Coffee Shop Feasibility Study

A series that begins in FY2001

The coffee record cannot be read from FY1991. Every year from FY1991 to FY1997 is not measured, FY1998 shows a single 7(a) loan whose dollars are suppressed, and FY1999 and FY2000 drop out again. The note under the table gives the reason: when SBA's files leave the industry code blank or unplaceable on too many loans, the class count becomes a floor, and when too few loans carry a code the year is not measured at all. The usable series opens in FY2001 with 241 7(a) loans, dollars withheld, and 5 under 504 for $827,000; FY2002 is the first year with both 7(a) columns printed, 298 loans for $43.2 million. A market section that quotes coffee lending growth since FY1991 would be measuring how SBA coded its files rather than how lenders behaved, and a reviewer who knows the data will catch it. Our studies date any long-run coffee trend from FY2001 for counts and FY2002 for dollars.

The first climb and the break

From FY2002 the class grew quickly. 7(a) approvals went from 298 to 427 in FY2003, 680 in FY2004, 824 in FY2005 and 884 in FY2006, when dollars reached $137.3 million. FY2007 held close at 855 loans and $129.6 million and FY2008 slipped to 668 and $107.6 million; then the volume fell away, to 273 loans and $39.9 million in FY2009 and 261 loans and $39.3 million in FY2010, a lower 7(a) dollar figure than any other year shows. 504 followed the same path down, from 26 loans in FY2005 to 7 loans and $1.3 million in FY2009. For a lender this is the class's stress history: coffee lending can lose most of its volume within a few years of a peak, and the shops already open when it happens still have to carry their debt. A study that shows only a base case leaves that history unanswered, so ours runs the projection through a stretch of lower ticket counts and asks whether the loan still services.

Dollars came back before the count did

The recovery ran at different speeds. Counts rose, with a small dip in FY2013, from 352 in FY2011 to 548 in FY2016 and 679 in FY2018, but the FY2006 peak of 884 was not passed until FY2023, when 928 loans were approved. Dollars recovered much sooner: FY2018's $140.4 million already topped the $137.3 million of FY2006. When dollars regain their high years before the count does, the loans coming back are larger ones, and no year since FY2018 has returned to the smaller tickets of the FY2006 book. The practical consequence for a borrower is that a coffee request today is compared with a book of bigger loans than the class carried before FY2009. A lender will ask whether the project's cost sits in line with that book or above it, and the study has to answer with a budget broken down into land or lease, building or build-out, equipment and opening costs.

FY2020 and FY2021: fewer loans, more money

The middle of the record moved in two directions at once. 7(a) approvals fell from 631 in FY2019 to 515 in FY2020, yet dollars held at $128.0 million, and in FY2021 573 loans carried $216.4 million, a dollar total well above any earlier year on the table. 504 counts kept rising through the same years: 27 loans in FY2019, 31 for $16.2 million in FY2020 and 40 for $16.1 million in FY2021. These years matter beyond their size, because they close the cohort behind every charge-off floor on this page, which counts loans disbursed from FY2017 to FY2021. The losses a lender reads for coffee are therefore measured on a book that ended with its largest tickets of the period. A study for a coffee loan above the median should show how its coverage holds if early sales fall short of the opening forecast, because a lender reading these floors will test exactly that case.

FY2022 to FY2026: the busiest years on record

7(a) coffee dollars set a new high every year from FY2021 to FY2025: $271.6 million on 723 loans in FY2022, $363.7 million on 928 in FY2023, $376.0 million on 973 in FY2024 and $518.1 million on 1,241 in FY2025, the peak in both columns. FY2026 eased to 1,074 loans and $452.3 million, still second only to FY2025. The years from FY2017 to FY2026 hold 7,958 of the 15,854 7(a) coffee loans recorded since FY1991, and $2.74 billion of the $4.00 billion. None of the FY2022 to FY2026 loans sits in the charge-off cohort yet, so the busiest stretch of coffee lending has no measured loss rate. A lender looking at that volume asks the saturation question first: how many coffee units have opened in the trade area over the same years, and how many more the morning commute can feed. Our study counts them on the ground, from national chains to single-site independents, and sets the subject's projected tickets against the share of those trips the site can realistically capture.

