Amazon lived through this story a decade ago. Between 2015 and 2018, third-party sellers flooded the marketplace, gross merchandise value (GMV) exploded, and most of the new entrants lost money. Fees rose, and ad costs climbed. Tariffs and Chinese import competition took what remained of unit economics. Marketplace Pulse has tracked this kind of margin pressure as a recurring Amazon pattern, and they call today’s ongoing version the “Great Compression.”¹ Amazon’s third-party services sales grew from $16.1 billion in 2015 to $23 billion a year later, and by 2017 third-party sellers had crossed 50 percent of Amazon’s retail business.² Operators who remained profitable during that period treated catalog data as an asset and measured margin per unit rather than revenue.

TikTok Shop is running the same cycle on a shorter clock. eMarketer forecasts US sales of $23.4 billion in 2026, 48 percent growth over 2025, with 57.7 million US buyers shopping through the platform by year’s end.³ Every quarter adds more merchants, more creators, and more platform-funded promotions pressing down on unit economics.

Four leaks separate the accounts that scale profitably from those generating headlines alongside losses. Two sit inside the catalog, where feed disapprovals remove listings from discovery without flagging an error, and returns run higher than on intent-based channels. The other two sit within the commercial relationship, where creator commissions compound on orders that later return and discount drift trains shoppers to wait rather than buy at full price.

Some catalogs shouldn’t be on TikTok Shop at all. High-ticket furniture, heavily regulated categories, and low-margin commodity goods without a visual story struggle with the platform’s content model and fee structure. If your unit margin can’t absorb 15 to 25 points of platform and creator costs, skip the scaling conversation and decide whether the channel belongs in your mix.

Each leak erodes contribution margin, the revenue that survives after platform fees, creator commissions, and return-and-shipping costs. GMV reporting is designed to hide that number. Our earlier piece on mobile and social commerce made the case for why merchants need to be on these channels. This one covers how to make them pay back.

Leak 1: The feed disapprovals nobody’s watching

TikTok Shop’s feed specs don’t match those of Google Shopping or Amazon. Titles cap at roughly 34 characters to display cleanly in mobile layouts, images need to clear 800×800 pixels with a light background, and the product should fill at least 80% of the frame.⁴ Good-tier listings also need five images minimum and structured attribute fields, which gates visibility in browse and recommendation feeds.⁵

When a listing misses spec, TikTok Shop doesn’t always bounce it back with a clear error. Disapprovals roll out in batches after feed refreshes, and the first sign is often a drop in impressions that shows up days later in reporting. By then, the SKU has missed a promotional window or lost placement on a creator’s video.

Feed errors compound when they meet inventory data. TikTok Shop tracks a Seller-Fault Cancellation Rate that must stay at or below 2.5% of order volume; oversells from stock-sync lag count against the threshold.⁶ A high cancellation rate reduces algorithmic visibility and limits access to promotional tools. It can also trigger Order Volume Limits that cap your sell-through until the cancellation rate recovers.⁷

For a useful diagnostic, compare the SKUs your team listed on TikTok Shop last month with the SKUs serving impressions this week. The answer usually surprises the merchant asking. Profitable accounts close the loop with two habits. They reconcile listed, active, and in-stock SKUs weekly, and they validate feeds before submission rather than after disapproval. Automated feed management tools like the Athos platform handle the mechanics, running channel-specific validation and real-time inventory sync across 1,500+ channels. The habits matter more than the tool.

Leak 2: Returns that eat into the channel

Returns are the second leak, and they run higher on TikTok Shop than on intent-based channels. Apparel, beauty, and fashion accessories are the categories where creator video drives impulse purchases, and they see particularly high return rates. Industry reporting shows TikTok Shop returns for creator-driven categories well above those for the same products on Amazon, where buyers arrive with stronger purchase intent and more product-page research behind them.

Product data drives most of these returns. We covered the broader argument in our earlier piece on data as a storefront. Mismatched sizing, missing attribute fields, and idealized creator imagery lead shoppers to purchase items they regret when the package arrives. Most TikTok Shop buyers make their decision based on a 30-second video of someone else reacting to the product. When the listing doesn’t back that up with specs and realistic imagery, the return is already in motion at checkout.

