Small Business Cash Flow Benchmarks 2026: What 580 Funded Files Say About Who Falls Behind

Key Takeaways

• Cash cover decides, not credit. Files where the weekly remittance was 100% or more of the cash in the account needed the collections desk 59% of the time and missed their first payment 23% of the time; files with ten or more weeks of cover ran 23% and 4%. Median FICO was 554 in both groups.
• Operators take small tickets. Median funded amount across 468 matured files: $1,000. 49% took under $1,000; 86% took under $2,500; under 2% took $10,000 or more.
• Revenue barely separates payers from missers. Median monthly deposits were $8,397 for clean files and $7,886 for troubled ones — a 6% gap. Median live balance at decision was $351 vs $51 — a 7× gap.
• Over-borrowing doubles trouble. Funded at 80–100% of what the math supported: 15% troubled. Funded at 150%+ of it: 46%, with a 25% first-payment default rate.
• The first Friday tells you almost everything. When the first remittance bounced, 77% of those files ended up in collections; when it cleared, 30%.

Every number on this page comes from Quickie's own funded book — 580 small businesses funded between late July and September 7, 2026, read as aggregates with no merchant identified. It is the dataset we underwrite from, published because the owners it describes deserve to see it. Cite it freely; a link back keeps the numbers in context.

What this data is (and is not)

Quickie purchases a slice of a small business's future receivables and collects it as a predictable weekly remittance sized to a percentage of the business's sales. That is not a loan, and this is not a study of loans. It is a study of what happened after 580 approvals — which files remitted on schedule, which ones needed the collections desk, and what we knew about each file at the moment we decided.

Definitions used throughout:

• Matured file — a funded deal with at least one weekly remittance due. 468 of the 580 qualify; 112 are too new to judge and are excluded from every rate below.
• Clean — every remittance cleared, or a miss was cured within the week. 288 files.
• Troubled — the file needed the collections desk at least once, or was written off. 180 files. A troubled file is not a lost file; many are paying again.
• First-payment default (FPD) — the very first weekly remittance did not clear.
• Decision-time features — what the bank data said the day we decided: average daily balance, live balance, deposits per month, revenue trend, weekly remittance as a share of cash and revenue, plus credit, time in business, industry, and state.

Sample sizes are printed with every table. Anything under 20 files is a hint, not a benchmark. The book is small compared with a bank's; it is also unusually honest, because it is drawn from live bank feeds rather than what applicants typed on a form.

Cluster 1 — How much small businesses actually take

The first thing the data corrects is the size of the ask. Marketing for business funding lives in the $50,000–$250,000 range. Operators do not.

Funded amount	Files	Share of book	Troubled	First-payment default
Under $1,000	230	49%	39%	21%
$1,000 – $2,499	172	37%	39%	13%
$2,500 – $4,999	41	9%	39%	7%
$5,000 – $9,999	19	4%	37%	6%
$10,000 +	6	1%	0%	0%

Quickie funded book, 468 matured files, decision dates Jul 24 – Sep 7, 2026.

The median funded amount is $1,000. The average is $1,326, pulled up by a handful of five-figure files. Eighty-six percent of the book is under $2,500.

Two things stand out. First, the troubled rate is flat across size — 39% whether the ticket was $600 or $4,000. Ticket size is not what makes a file miss. Second, first-payment default falls steadily as the ticket rises, from 21% under $1,000 to 6% at $5,000–$10,000. That is not because bigger money is safer; it is because bigger tickets went to stronger accounts. The underwriting that sizes a file is doing its job at the top of the range and is most tested at the bottom, where the smallest businesses live.

For an owner, the lesson is to size the ask to the use, not to the maximum. Here is the sizing math we recommend before you apply — and the working capital formula that tells you how much you actually need this month.

Cluster 2 — Cash cover: the number that predicts a missed payment

If you read one table on this page, read this one. "Cash cover" is the live bank balance at decision divided by the weekly remittance — how many weeks of pulls the account could absorb today if nothing else came in.

Weekly remittance as % of live cash	Weeks of cover	Files	Troubled	First-payment default
Under 10%	10+ weeks	102	23%	4%
10% – 25%	4 – 10 weeks	90	31%	12%
25% – 100%	1 – 4 weeks	73	23%	15%
100% or more	Under 1 week	104	59%	23%

Quickie funded book, 369 matured files with a live balance read at decision.