504 remains a side channel

Under 504 the class never exceeded 54 loans in a year, a level reached in FY2022 and again in FY2025, and its dollar peak is FY2025's $37.1 million. Since FY1991 the program records 617 coffee loans for $258.8 million, against 15,854 under 7(a), and the 375 approved from FY2017 to FY2026 make up most of that 504 history. Its annual dollar total stayed at or below FY2003's $7.1 million until FY2015 reached $8.6 million. The program also kept its own rhythm: in FY2023 504 fell to 37 loans and $16.1 million while 7(a) climbed to 928. A coffee borrower using 504 is therefore in the smaller and less familiar file, and with cancellation at 15.9% it is the file more likely to fail before it closes. The study for a 504 request has to carry the property case in full: what the land and building are worth to another quick-service operator, and whether the drive-thru, access and parking would hold their value with the site if the coffee business closed.

Two metros level at the top

New York-Newark-Jersey City (435 loans) and Los Angeles-Long Beach-Anaheim (434) are level on count, but they are different coffee markets for a lender. Los Angeles carries $166.0 million at a $288,500 median, the most dollars of any metro, while New York's $128.0 million sits at a $150,000 median and ranks third in dollars behind Dallas-Fort Worth-Arlington ($146.6 million on 307 loans). Both print a charge-off floor, 6.2% in New York and 6.6% in Los Angeles, close to the 6.5% national 7(a) figure. The New York metro median also sits between its two states, New York at $100,000 and New Jersey at $187,500. Within California the metro medians run from $200,000 in San Diego-Chula Vista-Carlsbad to $332,000 in San Jose-Sunnyvale-Santa Clara, with Riverside-San Bernardino-Ontario at $223,500, San Francisco-Oakland-Fremont at $224,000 and Sacramento-Roseville-Folsom at $300,000. A study in any of these metros is read against a deep local record, and the reviewer will expect its comparables to come from the same metro rather than from the state.

The Texas metros and the Houston floor

The Texas metros on the list look alike on size: Dallas-Fort Worth-Arlington at $348,500, Houston-Pasadena-The Woodlands at $348,000, Austin-Round Rock-San Marcos at $349,450 and San Antonio-New Braunfels at $365,200, all near the $348,250 state median. Where they differ is loss. Houston's 257 loans carry a charge-off floor of 19.0%, the highest charge-off rate printed anywhere on this page, against 10.9% for Texas and 6.5% nationally, while the Dallas-Fort Worth-Arlington floor is withheld. A Houston coffee loan therefore reaches a lender whose local evidence is the weakest on record for the class, and the study has to rebuild the confidence that record takes away: the downside case, the sponsor's liquidity and operating experience, and the count of competing coffee units around the site carry more weight in Houston than in any other market on the page.

The large-ticket metros

The highest metro medians belong to North Port-Bradenton-Sarasota ($415,050 on 28 loans), Jacksonville ($377,000), Phoenix-Mesa-Chandler ($369,000), San Antonio-New Braunfels ($365,200) and Las Vegas-Henderson-North Las Vegas ($362,500), followed by Orlando-Kissimmee-Sanford ($352,400). The Florida metros run high across the board: Miami-Fort Lauderdale-West Palm Beach has 136 loans at $324,000 and Tampa-St. Petersburg-Clearwater 95 at $350,000. Phoenix-Mesa-Chandler, with 147 loans and $72.5 million, carries more coffee dollars than Chicago-Naperville-Elgin's 204 loans ($72.2 million). Nashville-Davidson--Murfreesboro--Franklin and Raleigh-Cary sit at $350,000 and Atlanta-Sandy Springs-Roswell at $349,950. In metros where the typical coffee loan is this large, the study has to support a large budget with traffic: the vehicle counts past the site in the morning hours, the stacking room for the drive-thru line, and a ramp to a stabilized ticket count that the competing units nearby make credible.

Ohio and the small-ticket metros

Ohio holds more of the bottom of the median ranking than any other state. Cleveland (101 loans, $100,000), Columbus (100, $123,200), Cincinnati (64, $129,300), Akron (28, $55,250) and Toledo (26, $67,650) all sit among the smallest medians on the list, alongside Urban Honolulu at $50,000, Buffalo-Cheektowaga at $91,000, New Orleans-Metairie at $106,000, Portland-South Portland in Maine at $117,700 and Portland-Vancouver-Hillsboro at $125,000. The Ohio pattern matches the state row, 383 loans at a $100,000 median. A coffee project in Columbus or Cleveland that needs a loan far above the local median stands apart from its own market, and the study has to say why: a freestanding drive-thru where the local record holds mostly smaller formats, the purchase of an established business, or a build-out with costs the local comparables do not carry. Without that explanation a reviewer sees only a loan much larger than its neighbours.

