Tax Season Mileage Log: How to
Rebuild an IRS-Ready Deduction
The IRS business mileage rate hit 72.5 cents per mile in 2026 — the highest it's been in over a decade. For a self-employed worker driving 15,000 business miles, that's a $10,875 deduction on Schedule C. But there's a catch: the IRS will disallow every dollar of it unless you can produce a mileage log with five specific elements for each trip. Every February through April, millions of independent contractors, gig workers, and small business owners face the same grim math — they drove all year but never kept a clean log. Now the filing deadline is approaching, and they need to reconstruct one from whatever records survive. This article is about how to do that correctly, what evidence the IRS and Tax Court actually accept, and how much it costs to get it wrong.
Key Takeaways
- The IRS does not negotiate on mileage: if you cannot produce a log with date, destination, purpose, and miles for each trip, the deduction is disallowed in full — not partially reduced.
- Five Tax Court cases reveal the same dividing line between keeping and losing your deduction: whether your source records existed at the time of travel or were created from memory at tax time.
- ImageToTable.ai reads your calendars, MLS schedules, odometer photos, and assignment letters into one structured table — contemporaneous records become a compiled log without retyping a single address.
What the IRS Actually Requires — Five Elements No Mileage Log Can Skip
Under Internal Revenue Code §274(d), vehicle expenses are subject to strict substantiation rules. There's no room for estimation. The Cohan rule — a 1930 precedent that lets tax courts approximate deductions when records are incomplete — explicitly does not apply to mileage. Temp. Treas. Reg. §1.274-5T(c)(2) spells out what constitutes "adequate records": an account book, diary, log, statement of expense, trip sheet, or similar record — combined with documentary evidence — that establishes each element of the use.
For each business trip, five elements must be substantiated:
- Amount — mileage for that specific business use. Round numbers (e.g., "about 30 miles") are a red flag. GPS-measured or odometer-based distances carry more weight.
- Time — the date of each trip. A weekly log is explicitly considered timely under IRS Publication 463. A log reconstructed months later from memory is not.
- Place — the business destination, at minimum the city or town. "Client meeting" without a location is too vague. You need enough specificity that an auditor could verify the distance matches.
- Business purpose — what was the business reason for the trip. "Travel for work" doesn't cut it. "Property showing at 1420 Park Ave" does.
- Business relationship — who you met with, if applicable.
The regulation also requires you to report total mileage for the year — business, commuting, and personal — not just business miles. The IRS needs to see the denominator to evaluate the business-use percentage. A log that shows 12,000 business miles with no mention of total mileage invites the question: how many personal miles were driven on the same vehicle?
IRS Publication 463 states that a weekly-updated log is explicitly considered timely. The line between "acceptable late reconstruction" and "rejected reconstruction" is not whether you wrote it down the same day — it's whether your records existed at or near the time of travel and were later compiled, versus created from memory at tax time.
The Tax Court Record — Who Kept Their Deduction and Who Lost It
Tax Court opinions are the best available guide to what the IRS actually challenges and what judges actually accept. The pattern across multiple cases is unambiguous: the taxpayers who win keep some form of contemporaneous record — a calendar, a logbook, an app — and update it regularly. The taxpayers who lose have nothing but memory and a reconstructed spreadsheet built at tax time.
Here are the cases that define the line:
| Case | What the taxpayer had | Outcome | Lesson |
|---|---|---|---|
| Ressen v. Comm'r T.C. Sum. Op. 2015-32 | Weekly-updated calendar with mileage + daily logbook recording start/end odometer readings | $28,504 deduction upheld | A calendar updated weekly plus a logbook with odometer values is sufficient. The records don't need to be elaborate — they need to be consistent. |
| Renner v. Comm'r T.C. Memo. 2015-102 | No log during the year. Reconstructed a log after audit began. No customer names, no addresses, no business purpose. | $22,000+ deduction disallowed entirely | A log built from memory after the IRS opens an audit has zero credibility. The court won't even entertain it. |
| Velez v. Comm'r T.C. Memo. 2018-72 | iPad calendar used to reconstruct mileage after the fact | Deduction disallowed | Even a digital calendar — if it wasn't maintained contemporaneously and lacks business purpose details — won't survive scrutiny. |
| Chappell v. Comm'r T.C. Sum. Op. 2024-2 | Partial records. Actual expense method rejected, but standard mileage method accepted for documented trips backed by tracking app data. | Partial deduction allowed | Even if records are incomplete, the court may allow a deduction for trips that are properly documented. A partial log beats no log. |
| Khan v. Comm'r T.C. Sum. Op. 2025-5 | Reconstructed spreadsheets and generalized explanations. No contemporaneous log. | Deduction disallowed | "Generalized explanations and reconstructed spreadsheets" is the exact phrase the court used to reject the claim. This is the most common taxpayer mistake. |
The pattern is clear enough to state as a rule: records that existed at or near the time of travel, even if partial and imperfect, can be compiled into an acceptable log. Records created from memory at tax time cannot. Ressen won with a calendar and a logbook. Renner, Velez, and Khan lost with memory and spreadsheets. The difference wasn't the quality of their tax preparer — it was the existence of contemporaneous source material.
