Every Minute of Application Time Has a Price. Here's How to Calculate Yours
A model for pricing the minutes in your application flow: volume, drop-off sensitivity, value per file. With a worked example for your own numbers.
Alfred BEditorial Reviews
Every lender has the same argument once a year. Product says the application is fine. Sales says it is bleeding. Nobody has a number, so the longest-serving voice wins and the form stays as it is.
A minute of application time costs the applications it loses, times the money each surviving application eventually earns. Multiply monthly starts by the drop in completion rate per extra minute, then by approval rate, close rate and contribution per funded file. The result is a monthly dollar figure per minute. Everything else in this calculation is measurement.
The arithmetic is ordinary. It rarely gets done because four of the six inputs live in different systems and one has to be measured rather than looked up.
What is an application abandonment cost calculator?
An application abandonment cost calculator is a model that converts drop-off during an application into an annual dollar figure, by multiplying the applications lost against the contribution a funded application produces. It is not a benchmark and it is not portable. Two lenders in the same vertical will land an order of magnitude apart, because the variable doing most of the work is contribution per funded file.
The output is a rate, not a total. Cost per minute, per month, at current volume. That is the form a product decision takes: this change removes three minutes, and here is what that is worth.
Which inputs does the calculation need?
Six numbers, plus one that has to be earned through a test.
| Input | What it means | Where it usually lives |
|---|---|---|
| Starts per month (S) | Applicant entered at least one field | Product analytics, not origination |
| Completion rate (c) | Share of starts that reach submitted | Product analytics |
| Median time to complete (t) | Median minutes from first field to submit | Product analytics, timestamped |
| Approval rate (a) | Share of submitted files approved | Loan origination system |
| Close rate (f) | Share of approvals that fund | Loan origination system |
| Contribution per funded file (m) | Revenue on a funded loan less variable production cost | Finance |
| Sensitivity (Δc) | Completion points gained or lost per minute | Measured, never assumed |
Table 1: the inputs an application abandonment cost calculator requires, and the system each one normally sits in. The last row is the only one that cannot be pulled from a report.
Sensitivity is the input people want to skip. There is no published figure for it in lending, so the temptation is to import a checkout-abandonment number from ecommerce. That number describes a different decision made by a different person under different pressure. Measuring your own takes one release and four weeks.
Where does the contribution number come from?
Most lenders already hold contribution per funded file under another name.
For a public anchor, the Mortgage Bankers Association's Quarterly Mortgage Bankers Performance Report for the first quarter of 2026, released in May 2026, put pre-tax net production profit at $727 per loan for independent mortgage banks and bank mortgage subsidiaries, against total loan production expenses of $11,898 per loan. Note what that ratio implies: $727 of profit sits on top of an $11,898 cost base, so a single extra funded loan is worth many hours of the staff time that produced it.
Non-mortgage lenders run different structures. A merchant cash advance, a used-car retail contract and an unsecured personal loan produce contribution figures that share nothing except the units.
How does the calculation run on real numbers?
The figures in this section are illustrative. They are constructed to demonstrate the method, not drawn from any study, survey or published dataset, and no lender's real numbers are represented here.
Take a lender running 1,000 application starts a month. Completion is 62%. Of submitted files, 45% get approved, and 80% of approvals fund. Contribution per funded file is $727, borrowed from the mortgage benchmark above for illustration.
That produces 620 submitted, 279 approved and 223 funded a month, worth about $162,000 in contribution.
Now the test. The lender removes an income-document upload step and pre-fills two manual fields, cutting median time to complete from 14 minutes to 11. Completion moves from 62% to 66%.
Run the same chain: 660 submitted, 297 approved, 238 funded. The gain is 14.4 funded files a month, about $10,500 in monthly contribution, roughly $126,000 a year.
Three minutes bought that. Divide it out and a minute of application time is worth about $3,500 a month at this volume.
What is a minute worth at other assumptions?
The answer swings on two inputs, sensitivity and contribution. Holding the illustrative volume at 1,000 starts, a 45% approval rate and an 80% close rate:
| Sensitivity | $400 contribution | $727 contribution | $1,500 contribution |
|---|---|---|---|
| 0.5 points per minute | $720 | $1,310 | $2,700 |
| 1.0 points per minute | $1,440 | $2,620 | $5,400 |
| 1.33 points per minute | $1,920 | $3,480 | $7,180 |
Table 2: illustrative monthly value of one minute of application time, at 1,000 starts per month. These are worked outputs of the formula above, not measured results from any lender.