This was the other shoe for Amazon third-party sellers during the 2016-2018 margin crunch. The sellers who survived treated returns as a product-data problem and fixed them at the catalog level, before they became an ops burden. TikTok Shop is now running the same dynamic, and its content-first discovery model makes impulse-return risk structurally higher than on Amazon.

For a useful diagnostic, pull the top 20 SKUs by TikTok Shop GMV and overlay return rate per SKU. Top sellers that are also top returners tell you the growth story on your board reports is subsidizing the returns operation. Fix it upstream by treating product data as a return-prevention asset. Structured attributes, accurate dimensions, and realistic imagery reduce the distance between what the creator video promised and what arrives in the box. Front-load catalog enrichment before you ramp creator spend.

Leak 3: Creator commissions that compound on returns

TikTok Shop lets sellers set creator commission rates anywhere from 1% to 80% of order value.⁸ Most brands run open-invitation programs in the low double digits, with higher negotiated rates for top-performing creators in beauty, fashion, and home. Commission stacks on top of the 6% platform referral fee⁹ and roughly 1-4% payment processing, so the combined platform and creator cost on a typical affiliate sale often runs above 20 points of gross revenue before cost of goods.

Refunds create the real cost of creator affiliates. Creators get paid on orders that later get refunded, and TikTok Shop’s commission recovery rules push clawback complexity through your finance system weeks after the original sale.¹⁰ Videos also keep selling after campaign windows close. The same paid-promotion video can drive commissioned sales for months, adding to working-capital drag and reducing your control over product positioning.

A $28 product with 60% gross margin looks like a winner on paper. Subtract a 6% referral fee, 12% creator commission, and 3% payment processing, then layer in the elevated return rate on creator-driven categories, and the contribution margin on a meaningful share of units turns negative. The planning conversation shifts from “how do we scale this SKU” to “should we pause it until the return rate comes down.”

Do you know your effective creator commission rate per SKU, net of returns? Most brands don’t have a clean answer. TikTok’s Seller Center reports creator commissions, your finance system reports returns, and brands that want a unified view at the SKU level usually need to build a custom pipeline to bridge the two. Three habits tighten the economics in the meantime. Track SKU and creator performance together, cap commission exposure per creator, and pause any creator whose return rate crosses a red-line benchmark.

Leak 4: Discount drift that trains your customer

TikTok Shop’s promotional structure pulls brands toward continuous discounting. Platform-funded programs like the Bonus initiative underwrite first-order discounts of up to 50%.¹¹ Creator-negotiated codes add another layer on top of the platform-funded discount. Once a brand’s products have appeared at a promotional price for a few weeks, the full-price listing stops converting, and retreating to full price is harder than entering at a discount.

The effect compounds in the algorithm. TikTok’s recommendation system learns from engagement, and “I found a deal” creator content tends to outperform full-price brand content in the feed. The longer your products appear at a discount, the more the algorithm associates your brand with discount intent, the more creators structure videos around the promotional price, and the more full-price visibility erodes.

Pull your last 90 days of TikTok Shop orders and calculate your average discount depth across SKUs that were promoted. Compare it to your own storefront over the same window. When the TikTok Shop figure runs deeper than the storefront figure, shoppers are being trained to wait for a promotion. Rolling the channel back to full-price conversion takes longer than the original ramp that created the expectation. Close this leak with promotion governance. Set pre-approved discount depths per channel and SKU, and time-box promotions with hard-stop dates rather than letting discounts become the default. Rotate which SKUs carry the promotional load across a quarter. A single product that stays on sale for months gets anchored to a discount position in the feed, and the algorithm starts treating the promotional price as that product’s baseline. Treat TikTok Shop as a retail partner with its own pricing strategy, and apply the same full-price discipline you use on your direct-to-consumer storefront.

What the profitable Amazon sellers figured out

One discipline defends against all four leaks. Operators who measure contribution margin channel by channel and act on the data before GMV reporting inflate TikTok Shop into a false growth story, coming out ahead of those chasing headlines. Catalog hygiene and return discipline on the data side, paired with creator attribution and promotion governance on the commercial side, all trace back to that same habit.