What Survives an IRS Review — The Evidence Hierarchy for Mileage
Not all evidence is equal. The regulation is explicit: "the probative value of written evidence is greater the closer in time it relates to the expenditure or use." This creates a practical hierarchy of what holds up and what doesn't:
Evidence Strength — Strongest to Weakest
- GPS-tracked mileage logs with timestamps — An app that records trips as they happen. Chappell (2024) confirms these are accepted by the Tax Court. The gold standard, but useless if you didn't use one.
- Odometer photos with EXIF timestamps — A photo of your dashboard at the start and end of a shift is contemporaneous by definition. The EXIF data records the date and time. The GPS metadata records the location. This is strong evidence that survives scrutiny.
- Calendar entries + appointment records — A calendar showing "Property showing: 1420 Park Ave, 2:00 PM" on a specific date establishes time, place, and business purpose. Ressen's accepted calendar was essentially this.
- Third-party documents showing addresses — MLS showing schedules, Uber trip summaries, assignment letters, client invoices with addresses. These don't show mileage but they establish where you went and why. Combined with route distances (Google Maps), they create a defensible record.
- Receipts from business destinations — Gas receipts, parking receipts, meal receipts at business locations. They establish presence at a location on a date. Weak on their own, but corroborating when paired with other records.
- Memory + reconstructed spreadsheet — The Renner/Khan outcome. This is not evidence. The Tax Court has rejected this pattern in every case where it appeared alone.
This hierarchy is the key to tax-season reconstruction. If you're sitting in March with no mileage log, your first step isn't to open a blank spreadsheet and start guessing. It's to inventory what contemporaneous evidence already exists — because that's what determines whether your reconstructed log will survive or collapse under review.
Scenario 1 — The Rideshare Driver: Turning a 1099-K Into a Full Mileage Log
Uber and Lyft send drivers an annual tax summary that includes online miles — the distance driven with a passenger in the car. On the surface, this looks like usable mileage data. It's not the whole picture.
The Gridwise Annual Gig Mobility Report found that rideshare drivers who rely solely on app-provided trip distances miss 30 to 40 percent of their deductible miles. These are the deadhead miles — driving from the last drop-off to the next pickup area, repositioning to a surge zone, returning to a busy part of town after a long trip. A driver who logs 48,000 total miles in a year may only see 26,000–29,000 reflected in Uber's online-mile summary. The other 19,000–22,000 miles are equally deductible — IRS Topic no. 510 confirms that all business-use miles count, including miles driven between trips while the app is on — but they're invisible in the app's report.
At the 2026 rate of 72.5 cents per mile, missing 19,000 miles means leaving $13,775 in deductions on the table. For a driver in the 22% federal bracket paying 15.3% self-employment tax, that's roughly $5,140 in extra tax paid — on miles that were legally deductible.
Reconstruction path for rideshare drivers:
For a deeper look at how odometer photos specifically bridge the gap between "I have evidence" and "I have an IRS-ready log," see our guide on turning odometer photos into a Google Sheets mileage log.
Scenario 2 — The Real Estate Agent: MLS Showing Schedules as Route Evidence
Real estate agents drive more business miles than almost any other self-employed profession. A full-time agent logs 12,000–16,000 business miles per year — showings, open houses, inspections, closings, client meetings, brokerage runs. At the 2026 rate, that's $8,700–$11,600 in deductions. The problem: most agents are terrible at mileage tracking, and their brokerages don't track it for them.