The spread across that table is ten to one. Anyone quoting a universal cost-per-minute for lending applications is quoting the middle cell of a table like this one and not checking the corners.
What does the handling side add?
The lost-application cost is the larger half for most lenders. It is not the whole bill. Time inside the application is also time a human spends touching the file.
The US Bureau of Labor Statistics reported a median wage for loan officers of $74,180 a year, or $35.66 an hour, in its May 2024 occupational data. At that rate a staff minute costs about 59 cents. On the illustrative lender above, nine minutes of staff handling per started application works out to 150 hours a month, near $5,350, before management time and before the rework that follows a file back and forth.
In the private credit years the file that ruined a Friday was never the complicated one. It was the straightforward file waiting on one more bank statement, sitting in a queue, being re-explained to a borrower who had already answered the same question three days earlier.
Does the evidence support the direction of this?
Borrower preference is documented. The Federal Reserve Banks' Small Business Credit Survey, 2020 Report on Employer Firms, found 46% of online lender applicants cited speed of decision or funding as a reason for applying where they did, and 54% of large bank applicants said the same.
Speed also does not appear to require looser credit. Federal Reserve Bank of New York Staff Report 836, published in February 2018, found fintech mortgage lenders processed applications about 20% faster than other lenders while their default rates ran roughly 25% lower.
Neither finding gives a cost per minute. They establish the sign, not the size.
What we don't know
Three gaps, stated plainly, because this topic attracts confident numbers that have nothing behind them.
There is no credible published abandonment rate for loan applications. We looked. Every percentage in circulation traces back to vendor marketing or to trade content recycling an unattributed figure, with no regulator, statistical agency or peer-reviewed study behind it. A calculator that starts from a borrowed abandonment rate is producing a number about somebody else.
There is no published elasticity of application completion to minutes in lending. The sensitivity input in Table 1 exists nowhere in public literature we could find, which is why it has to be measured in your own flow rather than cited.
And there is no Canadian data on time-to-decision or time-to-funding. The federal SME financing surveys publish approval rates and nothing on speed, so every timing figure available for this work is American or British.
How do you measure sensitivity in one quarter?
- Timestamp the first field interaction and the submit event. Median, not mean. One applicant who leaves a tab open for six days distorts an average and leaves the median untouched.
- Establish four weeks of baseline on completion rate and median time, split by channel and by product. Aggregates hide the segment that is actually stalling.
- Ship one change that removes measurable time, not three changes at once.
- Hold four more weeks, then divide the difference in completion by the difference in median minutes. That quotient is Δc.
- Multiply through the chain in Table 1 and put the annual figure in front of finance in the units finance uses.
Step two is where this usually dies, and the measurement gap is often the real finding. The calculation is easy. Owning the inputs is the work.
Common questions
What is an application abandonment cost calculator?
It is a model that turns drop-off during a loan application into an annual dollar figure. It multiplies monthly application starts by the completion lost per extra minute, then by approval rate, close rate and contribution per funded file, producing a cost per minute at current volume.
What percentage of loan applications are abandoned?
No credible published figure exists. Every number in circulation traces to vendor marketing or unattributed trade content rather than to a regulator, statistical agency or academic study. The only defensible abandonment rate is the one measured inside a specific lender's own application flow.
How do I find contribution per funded loan?
Finance already has it, usually as revenue per funded file less variable production cost. As a public anchor, the Mortgage Bankers Association reported pre-tax net production profit of $727 per loan in the first quarter of 2026, against $11,898 in total loan production expenses per loan.
Does a shorter application mean weaker credit quality?
The largest study available says no. Federal Reserve Bank of New York Staff Report 836, from February 2018, found fintech mortgage lenders were about 20% faster with default rates roughly 25% lower, and the authors found no evidence of laxer screening driving the speed.
Why measure median time instead of average time?
Averages in application data are dominated by outliers. One applicant who starts a form, walks away and returns four days later adds days to a mean and nothing to a median. Median time to complete describes the experience of a typical applicant, which is what completion rate responds to.
Carousel's intake flows are built to compress the time between application start and a decidable file. See how verification fits your flow