Close the four leaks over the next quarter with this tiered plan.

This week: audit your TikTok Shop listings for feed compliance and active status. Match the count of active, impression-serving SKUs to your ERP record of live listings. Divergence between those numbers is the first leak worth closing.

This month: measure contribution margin by channel, not GMV. Pull platform fees, creator commissions, and return-and-shipping costs into a single per-SKU view for your major channels (TikTok Shop, Google Shopping, Amazon) and compare those channels to your storefront economics. The comparison usually reveals which channels are actually paying you.

This quarter: cap creator commission exposure per SKU and put a pricing governance process in place across channels. Set pre-approved discount depths, time-box promotions, and rotate which SKUs carry the promotional load. 

Marketplace Pulse tracks today’s version of this pattern as the “Great Compression.” Amazon sellers who came out ahead of earlier compression cycles shared three habits. They treated product data as a revenue asset and maintained it like one. Channel-level contribution margin replaced GMV as its benchmark metric. When a category or platform stopped paying back, they walked away, even if the headline growth numbers still looked good.

TikTok Shop is early enough in its own cycle that the same playbook still works. Build catalog hygiene, return discipline, and channel-level economics now, and you stay profitable on the channel when growth slows and fees tighten. Everyone else will spend 2027 and 2028 relearning what Amazon’s third-party sellers learned in 2018.

Athos Commerce is a commerce platform that unifies feed management, channel performance, and inventory sync across the 1,500+ channels that matter for modern commerce. But the habits that close the four leaks belong to the operator. Tools make the math faster, and disciplined operators decide whether a channel is still profitable in 2029.

Sources & Further Reading

1. Marketplace Pulse. “The Marketplace Pulse Year in Review 2025.” Marketplace Pulse, 2026. https://www.marketplacepulse.com/reports/year-in-review-2025
2. Marketplace Pulse. “Amazon Third-Party Seller Services Sales 2015–2025.” Marketplace Pulse, 2026. https://www.marketplacepulse.com/stats/amazon-third-party-seller-services-sales
3. eMarketer. “TikTok Shop Makes Up Nearly 20% of Social Commerce in 2025.” eMarketer Press Release, 2025. https://www.emarketer.com/press-releases/tiktok-shop-makes-up-nearly-20-of-social-commerce-in-2025/
4. TikTok Seller Center. “Product Detail Pages & Listing Quality Guidelines.” TikTok Seller University, 2026. https://seller-us.tiktok.com/university/essay?knowledge_id=481891871868714&lang=en
5. TikTok Seller Center. “Product Detail Pages & Listing Quality Guidelines.” TikTok Seller University, 2026. https://seller-us.tiktok.com/university/essay?knowledge_id=481891871868714&lang=en
6. TikTok Seller Center. “Seller-Fault Cancellation Rate Requirements.” TikTok Seller University, 2026. https://seller-us.tiktok.com/university/essay?knowledge_id=7953001314830094
7. TikTok Seller Center. “Guide to Shop Performance Score (SPS).” TikTok Seller University, 2026. https://seller-us.tiktok.com/university/essay?knowledge_id=5513514539419405&lang=en
8. TikTok Seller Center. “How Affiliate Commission Works.” TikTok Seller University, 2026. https://seller-us.tiktok.com/university/essay?knowledge_id=6077860360177451&lang=en
9. TikTok Seller Center. “TikTok Shop Referral Fee Updates – 2024.” TikTok Seller University, 2024. https://seller-us.tiktok.com/university/essay?knowledge_id=5982454398175018&lang=en
10. TikTok Seller Center. “How Affiliate Commission Works.” TikTok Seller University, 2026. https://seller-us.tiktok.com/university/essay?knowledge_id=6077860360177451&lang=en
11. TikTok Seller Center. “TikTok Funded Program.” TikTok Seller University, 2026. https://seller-us.tiktok.com/university/essay?knowledge_id=5370507046324010&lang=en