The 9 Silent Profit Killers in Every Online Store

Most e-shop owners we work with at Miranda Media think they're profitable. The bank account says otherwise. Below are nine cost lines that are small in isolation and devastating in combination — the silent killers we see eating margins every day in 7- and 8-figure stores.

1. Payment-gateway fees and chargebacks#

Stripe, Shopify Payments and PayPal all advertise a clean 2.9 % + $0.30 per transaction. Add a 1–3 % cross-border surcharge for foreign cards, factor in interchange fees baked into the rate, and the effective cost is 50–100 % higher than the headline (Toolecommerce, 2025).

Then there are chargebacks. Mastercard's 2025 research puts the global chargeback bill at $33.79 billion, with U.S. merchants losing $4.61 for every $1 of disputed sale once labor, technology and recovered revenue are netted out (Mastercard, 2025).

Fix: measure cost per transaction by card type and country; route foreign cards through a local acquirer when possible; deploy 3-D Secure / Visa Compelling Evidence 3.0 (cuts disputes by 20 %).

2. Returns — the silent margin killer#

The U.S. National Retail Federation reported $890 billion in returns in 2024, 16.9 % of all retail sales — and online returns ran at 24.5 % versus 8.7 % in-store (NRF / Synctrack, 2024). Apparel sits at 25–40 %, German online fashion at 44 % due to bracketing.

The arithmetic is brutal: a 25 % return rate does not cut your contribution margin by 25 %, it cuts it by about 70 % (Eightx, 2026). The full cost of a return is 30–60 % of the selling price — reverse shipping ($6–12), restocking labor ($3–8 per unit), 10–50 % devaluation, plus the lost full-price sale (Okiela, 2025).

Fix: track returns per SKU, charge for return shipping on the worst offenders (rates drop to 12–15 % when paid), and price-in an explicit return cost line in unit economics.

3. Slow-moving inventory and dead stock#

Carrying cost is 20–30 % of inventory value per year for a standard retailer; for DTC and Shopify brands, 22–41 % (Branvas, 2024). It breaks down into capital cost (8–15 %), storage and handling (4–10 %), shrinkage (1–3 %) and obsolescence and markdowns (6–12 %).

Worse, dead stock alone can cost up to 11 % of revenue. SKUs sitting on the shelf for 90+ days cost roughly three times what fast-movers cost — and 70–80 % of DTC brands fail before year three, often because of inventory blind spots (Branvas).

Fix: target ≥ 8 inventory turns per year; cap dead stock at < 20 % of warehouse space; flag any SKU with > 90 days on hand for liquidation.

4. SaaS subscription creep#

The average Shopify app costs $58–67 per month, and the average store runs 6–8 apps (Craftberry / Meetanshi, 2025). Small stores typically spend $50–200/month, Shopify Plus stores $1,000–5,000/month — and the highest-listed app, Wholster, charges $999.99/month alone.

The pattern: each app feels small. The aggregate is the third-largest fixed cost line after rent and payroll. And nobody owns the audit.

Fix: quarterly app audit. List every subscription, its monthly cost, and the last metric it actually moved. Cancel anything that fails the second test.

5. Shipping below cost#

Industry-average DTC shipping cost is 8–12 % of order value; top performers stay under 7 % (Endlesscommerce). The trap is that the carrier label price is 25–40 % less than your true shipping spend once fuel surcharges, residential-delivery fees, dimensional-weight pricing and peak-season premiums land on the invoice (Gobolt, 2026).

The classic gap we see at clients: charge $5.99 for shipping while the carrier bills $8.50 — that's $2.51 of margin gone per order, every order.

Fix: reconcile actual shipping spend against charged-to-customer line on the order export weekly. Negotiate with carriers (15–30 % typical discount). Reduce dimensional weight by trimming box height — 2 inches saved = $1–3 per shipment.

6. Discounts set without unit economics#

A 20 % discount on a 40 %-margin item does not cut profit by 20 %. It cuts gross profit by 40 %. To keep total profit the same, volume has to triple (Saras Analytics).

Two more uncomfortable facts: 64 % of online consumers prefer to wait for a sale, training your base to never pay full price; and customers acquired via discount generate less than one-quarter of the profit of full-price customers over six months.

Fix: before any campaign, calculate the break-even discount = your contribution margin %. Anything past that is paying customers to take your inventory. Bundle and threshold discounts (free shipping over X) preserve more margin than blanket % off.

7. FX losses on cross-border sales#

Currency conversion fees from banks, gateways and card networks run 2–5 % per transaction (TransFi, 2025). Most are baked into the spread, not visible as a line item.

Two real cases: UK fashion brand Odd Muse loses an estimated £40,000/year on £2 M of international Shopify sales (2 % conversion). UK sportswear brand Castore loses an estimated £500,000/year on £50 M international revenue at a 1 % rate (blkfx.co.uk).

Fix: open a multi-currency account, settle in the customer's currency, and benchmark your effective rate against the mid-market rate weekly. Specialist FX providers will quote ~0.5 %.

8. Uncontrolled ad spend (real CAC, not platform ROAS)#

A 4× ROAS on a Meta dashboard typically translates to a 5–15 % real contribution margin once COGS, shipping, payment fees and returns are deducted (Bluewater Marketing, 2026). Real CAC runs 30–60 % higher than the dashboard shows once creative production, agency fees and iOS-attribution losses are added.

The blended e-commerce CAC is now $68–90, up 60 % in five years (Retainful, 2025). The ad-platforms also blend new and returning customers, with retargeting making the overall ROAS look prettier than new-customer economics actually are.

Fix: track POAS (Profit on Ad Spend) per channel, not ROAS. New-customer CAC should be calculated against first-time customer orders from the e-shop, not from the ad platform.

9. Wrong landed COGS#

71 % of emerging brands fail to track landed cost at the SKU level (Endlesscommerce). The result: gross margin is 8–15 % lower than the dashboard claims, every quarter, until somebody adds it up.

A typical landed-cost stack adds inbound freight ($1–4 per unit), duties (0–25 %), packaging, prep ($0.20–0.80 per unit), and warehousing. A reported 69 % gross margin closes to 44 % once these layers are included (NSTAR Finance).

Fix: recalculate landed cost per SKU quarterly using actual invoices. Brands with disciplined landed-cost models preserve 6–12 percentage points more contribution margin than competitors.

What it adds up to#

In our practice we rarely see one of these killers showing up alone. We see four or five at once: a 30 % return rate hiding behind 2-3 % "shipping below cost" hiding behind 1.5 % FX leakage hiding behind ~10 % landed-cost understatement hiding behind a Meta dashboard showing 4× ROAS. The compounded effect routinely turns a "30 % gross margin" e-shop into a 12-15 % net business.

You don't fix this by getting smarter about ads. You fix it by knowing which line item leaks, by SKU, by channel, every week.

Frequently asked questions#