The cliff is at one week. A file with less than one weekly remittance sitting in the account on decision day needed collections 59% of the time and missed its very first pull 23% of the time. Above ten weeks of cover, those numbers fall to 23% and 4%. First-payment default is the cleanest gradient in the entire dataset — 4%, 12%, 15%, 23% — and it moves in lockstep with cover.

The medians say the same thing in one line: the typical clean file's weekly remittance was 17.6% of its live cash; the typical troubled file's was 66.7%.

Live balance alone, before any ratio, tells a near-identical story:

Live balance at decision	Files	Troubled	First-payment default
Negative	18	56%	17%
$0 – $100	145	53%	27%
$100 – $500	101	31%	14%
$500 – $1,500	73	25%	7%
$1,500 – $5,000	44	23%	5%
$5,000 +	31	26%	3%

Nearly a third of the matured book (145 files) was decided with between $0 and $100 in the account. Those files ran 53% troubled and 27% first-payment default. The median clean file had $351 on hand at decision; the median troubled file had $51.

This is the finding that changed our own policy. As of September 2026, Quickie's sizing engine caps the weekly remittance at one quarter of live cash — four weeks of cover on funding day — and a live balance under $100 drops a file to the floor tier regardless of how healthy the average balance looks. If you want to be approved for more, the fastest lever you control is not your credit score. It is the balance in the account on the day you connect it.

What underwriters actually read in a bank statement walks through cover, cadence, and the other lines below in detail.

Cluster 3 — Average daily balance benchmarks

Average daily balance (ADB) is the smoothed version of the same idea — what the account typically holds across the review window, not just on decision day. It matters, but less than the live number.

Average daily balance	Files	Troubled	First-payment default
Under $250	63	59%	31%
$250 – $500	84	43%	19%
$500 – $800	71	31%	20%
$800 – $1,500	86	34%	12%
$1,500 – $3,000	67	36%	10%
$3,000 – $8,000	70	39%	13%
$8,000 +	27	19%	4%

Quickie funded book, 468 matured files.

Median ADB: $998 for clean files, $668 for troubled files. The bottom bucket is brutal — under $250 of average balance ran 59% troubled and 31% first-payment default — and the top bucket is the only one that reads like a bank customer (19% and 4%). Everything between $800 and $8,000 sits in a broad middle where ADB alone does not decide the outcome; cover and trend do.

A more useful benchmark for owners is ADB as a share of monthly revenue — how much of a month's deposits the account keeps rather than spends through:

ADB as % of monthly revenue	Files	Troubled	First-payment default
Under 3%	33	42%	27%
3% – 6%	64	45%	22%
6% – 10%	105	33%	8%
10% – 20%	121	35%	17%
20% – 40%	88	44%	17%
40% +	38	16%	11%

The median business in this book keeps roughly 11% of a month's deposits as average balance — about three and a half days of revenue. That is thin. The JPMorgan Chase Institute's well-known figure puts the median U.S. small business at about 27 cash-buffer days; the businesses that apply for weekly working capital are, by selection, the ones running well under that. Keeping 40% or more of a month's revenue as balance — roughly twelve days — was the one band that behaved like a bank customer (16% troubled).

Cluster 4 — Credit score barely moves the needle

Here is the table that surprises owners who have been declined by a bank.

Personal FICO at decision	Files	Troubled	First-payment default
Under 500	62	44%	13%
500 – 549	132	33%	10%
550 – 599	132	39%	24%
600 – 649	56	46%	16%
650 – 699	21	14%	10%
700 +	11	27%	9%

Quickie funded book, 414 matured files with a soft credit pull on file. The 650+ bands are small.

The median FICO was 554 in both groups — clean and troubled. The 500–549 band (33% troubled) outperformed the 600–649 band (46%). Derogatory tradelines, collections accounts, and revolving utilization were just as flat: files with 90%+ utilization ran 37% troubled versus 40% for files under 30%. The median clean merchant carried 12 derogatories; the median troubled merchant carried 13.

