# The Variance Audit: Why Your Real Margins Don't Match Your Estimated Margins (And Where the Gap Hides)

_By GradeThread Team · Published August 23, 2026_

> Your spreadsheet says 42% margin. Your bank account says 28%. Here's the variance audit that finds exactly where the gap hides.

# The Variance Audit: Why Your Real Margins Don't Match Your Estimated Margins (And Where the Gap Hides)

The gap between your estimated margin and your actual margin almost always comes from five places: payment processing fees you didn't model, shipping cost overages, promoted listing spend, partial refunds that don't return the platform's commission, and returns tied to condition disputes. Run a side-by-side comparison of comp-based estimate versus real payout on 20 recent sales, and the leak shows up in one or two of those five lines within an hour.

Most resellers price off a simple formula: sold comp price, minus item cost, minus "eBay's cut," equals margin. That formula is directionally right and specifically wrong. It's right that fees eat a big chunk. It's wrong about which fees, how much, and when they hit. Run that formula on 100 items a month and the error compounds into a number that doesn't match what actually lands in your bank account — often by 10 to 15 margin points.

## The Gap, in Real Numbers

Take a vintage Carhartt chore coat. Sourced for $12. Comped against 15 sold listings on eBay at an average of $85. Quick mental math: $85 minus $12 minus "15% for eBay fees" (about $12.75) leaves an estimated profit of $60.25 — a 71% margin on sale price.

Here's what actually happened on the payout report:

| Line item | Estimated | Actual |
| --- | --- | --- |
| Sale price | $85.00 | $85.00 |
| Item cost (COGS) | $12.00 | $12.00 |
| eBay final value fee | $12.75 (15% flat guess) | $11.56 (13.6% + $0.30) |
| Promoted listings fee (6% ad rate) | $0 | $5.10 |
| Shipping label (charged $9.99, actual cost) | $0 assumed break-even | $2.40 overage (weight rounded up a class) |
| Packaging materials | $0 | $1.10 |
| Poly mailer + tag replacement | $0 | $0.60 |
| **Net profit** | **$60.25 (71%)** | **$52.24 (61%)** |

Ten margin points on one item is $8.01. Multiply that pattern across 400 sales a month and you've got a $3,200 gap between the P&L you think you're running and the one your bank statement shows. That's not a rounding error — that's a full week of sourcing budget disappearing into line items nobody logged.

## Where Estimated Margin Math Breaks Down

The formula breaks in the same five places for almost every reseller we talk to. In order of how often they show up in a variance audit:

- **Payment processing isn't flat.** eBay's managed payments, Poshmark, and Mercari each fold processing into their take rate differently, and it shifts slightly by payment method the buyer chooses. A flat "15%" estimate misses the real blended rate, which usually runs 12.9%–13.6% on eBay clothing and 20% flat on Poshmark over $15.
- **Promoted listings are invisible until payout.** If you run Promoted Listings Standard at even a modest 5–8% ad rate, that's a direct hit to margin that never shows up in a sale-price-minus-fee estimate, because most sellers budget it separately (if at all).
- **Shipping estimates assume perfect weighing.** Charge $9.99 for shipping, estimate the label costs $9.99, call it a wash. In reality, dimensional weight rounding, wrong package type, or an extra ounce over a weight class tier can add $1.50–$4.00 per package that never gets logged against the sale.
- **Partial refunds don't refund the platform's cut.** Poshmark and Mercari keep their full commission on the original sale price even when you issue a partial refund to the buyer — you eat the entire deduction. That's a margin hit an estimate never anticipates because it assumes full-price, no-refund sales.
- **Returns tied to condition disputes cost more than the refund line.** A return isn't just lost revenue — it's the return shipping label, the item going back into inventory at a lower resale value, and the time to reprocess it. None of that lives in a per-item margin estimate.

## The Variance Audit: A Step-by-Step Procedure

Run this once a month on a sample of your sales. It takes about 45 minutes for 20–30 items and tells you exactly which line is bleeding.

