A rejected mileage log doesn’t usually mean the IRS thinks you didn’t drive.
It means they think you can’t substantiate what you drove.
And in an audit, “I have an app report” is not the same thing as substantiation.
The moment it got rejected: what the examiner actually said
The letter wasn’t dramatic. That’s part of what makes it nasty.
It was a correspondence audit for Schedule C vehicle expenses. They asked for three things in plain language:
- the mileage log
- total miles for the year
- “documentation to support business purpose”
I sent what I thought they wanted: a PDF export from my mileage app (date + miles + start/end as city names) and a spreadsheet summary with monthly totals.
Two weeks later, I got the response that matters in your stomach more than your mailbox:
“Insufficient substantiation. Entries do not establish a business purpose. Total miles not verified.”
Then they listed a few examples from my own log. Not hypotheticals. Specific lines.
One was like: 03/14 – 18.6 miles – ‘Work’ – [City] to [City].
The examiner’s note basically said: What work? Where exactly? Who? Why?
Another was a batch of trips that started at my home, ended at the “Office,” and had no other details. I don’t have an “office” anyone can point to. It was my storage unit and a couple of client sites.
They weren’t calling me a liar. They were saying my record could’ve been created last week.
My tax bill jumped because they proposed disallowing a chunk of the deduction. Interest too. And the real cost was the scramble: nights trying to remember what I did on random Tuesdays eight months ago, digging through old texts like a detective.
What was “missing” (the specific fields they cared about, not the blog checklist)
Most advice says “date, miles, destination, business purpose.” True, but too vague to be useful when you’re staring at a rejection.
Here’s what the examiner kept circling back to in my case.
They wanted a business purpose that reads like a sentence, not a label
“Work.” “Client.” “Errands.” Those were basically poison.
What they wanted was the thing that made the trip business.
Not: “client.”
More like: “drop off signed contract to Smith Plumbing,” “pickup materials for 124 Oak remodel,” “meet CPA to review Q3 estimates.”
If you can’t say it in one line, it tends to fall apart under questions.
They wanted locations that could exist in the real world
My export had a lot of:
- “Downtown”
- just a city name
- or a saved place name that meant something only to me
The examiner didn’t say “you need full addresses for every trip,” but they did push back on anything that couldn’t be independently understood.
One back-and-forth was literally: “What is ‘Warehouse’?”
It was a Home Depot.
I had named it “Warehouse” in my app years ago.
They cared about total miles for the year (and how I got that number)
This surprised me. I thought the whole fight would be about business miles.
Nope. They wanted the total miles and how I verified it.
I didn’t have a clean starting odometer reading for Jan 1. I had an oil change receipt in February and a state inspection printout in November, but nothing that clearly boxed in the year.
That gap made my business percentage look like it was floating in space.
Install the MyCarTracks app and start tracking your mileage now
The mistake that made my whole log look reconstructed
I let “unknown trip” entries pile up.
Then, because I’m a functioning adult with bills and work, I tried to clean them up in one sitting months later.
That single decision made everything look worse.
When you add purpose notes long after the fact, you don’t add detail – you add generic words. That’s what I did. “Work.” “Supplies.” “Client meeting.”
From the examiner’s point of view, it’s the same vibe as someone trying to reverse-engineer a log to match a deduction.
Blunt truth: a clean PDF can still be a weak record.
What I pulled together to salvage part of the deduction (and what still got disallowed)
I couldn’t rebuild an entire year perfectly. So I stopped trying.
Instead, I tried to save what I could save – by making a smaller set of trips defensible.
Here’s what actually moved the needle.
1) I anchored the year’s total miles with ugly, real-world documents
I found:
- the closest service receipt to the beginning of the year with an odometer reading
- the closest one near the end
- plus a photo of my odometer I’d taken for a marketplace listing (accidentally useful)
It wasn’t perfect. But it gave the examiner something that wasn’t “trust me.”
2) I picked problem trips that the examiner already flagged and rebuilt just those
They had highlighted a few entries that looked like commuting or personal driving.
One was a repeated route from home to a consistent address in the morning. In my head, that was “first job site.” On paper, it looked like commuting to a regular place of business.
I pulled:
- The invoice for that client
- The text thread confirming the date/time
- and a calendar entry that showed it was a job site, not my own office
Did it save every mile? No.
A handful still got disallowed because I couldn’t tie them to anything concrete. Some were legitimately mixed-use trips where I “kind of” did a work thing and “kind of” did personal stuff. I didn’t have clean splits.
And when you don’t have clean splits, you lose.
3) I stopped over-explaining and started labeling
This sounds small, but it mattered.
Instead of sending a narrative essay, I sent a short list like:
- Trip 3 (3/14): matches Invoice #1842, materials purchase receipt attached
- Trip 9 (5/02): matches calendar entry “124 Oak final walk-through” + client email
Examiners are reading fast. You’re not trying to impress them – you’re trying to make the file easy to close.
The subtle insight: fewer miles can be easier to defend than “maximizing”.
Common advice is basically: track everything and claim the maximum.
Here’s the part people don’t say out loud:
The mile that hurts you is the one you can’t explain cleanly.
If you stuff your log with fuzzy trips because “it probably counts,” you don’t just risk losing those miles.
You risk making the whole method look unreliable.
What ended up working better for me was conceding the gray stuff and defending the clean stuff. It felt like losing. It also reduced the examiner’s suspicion.
How do I keep next year’s log from turning into a reconstruction project
I don’t “track mileage.” I close it out.
- Once a week (usually Sunday night), I scan for anything labeled “unknown” or “other,” and I fix the purpose while I still remember what it was.
- Once a month, I take an odometer photo. Not because it’s fun, but because total miles is the trap people fall into.
- Every quarter, I export my log and store it with my tax folder. I learned the hard way that an app report you can’t export cleanly (or can’t annotate later) turns into panic when you’re under a deadline.
In my case, part of the mess was that my original app export didn’t include the level of purpose detail I needed, and editing/annotating after the fact was clunky enough that I just… didn’t do it until it was too late.
After that audit, I switched my own workflow to MyCarTracks mainly because the export/backup habit was easier for me to keep up with, and I could add notes while the trip still had context. That didn’t “solve audits.” It just made my records look like they were created when the driving happened – which is what the IRS is testing for.
If your IRS mileage log was rejected, don’t start by arguing about the total miles.
Start by making a smaller, believable story: specific purpose, believable locations, and a way to anchor total miles in something outside your own spreadsheet.