This is not an argument that credit is meaningless. It is an argument that for a weekly remittance sized to deposits, credit history is the wrong instrument. A 620 with an empty account is a worse file than a 520 with four weeks of cover, and the data says so at every revenue band we can test. It is also why a soft pull is enough — the score informs pricing at the margin, but the decision lives in the bank feed. The full operator guide to funding with bad credit and the bad-credit funding hub explain what does and does not get approved, and soft vs hard pulls covers why the inquiry itself should not cost you points.

Cluster 5 — Revenue trend and time in business

Revenue level does not separate the groups. Revenue direction does.

Revenue trend (per month, trailing)	Files	Troubled	First-payment default
Sliding more than 15%	50	60%	22%
Sliding 5% – 15%	69	33%	16%
Flat (−5% to +5%)	98	34%	9%
Growing 5% +	184	35%	16%

Quickie funded book, 401 matured files with at least three months of deposit history.

A business whose deposits were falling more than 15% a month at decision ran 60% troubled — the worst single band in the dataset outside of creators. Everything from a gentle slide to healthy growth sat in a tight 33–35% range. The median clean file was growing about 4.6% a month; the median troubled file, 1.5%.

Revenue level, by contrast, is nearly useless on its own:

Monthly deposits	Files	Troubled	First-payment default
Under $3,000	57	53%	27%
$3,000 – $5,000	77	40%	21%
$5,000 – $8,000	95	34%	16%
$8,000 – $12,000	73	34%	14%
$12,000 – $25,000	113	35%	14%
$25,000 +	53	42%	8%

Below $3,000 a month, half the book struggled — a weekly remittance of any size is a large share of a small account. Above $5,000, the troubled rate flattens at roughly a third all the way up. A $30,000-a-month business is not safer than an $8,000-a-month business if both keep $80 in the account on Friday.

Time in business behaves the way lenders expect, with one wrinkle:

Time in business	Files	Troubled	First-payment default
Under 12 months	56	45%	27%
1 – 2 years	81	47%	17%
2 – 5 years	133	38%	17%
5 + years	198	34%	12%

Median tenure: 51 months for clean files, 42 months for troubled. Under-two-year businesses ran 45–47% troubled, and the newest cohort carried a 27% first-payment default rate — double the five-year cohort. Since September 2026 Quickie caps any file under twelve months in business at a B grade regardless of cash, because youth plus thin cover was the worst combination we found: under a year old and under a week of cover ran roughly seven in ten troubled. Can a new LLC get funding? covers what a young business can realistically qualify for.

Cluster 6 — Deposit cadence, real income, and who is on the filing

Three more account-level signals, each smaller than cash cover but real.

Deposits per month. A business that lands money on many days is easier to remit from than one that lands it twice.

Deposit days per month	Files	Troubled	First-payment default
Under 4	37	49%	22%
4 – 8	62	45%	18%
8 – 12	113	34%	19%
12 – 18	116	37%	16%
18 +	140	38%	12%

Fewer than eight deposit days a month ran 45–49% troubled; eight or more sat in a 34–38% band. The median for both groups was 13 deposit days — cadence separates the bottom of the book, not the middle. Quickie's top grade now requires eight or more deposit days a month for exactly this reason.

A business we can find. When the business had a verifiable public footprint — a state filing, a storefront, a review presence — the troubled rate was 30% (n=107). When we could not find it, 40% (n=108). An invisible business is not a fraud signal by itself, but it removes the corroboration that lets an underwriter believe the deposits are what the applicant says they are.

EIN vs sole proprietor. Files filed under a business EIN ran 34% troubled (n=291); sole proprietors ran 45% (n=177). First-payment default was identical at 16%. The gap is less about entity type than about what tends to come with it — a separate operating account, commercial deposit rails, and a business that exists on paper somewhere. Funding with an EIN only covers what changes when the business, rather than the owner, is the applicant.

Cluster 7 — The most expensive habit: taking more than the math supports

Every file Quickie funds has an engine number — the amount the deposits, cover, and cadence support. The desk can fund above it. When it does, the data is unforgiving.

Funded amount vs. engine-supported amount	Files	Troubled	First-payment default
Under 50% of supported	55	29%	13%
50% – 80%	42	29%	7%
80% – 100%	34	15%	6%
100% – 150%	64	31%	6%
150% +	151	46%	25%

Quickie funded book, 346 matured files with an engine amount on record.

Funded at 80–100% of the supported amount: 15% troubled, 6% first-payment default. Funded at 150% or more: 46% and 25%. The median troubled file was funded at 183% of what its own numbers supported; the median clean file at 110%.