Mid-size metros in the Mountain West and Midwest

Below the largest markets the list reaches mid-size metros with real coffee books. Utah appears through Salt Lake City-Murray (68 loans, $213,600), Provo-Orem-Lehi (61, $279,400) and Ogden (34, $246,000), with the Provo median above Salt Lake City's. Boise City has 50 loans at $257,000, Denver-Aurora-Centennial 116 at $319,800 and Colorado Springs 39 at $329,200, where the charge-off floor is not reported because fewer than 30 of its loans fall in the cohort. Kansas City (75, $349,200) and St. Louis (65, $317,000) sit in the upper half of the size ranking, while Minneapolis-St. Paul-Bloomington (91, $200,000), Indianapolis-Carmel-Greenwood (63, $182,600) and Milwaukee-Waukesha (34, $192,000) sit lower. A coffee site in these metros can draw on a local comparable set of useful size, but its loss evidence comes from the state row or the nation, because no metro floor is printed here.

What the metro floors withhold

The metro charge-off floor prints only for New York-Newark-Jersey City (6.2%), Los Angeles-Long Beach-Anaheim (6.6%) and Houston-Pasadena-The Woodlands (19.0%). Colorado Springs shows not reported, and every other metro on the list shows withheld. Withheld does not mean low, and our study never fills the gap with a number the data does not publish. In a metro whose floor is withheld, our study cites the state floor where the state has one and the national 6.5% for 7(a), and then tests the subject's own coverage, so the lender's loss assumption rests on the project rather than on a figure borrowed from another market. The list ends at New Orleans-Metairie with 25 loans; a site outside every listed metro is studied against its state row and fieldwork in its own trade area.

Quartiles: where the programs overlap

Under 7(a) a quarter of coffee loans are $69,925 or less, the median is $225,000 and the upper quartile $443,750. Under 504 the quartiles are $254,500, $415,000 and $629,500. 504's lower quartile is higher than the 7(a) median, and the 7(a) upper quartile is higher than the 504 median, so the two programs overlap across a wide middle range of loan sizes. A coffee request in that range could plausibly arrive under either program, and the choice turns on what the money buys and on the lender's preference. The study is written for the structure the lender has chosen, because a reviewer reads a property-led file and an operations-led file for different things: the first for site value and reuse, the second for the operator's cash flow and the ramp to it.

Small loans by count, large loans by dollars

The 7(a) band at or below $150,000 is the largest by count, 3,323 loans, and carries $220.9 million, less money than any other occupied band. The 88 loans in the highest occupied band carry $256.9 million, more than all 3,323 small loans together. The band just above $350,000 holds 2,055 loans and $1.16 billion, the most money of any band, and the band above that holds 386 loans for $552.3 million, almost as much as the 2,106 loans between $150,000 and $350,000 ($555.2 million). No coffee loan under either program falls in the top band of the table. In dollars, then, a lender's coffee exposure sits in loans above $350,000, and a study supporting a loan in those bands is reviewed as a material credit, with its market and cost sections checked line by line.

The 504 bands

504 coffee loans sit higher on the scale. Only 31 are at $150,000 or less, for $3.3 million, and 116 fall between $150,000 and $350,000 for $29.0 million. The largest group, 207 loans for $120.7 million, sits in the band above $350,000, and the bands above it are thin: 17 loans with their dollars withheld and 4 more whose dollars are suppressed because fewer than 5 loans share the band. A typical 504 coffee project is a pad site or a small building rather than a large property, and the study that supports it carries a property analysis: land and building cost, the value of the site to another quick-service operator, and the resale case for the drive-thru.

Floors by program

Under 7(a), the cancellation floor on FY2017 to FY2024 coffee approvals is 11.6%, and the charge-off floor on loans funded in FY2017 to FY2021 is 6.5% to date and 3.8% within 5 years. Under 504, cancellation is 15.9% and both charge-off measures stand at 0.00%. The 7(a) gap between the five-year figure and the to-date floor says that coffee losses did not stop at year five: some loans in the cohort failed after their fifth year. A projection that shows coverage only through stabilization therefore misses part of the risk, and our debt service test runs across the term, with a slow-month ticket count and rising labor and occupancy costs. The 504 figures point the other way: the property loans fail more often before they close and show no recorded charge-off after.

What size means for the study

Size also tells the study where to put its weight. A request near the $69,925 lower quartile or the $225,000 median is compared with the thousands of small 7(a) coffee loans, and the reviewer's questions are about the operator: experience, the ticket forecast, labor, and the working capital to reach break-even. A request in the band above $350,000, or under 504, adds the real estate: site value, construction or acquisition cost, and the reuse value of the building and drive-thru. A request in the highest occupied 7(a) bands, which hold 386 and then 88 loans, is rare enough that the lender will want the market and cost evidence in full. Our scope request asks for the loan amount first for this reason, since it tells us which of these studies the lender expects.