How we calculate the numbers

1. How we calculate Real Profit

“Real Profit” is what lands in your pocket after every fee, expense, and return has been accounted for. We compute it per-order from TikTok Shop API data, then sum across your date range.

Where the numbers come from

• Order revenue, referral fee, transaction fee, affiliate commission, fulfillment fee, campaign fee, refund admin fee — pulled directly from the TikTok Shop API via /order/202309/orders and /finance/202507/orders/unsettled. We do not estimate these — they are the exact amounts TikTok charged you.
• COGS — whatever you entered in the Products section. If you haven't entered COGS for a SKU, we leave it at zero and lower your confidence score accordingly (see section 5).
• Operating expenses — whatever you entered in Settings → Expenses. We subtract them proportionally across the selected date range (daily, weekly, monthly, yearly, one-time).
• Ad spend — pulled from the TikTok Ads API if you've connected that account. If not connected, ad spend is treated as zero and the Real Profit tooltip notes that explicitly.

What is not included

• Chargebacks and disputes that haven't settled yet — TikTok reports these only once finalized.
• Off-platform expenses you haven't entered (tax, legal, insurance, etc.). Add them in Settings → Expenses and they'll roll in.
• Sales tax — it's pass-through and not part of your profit.

Arithmetic

All monetary math happens in integer cents (or the local-currency subunit), not floating-point dollars. Rounding happens once, at the display layer. This eliminates the drift you see in spreadsheet-based profit trackers that do everything in floats.

2. How we calculate Return Cost

A $30 refund doesn't cost you $30. For each return we sum six distinct components, every one of which is a real line item on your TikTok statement or an expense we know about:

Worked example — $30 refund on a $30 cosmetic return

• Refund amount: $30.00
• Refund admin fee (20% × $30 referral fee × 6% = $0.36, capped): $0.36
• Return shipping (seller-paid): $6.50
• Lost affiliate commission (15% of $30): $4.50
• Lost transaction fee (already reversed — net 0 here): $0.00
• Labor allocation (5 min @ $20/hr): $1.67
• Total return cost: ~$43.03 (1.43x the refund)

Your actual multiplier will land in the 1.3–2.0x range depending on which components apply to your category. We show the number per-return in the returns detail view so you can see exactly which components drove it.

3. How we calculate Money Saved

The “Money Saved” counter on your dashboard is the sum of three distinct sources:

1. Rejected returns — when you (or an auto-rule) reject a return, we count the refund amount at face value. We do not apply the 1.3–2.0x multiplier here because we don't want to overstate savings.
2. Deadline saves — when you act on a return within 12 hours of the TikTok auto-approval deadline and the outcome is rejection or partial refund, we attribute that save to Foregrade's alerting.
3. Fee anomaly corrections — when our fee anomaly detector surfaces a TikTok over-charge and you successfully dispute it, we add the recovered amount.

Attribution caveat — read this

“Money Saved” is an upper bound, not a causal claim. We cannot distinguish “Foregrade made you reject this return” from “you would have rejected this return anyway.” We include every rejection because most sellers tell us they weren't reviewing returns at all before Foregrade — but if you were already diligent, the number will overstate your lift from Foregrade specifically. Use it as a rough measure of the value flowing through the dashboard, not as ROI.

If you want a tighter number, compare your reject rate in the 30 days before Foregrade to the 30 days after — the delta is the clean causal estimate.

4. Why we don't guarantee prevention rates

Our ROI calculator uses typical bands of 30% (Starter), 40% (Growth), and 45% (Scale) when estimating how many returns you'd catch. These are observed ranges across our seller base — they are not guarantees, and we intentionally do not commit to a specific rate.

Your actual rate depends on:

• Your product category (fashion returns differently from electronics)
• Your return policy (tight vs. lenient)
• How consistently you use auto-rules
• Your baseline process before Foregrade (high-diligence sellers see less lift)
• Your return-reason mix (buyer-remorse vs. damaged vs. not-as-described)

In practice we've observed prevention rates anywhere from ~20% to ~50% depending on the seller's situation. The calculator is an estimation tool, not a commitment.

5. What the confidence score means

The “data accuracy” badge on your dashboard is a single number between 0 and 1, representing how much of the input Foregrade needs to compute your Real Profit is actually present. It is about data completeness, not profit accuracy.

We show a green badge at ≥0.85 (most inputs present), an amber badge between 0.60 and 0.85 (some inputs missing — the number is directional), and a red badge below 0.60 (the number has enough holes you shouldn't trust it yet). The biggest input to this score is COGS coverage — enter costs for your top SKUs and the badge jumps fast.

Note that a 0.85 confidence score does not mean your profit is within 15% of the truth — it means 85% of the input fields Foregrade expects are populated. The profit itself is exact given the inputs; the score just tells you how many inputs made it in.