This is the single most actionable finding in the dataset, because it is the one both sides control. Owners ask for a round number; desks want to say yes. Every stretch past the window in our book carried under a week of cover — the two habits travel together. Quickie's Decision Snapshot now prints this exact table's odds next to any override, so the cost of a stretch is stated in numbers before anyone signs. If you are offered less than you asked for, the counter is not an insult. It is the amount the account can carry. When not to take funding covers the cases where the right answer is a smaller number or none.

Cluster 8 — Industry benchmarks

Industry matters less than owners fear and less than the trade press implies, with two exceptions.

Industry	Files	Troubled	First-payment default
Health & fitness	14	7%	0%
Retail	21	24%	14%
Trucking & transport	47	34%	17%
Beauty & salon	56	36%	13%
Restaurant & food	62	39%	19%
E-commerce	10	40%	40%
Construction	25	44%	12%
Professional services	112	44%	16%
Creator & influencer	12	50%	42%
Music & creator	16	75%	33%

Quickie funded book, 468 matured files, industries with 10+ files. Bands under 20 files are directional.

A few readings:

• Restaurants (39%), salons (36%), and trucking (34%) — the three verticals the industry treats as high-risk — sat at or below the middle of the book. Their cash is lumpy, but it is real, it arrives on commerce rails, and it arrives often. Above four weeks of cover, restaurants in our book essentially did not miss. Restaurant funding, salon and spa funding, and trucking funding have their own hubs, and the restaurant cash-flow playbook explains why the "risky restaurant" reputation is mostly about cover.
• Retail (24%) and health & fitness (7%) ran cleanest — recurring memberships and daily card batches are the friendliest deposit shapes for a weekly remittance. See retail funding and gym and fitness studio funding.
• Construction (44%) and professional services (44%) ran above the middle. Both are invoice businesses: money arrives in large, irregular lumps that leave the account almost as fast as they land. The contractor bridge guide and construction funding cover how to size around draws.
• Creators, musicians, and influencers (50–75%) were the one true outlier — platform payouts are irregular, personal and business money mix, and the account is often near zero between payouts. These files are sized as floor tickets or declined, and the data says that is right.
• E-commerce is a small sample (10 files) with a startling 40% first-payment default. Ad spend and inventory leave the account before payouts settle. E-commerce working capital and the e-commerce seller hub cover the timing problem.

Our own industry scoring was recalibrated on these numbers in September 2026: the risk adjustment that used to apply to restaurants, trucking, beauty, and retail was removed, and it now applies to construction, e-commerce, and creators. How underwriters score industry risk has the longer discussion.

Cluster 9 — State snapshot

Geography is not a credit factor and Quickie does not price by state. The distribution is still interesting, because it tracks where small-ticket demand concentrates.

State	Files	Troubled	First-payment default
North Carolina	24	21%	13%
Texas	63	22%	16%
Illinois	19	26%	16%
Ohio	20	35%	20%
Florida	65	38%	19%
Pennsylvania	24	42%	21%
Georgia	47	45%	21%

Quickie funded book, states with 19+ matured files.

Florida, Texas, and Georgia together are 37% of the book. Texas ran cleanest of the large states (22%); Georgia ran hottest (45%). We would not read anything causal into a state row — the industry and cover mix inside each state explains most of it — but it is a useful sanity check for owners comparing notes. State pages for Florida, Texas, and Georgia carry the disclosure-law specifics that do differ by state, and the full state index covers the rest.

Cluster 10 — The first Friday tells you almost everything

Weekly remittances at Quickie run on Fridays. The first one is a referendum on the underwriting.

First weekly remittance	Files	Eventually troubled
Cleared	386	30%
Did not clear	75	77%

Quickie funded book, 461 matured files.

Sixteen percent of matured files missed their first pull. Of those, 77% went on to need the collections desk; of the files whose first pull cleared, 30% did. A first-payment default is not a bad week. It is the account telling you the remittance was never going to fit.

Two operational consequences followed. First, every desk pull now reads the live balance immediately before it submits and holds the row when cover is short — so the merchant's bank never sees an authorization it would decline, and a short week becomes a conversation instead of a returned item with a fee on both sides. Second, cover on funding day is now checked separately from cover at decision, because the two can be days apart and the account can drain in between. What happens when you cannot pay back an advance covers the merchant side of a miss honestly.