1. Pull your last 30 days of completed sales from each platform's payout report (eBay Payments, Poshmark's My Sales, Mercari's Transaction History) — not your listing tool's projected margin, the actual settled payout.
2. For each sale, log five numbers: sale price, item cost, platform fee actually charged, shipping cost actually paid, and any refund or deduction issued.
3. Calculate actual net margin per item: (sale price − item cost − platform fee − shipping cost − refund) ÷ sale price.
4. Pull the estimated margin you used when you priced the item — from your comping notes, spreadsheet, or listing tool's projection.
5. Subtract actual margin from estimated margin for each item. This is your variance, expressed in margin points, not dollars.
6. Sort the list by variance size, largest gap first. The top five items usually share a root cause — same fee type, same platform, same category.
7. Assign each variance to one of the five leak categories above, then total the dollar impact by category across the full sample.
8. Fix the estimate formula for the category with the biggest total leak before you run the audit again next month.

Step 7 is where most people stop early and miss the point. A single $8 variance on one jacket is noise. A $3-per-item average variance across 60 sales in the same category, all coming from the same fee type, is a formula error you can fix once and stop bleeding on every future listing.

## Building the Monthly Margin Variance Audit Spreadsheet

You don't need software to start. A spreadsheet with these columns catches 90% of the gap:

- SKU / item ID
- Category
- Platform
- Sale price
- Item cost (full COGS, not just purchase price)
- Estimated fee %
- Actual fee $ (from payout report)
- Estimated shipping cost
- Actual shipping cost (from carrier receipt, not the flat rate you charged)
- Refund/deduction $
- Estimated margin %
- Actual margin %
- Variance (points)
- Leak category (processing / promoted / shipping / refund / return)

Sort by variance and leak category monthly. Once you've identified which leak is biggest — for most sellers it's promoted listings or shipping weight overages — you correct the estimate formula going forward and re-audit next month to confirm the gap closed. If it's still wide after two fixed months, the leak is probably operational, not mathematical: mis-measured items, wrong category fees, or a returns problem specific to one product line.

## The Return Line Item: How Condition Disputes Hide Inside Your Gap

Returns are the leak category that costs the most and gets tracked the least, because the loss doesn't show up as a single fee line — it shows up as a chain of smaller costs across weeks. A buyer files "item not as described," you eat return shipping, the item comes back, and it often no longer qualifies for the condition tier you sold it at.

This is where condition grading language matters for your P&L, not just your listing copy. If you sold a jacket as Excellent and a buyer disputes based on a fabric flaw you didn't catch — a repair affecting Structural Integrity, a stain affecting Cosmetic Appearance, or a smell affecting Odor & Cleanliness — that item frequently comes back and has to be relisted a tier down, at Very Good or Good, at a lower comp price. The original estimate assumed one sale at full grade value. The actual outcome is a return cost plus a second sale at a discount. That's a two-line hit your original margin math never modeled.

Track your return rate by condition tier for a quarter (NWT, NWOT, Excellent, Very Good, Good, Fair, Poor) and you'll usually find disputes cluster in one or two tiers — most often items sold as Excellent that had a defect closer to Very Good. Standardizing how you grade against the five factors — Fabric Condition, Structural Integrity, Cosmetic Appearance, Functional Elements, Odor & Cleanliness — before listing closes that gap at the source, instead of after the return has already cost you shipping, relisting time, and a lower comp.

## What a Healthy Gap Looks Like

You will never hit zero variance — comps shift, buyers negotiate, weight classes round unevenly. A variance of 3–6 margin points between estimate and actual, averaged across a monthly sample, is normal operating noise. A variance above 8–10 points, especially if it repeats month over month in the same category, is a formula problem or a return problem, and it's costing you real money every cycle you don't fix it.

Run the audit quarterly at minimum, monthly if you're doing more than 100 sales a month. The math takes less time each cycle once your spreadsheet columns are set, and the fix — fee-formula correction or grading-accuracy correction — usually pays for the audit time in the very next batch of listings.

If returns tied to condition disputes are one of your leak categories, that's a fixable input, not a cost of doing business. Grade one item against the five factors before you list it and see whether your description would have survived a buyer dispute — that's the cheapest variance-closing move you can make this week.

---

Canonical: [https://gradethread.com/blog/variance-audit-actual-vs-estimated-margin-gap-reseller](https://gradethread.com/blog/variance-audit-actual-vs-estimated-margin-gap-reseller)