Most ecommerce sellers I work with treat shrinkage like weather. It happens, you can’t really stop it, you write down the number once a year and move on. Then they’re surprised when their tax bill comes back higher than expected, or when an investor poking at their P&L asks why COGS is so noisy quarter to quarter.
Shrinkage is one of those line items that sits quietly on your books until it doesn’t. The retail industry’s average is around 1.44% of sales according to the National Retail Federation, which sounds small until you do the math on a $5M revenue brand. That’s $72,000 a year disappearing somewhere between your supplier’s dock and your customer’s doorstep.
The good news is that the IRS lets you deduct most of it if you document it properly. The bad news is that “properly” is doing a lot of work in that sentence.
Where the loss is actually happening
For DTC and marketplace sellers, “shrinkage” covers more than just shoplifting (which doesn’t really apply to most of us anyway). The buckets that matter are:
Warehouse and 3PL discrepancies are usually the biggest one. You shipped 1,000 units in, the 3PL’s system shows 987, the remaining 13 are someone’s problem to track down, and 80% of the time, nobody does. This is the single biggest leak I see in client books.
Carrier loss and damage is the next bucket. Packages disappear in transit. UPS and FedEx will reimburse you, but the process is slow and the recovery rate is well under 100%. The gap between insured value and what you actually recover is real shrinkage that ends up on your P&L.
Customer-side fraud has been getting worse for several years. “Item not received” claims, friendly fraud chargebacks, return swaps where the customer ships back a different SKU than the one they kept. The platforms are not particularly helpful with any of this.
Cycle count adjustments are the catch-all. Your inventory system says 100 units, your physical count says 92, the 8 units have to go somewhere on your books, and that somewhere is usually a vague “shrinkage” expense.
If your bookkeeper is dumping all four of these into one bucket, you’re losing visibility into what’s fixable and what isn’t. Carrier loss is fixable with claims discipline. Warehouse discrepancies are fixable with a better 3PL or better receiving controls. Customer fraud needs different tools entirely.
How most sellers account for it (and why it’s wrong)
The default approach I see is something like this: at year end, the bookkeeper compares the inventory system balance to the actual physical count, books a journal entry for the difference, and calls it shrinkage. Done.
This is fine for the books. It is not fine for the tax return.
The IRS wants documentation tied to specific events. A single year-end journal entry of $40,000 to “shrinkage” with no supporting detail is exactly the kind of thing that gets flagged in an audit. The deduction may still be valid, but you’ll spend hours reconstructing what the entry actually represented, and some of it will be disallowed because you can’t prove it.
The fix isn’t complicated. It’s:
- Book shrinkage when you discover it, not at year end
- Tie each adjustment to a SKU, a quantity, a discovery date, and a cause
- Keep the supporting evidence (3PL exception report, carrier claim docs, security footage if applicable, police report for any meaningful theft)
- For larger losses, file an insurance claim even if you don’t expect to collect, because the documentation supports the deduction either way
A 21% C-corp recovers about $0.21 per dollar of properly documented shrinkage. On that $72k example, that’s roughly $15,000 in tax. Worth getting right.
The casualty loss question
For losses from theft, fire, flood, or other discrete events, you have two paths: deduct the loss as a cost of goods sold adjustment, or claim it as a casualty loss under §165. For inventory specifically, the COGS adjustment is almost always cleaner. You’re already valuing inventory, the loss writes down the asset, and the deduction flows through gross profit.
Where it gets interesting is partial recovery. If you lose $50k of inventory in a warehouse incident and your insurance pays $35k, you can only deduct the $15k difference. People miss this and overstate the deduction, which is exactly the kind of error that triggers a notice if your numbers are otherwise tidy.
There’s also a less-known accounting method change available for retailers with documented theft losses, which lets you accrue an estimated theft loss based on historical data rather than waiting for the actual count adjustment. PwC has written about this for the larger retail clients, but it’s available to smaller sellers too if the volume justifies the analysis. We’ve used it for a couple of clients with persistent organized retail crime exposure on Amazon FBA returns.
What “rigorous documentation” actually looks like
I know “documentation” sounds like accountant speak for “more spreadsheets.” It kind of is. But the bar isn’t high. For a 3PL discrepancy, you need:
- The receiving report or ASN showing what was sent
- The 3PL’s putaway record showing what was received
- The variance reconciliation showing the difference
- Any communication with the 3PL about the variance (Slack, email, ticket — whatever)
- The journal entry tying the variance to your inventory adjustment
Most 3PLs will give you the first three. The communication trail is on you. The journal entry is on your bookkeeper. None of this is hard, but I almost never see it done end to end.
For carrier loss, the bar is similar: shipping label or tracking, customer report or proof of non-delivery, claim filed with the carrier, claim outcome, journal entry. The claim outcome is the part people skip. If you don’t file a claim, the IRS can argue you didn’t actually sustain the loss, just declined to recover it.
Where to start
If you’ve never actually looked at your shrinkage number, here’s a 30-minute exercise:
- Pull your COGS detail for the last 12 months and find the shrinkage entries
- Total them, then divide by gross sales
- Compare to the NRF benchmark (1.44%)
- If you’re meaningfully above, you have a process problem worth investigating
- If you’re meaningfully below, you probably have a documentation problem and you’re losing deductions
I’d guess about 70% of the brands I look at are in the second bucket. Their actual losses are higher than what they’re booking because nobody is tracking the small leaks, and at tax time they’re leaving money on the table.
The fix is mostly process: better receiving controls at the 3PL, a real claims discipline with the carriers, a monthly cycle count instead of annual, and a journal entry standard that captures the cause of every adjustment. None of it is fun. All of it pays for itself.
If your books look noisy in a quiet way and you suspect shrinkage is part of the story, that’s a worthwhile thing to dig into before year end. The earlier you start tracking it cleanly, the better the deduction story you can tell.