Cluster 11 — Bank data is not bank truth: the negative-days artifact

The most useful thing this dataset taught us was about the data itself.

Every cash-flow underwriter reads a balance history. Most of us read it from an aggregator's asset report rather than the bank's PDF. The catch: aggregators compute historical balances by walking backward from today's balance through the transactions they managed to pull. Every posted debit the feed missed makes every earlier day look poorer than it was. Feeds from certain banks and fintech accounts drop card swipes and sub-account transfers routinely, so the past can read months more negative than reality — and the deepest "negative" is always the oldest day.

In a September 2026 audit of 382 declined files with $3,000 or more in monthly revenue, 156 carried negative days the bank itself never recorded, and 104 had zero real negative days once the statement ending balances were checked. One trucking file showed 67 negative days and a −$13,269 low in the feed; the bank's own statements showed zero negatives and a positive four-figure average. Merchants had been telling us "I don't have negative days." They were right.

The raw negative-day count, unsurprisingly, predicted nothing in the funded book:

Reported negative days (raw feed)	Files	Troubled
None	242	38%
1 – 4	99	40%
4 – 8	50	38%
8 – 15	37	30%
15 +	40	48%

Neither did NSF count — files with four or more NSF or overdraft fees in the window ran 22% troubled, better than files with none (41%), because a bank that charges NSF fees is a bank that lets a merchant run negative rather than declining the debit, and those merchants tend to have real operating accounts.

What we changed: a balance series from an institution that cannot carry a negative balance, or one showing dozens of negative days with no fee lines, is now corrected to a floor before it is graded, and the correction is printed on the file. What you can do: if any funder declines you for negative days you did not have, ask them to read the ending balance on each statement. It takes five seconds, and the statement beats the feed every time. Reading bank statements like an underwriter shows exactly where to look.

Cluster 12 — Existing funders: a paying stack is not a kill

Conventional wisdom says a merchant already remitting to another funder is a decline. The data disagrees, with a condition.

Weekly load to other funders at decision	Files	Troubled	First-payment default
None detected	326	41%	20%
$1 – $150 / week	50	30%	18%
$150 – $400 / week	44	39%	2%
$400 + / week	48	31%	4%

Quickie funded book, 468 matured files. Loads detected from named funder debits on the bank feed.

Merchants already carrying another funder ran 30–39% troubled versus 41% for merchants with none, and their first-payment default rates were dramatically lower (2–4% at the heavier loads versus 20%). They carry more cover, they have already proven a weekly can clear, and they know what Friday feels like.

The condition is the word paying. A funder debit that landed and then stopped — remittances that died well short of a payoff — is a different file entirely and is an automatic decline on our desk. The stack itself is a sizing input, not a verdict: the combined weekly load across every funder has to fit inside the same cash-cover rule, which usually means a smaller ticket, not no ticket. MCA stacking, the quiet killer explains the difference between a stack that is serviced and one that is not.

Cluster 13 — Speed and rails: what "same-day" looks like in practice

Speed is the reason this product exists, so here is what it actually looks like at volume.

• Decision: minutes after the bank connects. The underwriting above — cover, cadence, trend, stack, credit, footprint — runs in parallel while the asset report generates.
• Release: same-day ACH is the default rail once the agreement is signed and the desk clears the file. Funds released after banking hours land the next business day.
• Remittance: one weekly ACH, Fridays, sized to a percentage of sales and disclosed in total dollars before signing. On the Friday before this post was written, the book ran 282 remittances — 259 by bank ACH and 23 on a card the merchant chose to keep on file as a backup.
• Volume: in August 2026 the funnel processed roughly 2,600 applications and 3,700 bank connections and funded 431 files — about one in six applicants. That ratio is the honest answer to "how hard is it to get approved": most declines are cash-cover declines, and most of those are fixable by the owner in a month.

Same-day funding, the reality walks through the difference between a decision in minutes and money in the account, and the same-day business funding hub compares how the major online funders handle the same clock.

How to use this if you run a business

The data reduces to a short checklist. None of it requires a better credit score.

