What the range hides

An Item 7 range is not a probability distribution. The midpoint is not a quoted “typical” project unless the filing says so, and the low end is not the price a qualified buyer should expect. The endpoints combine assumptions about format, site, market, payment timing, and operating period that can move together.

Different spaces produce different scopes

Construction is the obvious source of spread. A compliant second-generation restaurant may preserve infrastructure that a shell needs from scratch. Landlord work and a tenant-improvement allowance can change who writes the check and when. Local labor, permit, utility, and accessibility requirements can move the project outside a generic estimate. A wide construction range may therefore describe several site conditions, not contractor uncertainty around one defined scope.

Footprint compounds the difference. The 800–1,500-square-foot outlet in one filing and the 2,000–4,000-square-foot full restaurant in another do not need the same seating, finishes, HVAC, refrigeration, or working inventory. Comparison starts with the format statement printed above the line items.

Labels bundle unlike costs

The public filings show the problem directly. Minnesota’s filed Shah’s Halal Item 7 uses a fixture package alongside a separate build-out line. A filed Great Greek disclosure puts equipment, furniture, supplies, and fixtures together. In this index, German Doner Kebab separately discloses restaurant equipment, fit-out materials and installation, smallwares, furniture and fixtures, and mechanical, electrical, and plumbing.

Those tables may all be internally valid. They are not row-for-row substitutes. Adding every relevant row within one filing is safer than comparing a single “equipment” label across filings. When a package cannot be unpacked from public data, the package stays intact.

Discounts and zeroes need context

A low total can include a fee discount available only to a defined class of buyer. A $0 low value can mean the cost is not expected in a qualifying site, could be covered by another party, or is estimated as no additional payment under the stated assumption. It is a disclosed zero, not a blank. Conversely, a brand with a published total but no line-item source in this dataset remains total-only; no rows are reverse-engineered from the difference.

Arithmetic deserves the same restraint. If a filing’s line items do not add to its printed total, report the printed total and the sum. The Shah’s Halal page in this index does that. Quietly changing one of the values would make the directory cleaner and the source record less accurate.

Time changes the comparison

Filing years matter because labor, equipment, freight, insurance, and rent assumptions age. The initial period matters because three months of additional funds is not six months. Payment timing matters because a project can exhaust cash before every reimbursement or loan draw arrives even when the final total fits the estimate.

The FTC’s consumer guide to buying a franchise recommends comparing the filing with what franchisees in the system and competing systems actually paid. That is the right use of the range: a starting hypothesis to test against recent openings, site quotes, and a dated cash-flow model.

Read the high end as exposure under the franchisor’s assumptions, not as a cap. Read the low end as a conditional endpoint, not a target. Then document which conditions the real project satisfies.