What most agents don't realize is that they already have one of the best mileage-substantiation documents in existence: their MLS showing schedule. Every confirmed showing in the MLS includes the property address, the date, and the time. That's three of the five IRS-required elements — place, time, and business purpose — documented by a third-party system that neither the agent nor the IRS can alter.
Reconstruction path for real estate agents:
An agent who did 300 showings in a year at an average of 15 miles round-trip per showing has 4,500 documented miles from the MLS alone. Add open houses, inspection visits, and closing trips, and the documented total can easily reach 8,000–10,000 miles — all backed by third-party records the agent already has.
Scenario 3 — The Traveling Nurse: Assignment Letters to Tax-Ready Facility Mileage
Traveling nurses and allied health professionals face a different mileage problem than rideshare drivers or real estate agents. Most are W-2 employees of their staffing agency, which means unreimbursed employee business expenses — including mileage — are not deductible on federal returns under current tax law (the Tax Cuts and Jobs Act suspended the miscellaneous itemized deduction for employees through 2025). However, there are two critical exceptions where mileage tracking still matters:
- 1099 contractors — a small but real segment of traveling healthcare workers operate as independent contractors. They file Schedule C and can deduct mileage at the standard rate.
- State-level deductions — several states, including California, New York, Alabama, Hawaii, and Arkansas, still allow employee business expense deductions on state returns. A nurse who worked assignments in multiple states may have deductible mileage on one or more state filings even if the federal deduction is unavailable.
For those who can claim mileage, the reconstruction path starts with assignment documentation the nurse already has:
For nurses who rotate through multiple facilities during a single assignment (floating between clinic locations, covering shifts at affiliated hospitals), the assignment letter typically lists all expected work sites. Each facility-to-facility trip is a separate deductible trip — and the assignment letter documents the addresses.
The Cost of Getting It Wrong — Dollar Math That Makes the Case for Accuracy
The IRS doesn't negotiate on mileage. Unlike other Schedule C expenses where an auditor might allow a partial amount, mileage deductions subject to §274(d) follow a binary pattern in Tax Court: substantiated or disallowed. Here's what that means in dollars across the three scenarios:
| Scenario | Estimated Business Miles | Deduction at 72.5¢/mile | If 20% Is Disallowed | If 100% Is Disallowed |
|---|---|---|---|---|
| Rideshare driver Full-time, 48,000 total miles | 32,000 | $23,200 | $4,640 lost deduction ≈ $1,730 extra tax (22% + SE) | $23,200 lost ≈ $8,650 extra tax |
| Real estate agent 300 showings + misc. trips | 15,000 | $10,875 | $2,175 lost deduction ≈ $810 extra tax | $10,875 lost ≈ $4,050 extra tax |
| Traveling nurse (1099) 3 assignments, 4 facilities | 8,000 | $5,800 | $1,160 lost deduction ≈ $430 extra tax | $5,800 lost ≈ $2,160 extra tax |
Tax impact calculated at 22% marginal federal rate + 15.3% self-employment tax (37.3% combined). Actual amounts vary by total income, filing status, and state tax rates.
These numbers explain why the IRS scrutinizes mileage so closely. A $10,875 deduction claimed without adequate records is $10,875 the IRS can disallow in full — plus penalties and interest if the understatement is substantial. The cost of getting it wrong isn't just the lost deduction. It's the lost deduction plus the audit risk plus the time and professional fees spent defending a position you could have documented in 5 minutes a week.
15,000 business miles = $10,875 deduction. Missing 20% due to poor records = $2,175 lost. At a 22% marginal rate plus 15.3% self-employment tax, that's $810 in extra tax — for miles you actually drove.
How ImageToTable.ai Turns Scattered Records Into One Mileage Table
Reconstructing a mileage log from partial records creates a predictable bottleneck: you have evidence spread across multiple formats — screenshots of calendar entries, photos of your odometer, PDFs of assignment letters, exported CSV files from the MLS or Uber — but turning each piece into a row in a mileage spreadsheet is painfully manual. You're reading an address from one screen, typing it into another, switching to Google Maps to get the distance, typing that too. Multiply by 300 showings or 250 driving days and the reconstruction project starts to look like a full-time job.
ImageToTable.ai addresses this bottleneck by reading information from your existing records and placing it into a structured table — no manual transcription required. The tool uses Custom Column Extraction: you define the columns you need (Date, Destination, Business Purpose, Miles, Odometer Start, Odometer End) and the AI locates the corresponding information on each document by understanding what the data means, not by matching a fixed template.