1. Know your cover. Divide the balance in your operating account by the weekly remittance you are considering. Under one week is the cliff; four weeks is where the book behaves. If you are under four, the fix is timing — apply the week a large receivable lands, not the week after payroll.
2. Keep the account that pays you. A weekly remittance is easiest on an account that receives deposits on eight or more days a month. If your income lands in one account and your bills leave another, connect the one the income lands in.
3. Ask for the use, not the maximum. Every table above says the same thing: files funded past what the math supported carried the worst outcomes. Write down what the money buys, what it returns, and when. The working capital formula does this in ten minutes.
4. Watch the slope, not the level. A business sliding 15% a month should fix the slide before it takes on a weekly remittance. Growth of any size, or flat revenue, was fine.
5. Exist on paper. A state filing, a storefront listing, a business phone that rings — the businesses we could find ran ten points cleaner, and it is the cheapest lever on the list. Three documents that fund a file this week covers what to have ready.
6. Check your own negative days. Pull the ending balance from each of your last three statements before you apply. If a funder's read disagrees, say so.
7. Treat the first Friday as sacred. Cover the first remittance like rent. In the data, clearing it cut the odds of trouble by more than half.

How we use it

This is also a transparency note on how Quickie underwrites in September 2026, because the numbers above are the reason the rules look the way they do:

• Weekly remittance capped at one quarter of live cash on funding day; live cash under $100 or a negative live balance drops a file to the floor tier ($300–$1,500 over 16 weeks) or a decline.
• Under twelve months in business caps at a B grade; revenue sliding 15% or more a month caps at a C.
• The top grade requires a business we can find online, eight or more deposit days a month, and at least 60% of inflows classifying as real operating income. Cash on hand alone never buys it.
• Industry adjustments apply to construction, e-commerce, and creators — not to restaurants, trucking, beauty, or retail.
• A paying stack is sized around, not declined. A stopped-paying funder is an automatic decline.
• Balance histories that cannot be real are corrected before grading, and the correction is shown.
• Every desk pull reads the live balance first and holds when cover is short. Nothing debits an account on its own.
• Every Decision Snapshot carries a "book prior" — the troubled rate of funded files shaped like the one in front of us, from this dataset — so the desk reads the odds before it writes a number.

The priors refresh from the live book. When they move, this page moves with them; the updated date at the top is the tell.

Sources & methodology

Primary data. Quickie Business Services LLC funded book, 580 files funded July 24 – September 7, 2026; 468 with at least one weekly remittance due as of September 7, 2026. Decision-time features are read from the Plaid Asset Report and soft credit pull on file at the moment of decision; outcomes are read from the remittance ledger (cleared, returned, held, cured, collections, written off). All figures are aggregates; no merchant, business, or account is identifiable. Bands with fewer than 20 files are labeled directional. "Troubled" means the file required the collections desk at least once or was written off; "clean" includes files that missed and cured within the week. The underlying audit is the same one that sets Quickie's underwriting thresholds, and it is re-run weekly.

Context sources. Cash-buffer benchmarks for the broader small-business population are from the JPMorgan Chase Institute's Cash is King: Flows, Balances, and Buffer Days. Application and approval patterns across lender channels are from the Federal Reserve Banks' Small Business Credit Survey. The mechanics of aggregator balance histories are documented in Plaid's Assets product reference. ACH return handling follows Nacha Operating Rules. Quickie's own small business funding statistics page collects the public numbers this dataset should be read against.

What this is not. It is not a study of loans, APRs, or default rates on lending products; Quickie purchases future receivables and this is remittance behavior on that product. It is not a representative sample of U.S. small businesses; it is the population that applied for and qualified for $1,000–$25,000 of weekly working capital in a six-week window. Correlations are reported, not causes. Every file prices individually and nothing here is a promise of approval.

Bottom line

Five hundred and eighty files later, the pattern is not subtle. The businesses that fall behind on weekly working capital are not the ones with bad credit, small revenue, or the "wrong" industry. They are the ones that were funded into an empty account, or for more than the account could carry, or while revenue was sliding. The businesses that pay clean keep a few weeks of cover, get paid on many days a month, and take the number the math supports.

If that describes you, see what your deposits qualify for — soft pull, decision in minutes, every dollar disclosed before you sign, and a weekly remittance sized to your sales. If it does not describe you yet, the working capital funding hub and the checklist above are where to start; most of what moves these tables is inside an owner's control within a month.