ContentIQ working paperCIQ-WP-2026-53TikTok ShopVersion 118 Oct 2026
How to choose a TikTok Shop agency
51,843 videos tracked, 29 categories, 19 Sept–18 Oct 2026 (US TikTok Shop)
MediaLabs Research · in collaboration with Coherence Limited
Key findings
- 0166.4%of top-video revenue came with paid promotion, on 16.1% of views (a lower bound)
- 022.37median estimated ROAS of paid top-selling video snapshots (middle half 1.95–2.81); 27.7% were under 2
- 031.94–2.91range of category median ROAS, from Jewelry Accessories & Derivatives to Beauty & Personal Care
- 0421%of top video snapshots were organic (under 10% paid) but they took 11.9% of revenue
- 055.3%–78.7%share of category revenue sold through shoppable video, Collectibles to Baby & Maternity
Contents
Abstract
Choose an agency on evidence you can check, not on pitch: ask for category-specific benchmarks, a clear split between paid and organic attribution, and a pre-agreed test. In our analysis of US TikTok Shop, paid promotion came with at least 66.4% of top-video revenue on 16.1% of views, and the median estimated return on ad spend of paid top-selling videos was 2.37 (middle half 1.95–2.81), varying by category from 1.94 to 2.91. Sales channel mix also differs sharply by category, from 68.3% of Beauty & Personal Care revenue through shoppable video to 5.3% for Collectibles. We turn these benchmarks into a due-diligence scorecard, clearly labelled as a recommendation, not a finding; we do not rank agencies or observe agency results.
1. Introduction
The short answer for a brand deciding on an agency: pick the one that can show evidence specific to your category, separate paid from organic revenue when reporting, and agree a test with a pre-set success threshold before you scale spend. This paper gives benchmarks to judge such a proposal against; it does not rank agencies, and it observed no agency data, prices or client results.
The benchmarks matter because TikTok Shop revenue is heavily tied to paid promotion. Across the top-selling videos we tracked, at least 66.4% of revenue came with paid promotion, so an agency's claims are in large part claims about paid media, creator content and attribution. Attribution is hard: observational data make it difficult to separate the returns to advertising from sales that would have happened anyway [1]. Sales are also concentrated in a small number of very successful items of content, a long-tailed pattern known from online content generally [2].
Our question is: what do category benchmarks say about what to ask an agency, and which evidence should be treated as strong or weak? Findings are reported as associations. Recommendations are marked as such.
2. Data
The sample is 51,843 videos tracked on US TikTok Shop between 19 September and 18 October 2026, across 29 categories. This paper does not use in-depth creative analysis, so nothing here rests on video content analysis; the paper is about sales, paid share and channels.
The paid-promotion and economics statistics rest on 33,825 top-selling videos and 110,867 video snapshots (one row per video and observation date). The channel figures use each category's latest week, as of 12 October 2026. For each video snapshot we recorded weekly revenue, views, the share of revenue and views from paid promotion, estimated return on ad spend (ROAS), likes and shares per 1,000 views, and video age.
These are top-selling videos, so the data describe videos that already sold well. They are US TikTok Shop observations, not all user-generated content everywhere.
Weekly figures for the top-selling TikTok Shop videos in the US were collected from commercial data providers: each video’s estimated revenue, views and engagement in each week it ranked, the share of its revenue and views that came from paid promotion (reported per video and top-coded at 80%), its estimated return on ad spend and its publish date. One row is one video on one observation date (a video snapshot). Weekly category figures for the US (estimated revenue, its split between video, LIVE and the shop tab, affiliate revenue, shops and active products, and a year of weekly revenue) and daily rankings of each category’s top-selling products (unit price, commission rate, launch date and revenue by channel) were collected from several independent sources. Medians of price and commission are weighted by revenue; momentum compares the latest weeks with the same number of weeks before. This paper covers 2026-09-19 to 2026-10-18: 51,843 videos tracked.
Daily seven-day video snapshots can overlap and repeat the same video. Counts are observations, not independent weeks. Aggregate shares and medians are snapshot-weighted; summed revenue is not unique sales over the publication window.
3. Methods
Sales channels: each category's revenue in the latest week was split into shoppable video, LIVE and the shop tab, with any remainder treated as other or unattributed (zero in every category here). We also report the share credited to affiliate creators.
Paid share: each top-selling video's weekly revenue and views carry the share that came from paid promotion, reported per video and top-coded at 80%. Market and category shares weight videos by revenue (or views), so they are lower bounds. ROAS is estimated revenue per dollar of ad spend, summarised by its median and middle half. For this analysis, videos with under 10% of revenue paid are classified as mostly organic.
Video economics: revenue and engagement per 1,000 views are totals divided by total views, and each category's share of revenue through video comes from its weekly figures [2]. Whether a ROAS is profitable depends on margins we do not observe.
4. Results: paid promotion and organic content
Paid promotion accounts for most top-video revenue. At least 66.4% of revenue came with paid promotion, on 16.1% of views. In 56.8% of video snapshots, the paid share was at or above the 80% top-code, so the true market share is higher than reported.
Organic video snapshots (under 10% paid) were 21% of the total and took 11.9% of revenue; paid-led video snapshots (80% or more) were 56.8% of video snapshots and took 65.9%. Revenue per 1,000 views was similar across bands: $17.06 for organic and $16.02 for paid-led. Median weekly revenue was higher for paid-led ($3,685) than for organic ($1,550). Organic videos had far more engagement: 148.8 likes and 16.5 shares per 1,000 views, against 37.3 and 4.7 for paid-led.
