Your time log
Your numbers
One row per time entry, dates like 2026-03-14 or 3/14/2026, hours as 1.5 or 1:30. Months with no entries between a client's first and last month count as zero hours. Each client needs at least 6 months.
Upload or paste your time log, then click Analyse.
Result
Monthly hours, month by month
All clients at this retainer
| Client | Months | Mean h | Median h | 80th pct h | Rate at median | Rate at 80th | Underpays | Retainer for target | Hold-out |
|---|
How it is calculated
Monthly hours: your entries are summed by client and calendar month, from the client's first month to its last, and a month with no entries is zero hours. The median and 80th percentile are interpolated between ranked months. The range is the middle 80% of 300 seeded resamples of those months (a fixed seed, so the same log always gives the same range).
Under the retainer: effective hourly rate is the retainer divided by the hours worked, shown at the median month, the 80th-percentile month and your busiest month. Break-even hours are the retainer divided by your hourly rate. The retainer underpays in a month when those hours at your hourly rate would have come to more than the retainer; the share is counted over your months, with a bootstrap range. The retainer for a target rate is that rate times the hours at the coverage percentile, so it would have met the target in that share of your months.
Hold-out: with 9 or more months, the first two thirds set the percentile and the last third is held back. If average hours moved by more than 25% the history is called unstable. Fewer than 12 months is called weak evidence even when it passes. Under 6 months for a client, nothing is computed.
Not modelled: scope creep, unused-hour rollover, late payment, taxes, the client's willingness to agree, and work you did not log. Unlogged hours make a retainer look better than it was. Overlapping percentiles from a handful of months are rough, and 6 to 12 months can miss a busy season. Unverified: tested on hand-counted logs and simulated hours with a known distribution, not on a published dataset.
Common questions
Is a retainer better than hourly billing?
It depends on how many hours the client really uses in a month. A retainer pays the same however many hours you work, so it beats hourly in the light months and loses in the heavy ones. This tool counts, from your own time log, how many past months would have paid more under the retainer you propose, and what your effective hourly rate would have been in a typical, a busy and your worst month.
How do I price a retainer from my time log?
Take the hours you actually spent on that client each month, pick how many months you want the retainer to cover (80% is common), read the hours at that percentile, and multiply by the hourly rate you want to earn. Paste one row per time entry and the tool does it, with a range from 300 seeded resamples of your months.
What is effective hourly rate on a retainer?
The monthly retainer divided by the hours you worked that month. A 1,500 dollar retainer is 100 dollars an hour in a 15-hour month and 75 dollars an hour in a 20-hour month. The tool shows it at the median month, the 80th-percentile month and your single busiest month.
What is the break-even number of hours?
The retainer divided by your hourly rate. Below that many hours in a month the retainer pays more than hourly billing would have; above it, less. The tool also shows the share of your past months that fell above it.
How much history do I need?
At least 6 months for each client; a client with fewer is left out. Months with no entries between a client's first and last month count as zero hours. With 9 or more months the tool also holds the latest third back and checks whether the earlier months predicted it; with fewer it says the evidence is weak.
Why does it say my hours are unstable?
The percentiles are computed from the earlier months and compared with the latest ones. If average monthly hours moved by more than a quarter, a client may have ramped up or wound down, and old months may not describe the next year. Read the numbers as history, not a forecast.
Is my time log uploaded anywhere?
No. Client names, hours and amounts are read and analysed in this browser tab. Nothing is sent to a server, there are no accounts or cookies, and the tool does not store your log. It works offline once the page has loaded.
Does it replace a contract or advice?
No. It describes the hours in the log you supply. It does not model scope creep, rollover or unused hours, late payment, taxes or what a client would agree to, and it is not legal, tax or financial advice.