and the Dewitt Road junction to the left
No. 520 from the north, April 2025, its two bays facing Barton Street.

The full study

The full study runs to nine chapters and four appendices, and credits every source and photograph. This page is the short version.

Read the full study
On the web · 9 chapters, 4 appendicesDownload the PDF
78 pages · 11 MB