For a real estate agent reconstructing mileage from MLS schedules: upload screenshots or PDFs of the showing schedule, define columns for "Property Address," "Showing Date," and "Showing Time." The AI reads each entry and populates the table. The agent then adds a "Miles" column (pulled from Google Maps route distances, calculated once per route pattern) and the log is substantially complete. For a rideshare driver with odometer photos: upload the photos, define columns for "Date" and "Odometer Reading," and the AI extracts the digits from each photo — freeing the driver to focus on adding business purpose descriptions and route distances rather than transcribing numbers.
With batch processing, you upload all records at once — 50 MLS screenshots, 100 odometer photos, a year's worth of assignment letters — and get a single merged table. The output exports as Excel (XLSX) or CSV, ready to serve as the foundation of your mileage log. The tool doesn't replace the work of verifying distances or adding business purpose descriptions; it eliminates the transcription step that turns reconstruction from a weekend project into a multi-week ordeal.
To learn how odometer photos specifically work within this workflow, see our guide on batch-processing 30+ odometer photos into one tax-ready mileage log.
Frequently Asked Questions
Can the IRS accept a reconstructed mileage log?
Yes, but only if the reconstruction is based on contemporaneous source material — not memory. The IRS distinguishes between compiling existing records into a log (acceptable) and creating a log from memory at tax time (not acceptable). The Tax Court cases confirm this distinction: Ressen's calendar-and-logbook reconstruction was accepted because the underlying calendar was maintained weekly. Renner's post-audit memory reconstruction was rejected. If your reconstruction is built on records that existed at or near the time of travel, it can survive scrutiny.
How far back can I reconstruct a mileage log?
Generally three years — the standard IRS statute of limitations. If you understated income by more than 25%, the IRS can go back six years. State rules may differ. The key constraint isn't the time window — it's whether you still have the source records from those years. Calendar entries from 2023 are useful. Your memory of where you drove in January 2023 is not.
What's the difference between commuting and business mileage?
Commuting — driving between your home and your regular place of business — is not deductible. Business mileage is everything else: driving from your office to a client meeting, from one job site to another, from your home office (if it qualifies as your principal place of business) to a work location. For rideshare drivers, miles driven with the app on and available for trips count as business miles, including deadhead miles between rides. The IRS definition is in Publication 463, Chapter 4.
Do I need odometer readings for every trip?
No. The IRS accepts mileage calculated from reliable sources — Google Maps route distances, GPS tracker logs, odometer readings. What matters is that the mileage is verifiable, not how you measured it. The regulation only requires that you substantiate "the amount of each business use" — it doesn't specify that odometer readings are the only acceptable method. However, if you claim the standard mileage rate, you do need to know your total mileage for the year (business + personal), which requires at least a beginning-of-year and end-of-year odometer reading or a reliable estimate.
What if I have some records but not all five elements for every trip?
Partial records are better than no records — and Chappell v. Commissioner (2024) confirms the Tax Court will allow a deduction for trips that are adequately documented even if other trips in the same year are not. The strategy is to claim mileage only for the trips you can substantiate, not to fabricate records for the ones you can't. A partial deduction on a solid foundation beats a full deduction on a shaky one. An auditor who sees honest gaps is more likely to accept the documented portion than an auditor who sees a suspiciously perfect log.
Can I use Google Maps Timeline as mileage evidence?
Google Maps Timeline is useful as a corroborating source — it shows where you were and when — but it shouldn't be your only evidence. The GPS data is subject to the same drift issues as any GPS tracker, and the IRS views location history as supporting documentation rather than a primary mileage log. Use it to verify route distances and fill gaps in calendar-based records, but pair it with other contemporaneous evidence like odometer photos, calendar entries, or appointment records.
What happens if I get audited and my mileage log is rejected?
If the IRS disallows your mileage deduction during an audit, the consequences include: (1) the deduction is reversed, increasing your taxable income for that year; (2) you owe the additional tax plus interest from the original due date; (3) if the understatement is substantial (more than $5,000 or 10% of the correct tax), a 20% accuracy-related penalty applies. This is why the reconstruction effort is worth doing carefully — the cost of getting caught is significantly higher than the cost of doing it right.
Your odometer photos, calendar entries, and route maps are the raw material. The spreadsheet that turns them into a deduction doesn't have to be built by hand.
Try ImageToTable.aiNo sign-up required for your first file