Diffing the same offering

Subtracting one year’s Item 7 total from the previous year’s takes about ten seconds and is wrong more often than it is right. Two documents with the same brand on the cover can be selling different things, describing different formats, or issued by differently named entities, and in each of those cases the difference between the two totals is not a revision. It is an artefact of comparing two things that were never the same.

The checks below come before any arithmetic. They are unglamorous, and running them is the part of the comparison that produces a defensible answer rather than a number.

Check one: the two documents offer the same thing

Atomic Wings filed on 30 April 2024 and again on 29 April 2025, almost exactly a year apart. The two documents are different offerings. The 2024 document, marked “Atomic Wings FDD 2024 A1 (AR)”, is an area representative offering; the 2025 document is a unit franchise offering.

Their outlet tables count different things, which is the clearest way to see it. Item 20 Table 1 of the 2024 document counts an outlet type called “Area Representatives” — 1 in 2021, 1 to 5 in 2022, 5 in 2023 — alongside company-owned outlets of 0, 1, 2 and 3. The 2025 document counts “Franchised” outlets: 9 to 15 in 2022, 15 to 18 in 2023, 18 to 20 in 2024, with company-owned at zero throughout. Reading one table’s totals against the other compares a count of area representatives with a count of restaurants.

The investment figures separate for the same reason. This index excludes the 2024 Atomic Wings document from its line-item pages because its Item 7 is a laptop, a vehicle and three months of a home office. That is an entirely ordinary Item 7 for an area representative, who is buying a development right rather than building a restaurant, and it would be a category error to publish it as a restaurant buildout or to subtract it from a unit franchise total.

Neither document is deficient and neither supersedes the other. They are two different products, and the honest output of the comparison is that there is nothing to compare. A franchisor may offer a unit franchise, a multi-unit development agreement, an area representative or master franchise, a non-traditional or express format, and a conversion — each with its own Item 7, sometimes in the same document. Establish which one the earlier filing was before treating the later one as its successor.

Check two: the format count has not changed underneath you

Mad for Chicken’s 3 May 2024 filing discloses two Item 7 tables, a full restaurant at $320,125–$687,700 and an express format at $242,500–$466,700. Its 12 March 2025 filing discloses three: a Full Restaurant at $321,125–$691,700 on 2,000–4,000 square feet, an Express Model at $243,500–$470,700 on 750–2,000, and a Multi-Unit Development Agreement at $263,500–$711,700.

Two of those pair cleanly, full to full and express to express, and both moved by $1,000 at the low end and $4,000 at the high. The third pairs with nothing, and not only because the earlier document lacks it. It is not a store format: it prices entering a three-outlet development agreement and opening the first outlet, with a $55,000 development fee, and its footnote says the low end assumes that first outlet is an Express Model while the high end assumes a Full Restaurant. A table like that has no counterpart in a document of single-restaurant tables even when both years contain one.

Averaging the 2025 filing’s three ranges and comparing the result with the 2024 filing’s two would produce a movement of tens of thousands of dollars out of a revision of four thousand, and the whole difference would be an artefact of the arithmetic. Dividing the development range by three to make it look like a per-unit figure first would be worse; reading a multi-unit Item 7 is why.

The trap tightens when a format is renamed rather than added. A table headed differently from last year’s may still describe the same restaurant, and the only way to tell is the format statement above it — footprint, service model, site type, and whether the table covers one unit or a development schedule. Format drives the total is why that statement outranks the row labels underneath it.

Check three: the entity on the cover is the entity you think it is

375° Chicken ‘n Fries filed on 24 February 2023 and again on 30 April 2024. Between the two documents the reporting entity was renamed: the financial statement in the 2023 filing is headed 375 Ventures LLC and the one in the 2024 filing is headed 375 Enterprises LLC. The overlapping years carry identical figures, so this is one lineage under a new name rather than a different business — but the only reason that is knowable is that the numbers were checked against each other.

A renamed entity is not a change to the offering. It is a change to what a reader has to search for, in a state registration database, in a litigation history, in a franchisee list. So is a moved address: German Doner Kebab’s 20 July 2023 filing directs Item 19 reports to an address in Concord, Massachusetts, and its 3 September 2024 filing gives a principal business address in Auburn Hills, Michigan. Neither of those is an Item 7 fact, and both of them will break a search that assumes last year’s identifiers.

Checking the overlapping years is the test. Where two filings show the same period, they should show the same figures; where they do, the entities are the same lineage whatever the letterhead says, and where they do not, that discrepancy is the thing to ask about before anything else in either document is compared.

Check four: the window rolled, and the older year is only in the older copy

Item 19 in the 375° pair covers a rolling window: the 2023 filing reports calendar 2019 through 2022, and the 2024 filing reports 2020 through 2023. FY2019 — sales of $701,815 against a net loss of $42,106, a negative 6.0 percent margin — appears in the earlier document and is absent from the later one. The loss year did not get withdrawn; it aged out of a four-year window, which is how these tables are built.

That is a disclosure mechanic worth carrying to Item 7, where the equivalent happens quietly. A filing states what the franchisor estimates now. It does not carry last year’s estimate for comparison, it does not annotate what changed, and once the older document is out of circulation the earlier estimate is unrecoverable from the current one. The candidate who keeps the previous filing keeps the only copy of the previous answer.

Neither year’s figures are something a buildout budget can be drawn from, and in both filings the representation is a single aggregate income statement for the corporate outlets rather than per-unit figures — so unit economics cannot be derived from it in either year. The point here is only about the window: what scrolls out of view was disclosed once, and only once.

Check five: date the document by its issuance date, not by what is printed on it

Documents in a series are easy to mislabel. Mad for Chicken’s 12 March 2025 filing carries a page footer reading “Rev. April 2, 2024”, which is the wrong year for the document it appears on. The same brand’s 3 May 2024 filing introduces its Item 19 with prose describing “the 2022 and 2021 Gross Revenue” above tables that are headed 2023 and 2022. 375°’s 2024 filing prints a royalty footnote reading “five percent (6%)” against 6 percent in the Item 6 table, which is the figure this index uses.

Every one of those is a drafting slip in a long document assembled under time pressure, and none of them is evidence about a business. They matter for one practical reason: a diff is only as good as the labels on the two documents being diffed, and a footer, a header or a spelled-out number is not a reliable label. Cite the issuance date from the cover page and the Item number, and let the two documents disagree with themselves in the footnotes without letting that disagreement into the comparison sheet.

The preconditions, in order

1. Name the offering in each document — unit franchise, development agreement, area representative, express or non-traditional format — and stop if they differ.
2. Confirm the franchisor entity, and check any overlapping years to see that a rename is a rename.
3. Copy each Item 7’s format statement in full, and pair like with like.
4. Record formats present in one filing and absent in the other as additions or removals rather than as movements.
5. Confirm the working-capital period is stated in the same number of months in both filings before comparing the dollars.
6. Note the issuance date of each document from its cover page, and ignore footers, revision stamps and running heads.
7. Keep the earlier document. It is the only place the earlier estimate exists.

Once all seven pass, the comparison itself is short, and what changes between two filings is what to do with the result. Where the two documents are a year apart but there is only one of them — a first filing, a first offering — the reading is different again, and a first Item 7 covers it. For two brands in the same year, the sibling method is comparing two Item 7 tables, whose step one is this same disqualification applied sideways instead of backwards.

None of these checks is about catching anybody out. They exist because a franchise system is a moving object with a legal identity, several products and an annual filing obligation, and a table that looks like last year’s table is not automatically about the same restaurant. Prove that it is, and a four thousand dollar revision means something. Skip it, and a rename or an added format will look like a hundred thousand dollars of movement that never happened.