These are associations among videos that already ranked as top sellers, not evidence that paid spend caused the revenue. An agency that reports only total sales growth cannot show which of these bands it is operating in.
| Paid share of the video’s revenue | video snapshots (n) | Share of revenue | Revenue per 1,000 views | Median weekly revenue | Likes per 1,000 views | Median ROAS |
|---|---|---|---|---|---|---|
| Organic (under 10% paid) | 23,310 | 11.9% | $17.06 | $1.6K | 148.8 | — |
| Mixed (10–50% paid) | 6,933 | 5% | $16.12 | $2.5K | 96.2 | 2.34× |
| Mostly paid (50–80%) | 17,688 | 17.2% | $16.94 | $3.3K | 43.8 | 2.43× |
| Paid-led (80% or more) | 62,936 | 65.9% | $16.02 | $3.7K | 37.3 | 2.36× |
Note. A video snapshot is one video’s rolling seven-day metrics on an observation date. The paid share is reported per video and top-coded at 80%, so paid shares are lower bounds. ROAS is estimated revenue per dollar of ad spend.
5. Results: return on ad spend by category
The median estimated ROAS of paid video snapshots was 2.37, with a middle half of 1.95–2.81. 27.7% of paid video snapshots had a ROAS under 2. Category medians ranged from 1.94 in Jewelry Accessories & Derivatives (middle half 1.60–2.32) and 1.96 in Baby & Maternity (1.66–2.29) to 2.91 in Beauty & Personal Care (2.39–3.28) and 2.86 in Womenswear & Underwear (2.53–3.15).
Categories differ in paid dependence. Paid promotion came with 72.4% of top-video revenue in Household Appliances and 70.4% in Beauty & Personal Care, but 51.6% in Kids' Fashion and 54.7% in Food & Beverages; these are lower bounds. Organic video snapshots were 10.7% of Beauty & Personal Care but 31.6% of Kids' Fashion and 30% of Food & Beverages.
Organic and paid-led revenue per 1,000 views also diverge by category. Household Appliances earned $38.91 organic against $28.98 paid-led, while Baby & Maternity earned $12.73 organic against $22.87 paid-led. These gaps are descriptive and may reflect product or video mix.
| Category | video snapshots (n) | Paid share of revenue | Paid share of views | Mostly organic video snapshots | Median ROAS (middle half) |
|---|---|---|---|---|---|
| Beauty & Personal Care | 4,839 | 70.4% | 15.1% | 10.7% | 2.91× (2.4–3.3) |
| Womenswear & Underwear | 3,103 | 67.3% | 15.5% | 14.1% | 2.86× (2.5–3.1) |
| Household Appliances | 3,279 | 72.4% | 21.3% | 11.2% | 2.53× (2.1–3.0) |
| Phones & Electronics | 4,881 | 64.4% | 16.6% | 18.4% | 2.65× (2.1–3.0) |
| Sports & Outdoor | 4,676 | 66.1% | 21.4% | 20.1% | 2.56× (2.1–3.0) |
| Health | 3,789 | 70.3% | 33.5% | 14.3% | 2.41× (2.1–2.9) |
| Shoes | 3,443 | 67.5% | 31.7% | 15.7% | 2.49× (2.2–2.8) |
| Furniture | 4,048 | 66.1% | 20.5% | 16.4% | 2.20× (1.9–2.6) |
| Home Supplies | 3,755 | 66.6% | 13.8% | 15% | 2.67× (2.3–3.0) |
| Tools & Hardware | 4,585 | 67.9% | 10.6% | 14.9% | 2.43× (2.0–2.8) |
| Textiles & Soft Furnishings | 3,407 | 68.7% | 9.3% | 11.6% | 2.63× (2.3–2.9) |
| Automotive & Motorcycle | 5,017 | 67.8% | 13.6% | 17.9% | 2.51× (2.1–2.9) |
| Menswear & Underwear | 3,353 | 61.9% | 13.2% | 20.6% | 2.75× (2.4–3.1) |
| Fashion Accessories | 4,594 | 58.8% | 17.3% | 26.9% | 2.24× (1.9–2.7) |
| Home Improvement | 4,846 | 67.1% | 14% | 17.5% | 2.30× (1.9–2.7) |
Note. Mostly organic: under 10% of the video’s revenue paid. Paid shares are top-coded at 80% per video, so they are lower bounds.
6. Results: where sales happen
An agency's plan should match the category's sales channel mix. Shoppable video carried 68.3% of Beauty & Personal Care revenue, 71.9% of Womenswear & Underwear and 78.7% of Baby & Maternity in the latest week. It carried only 28.4% in Jewelry Accessories & Derivatives, 42.1% in Books, Magazines & Audio and 5.3% in Collectibles, where the shop tab (56%) and LIVE (38.7%) dominated.
LIVE took 23.8% of Health revenue and 21.2% of Kitchenware, but 4.9% of Menswear & Underwear. Affiliate creators were credited with 85.3% of Baby & Maternity revenue and 82.4% of Household Appliances, against 7.4% in Collectibles and 30.9% in Jewelry Accessories & Derivatives. LIVE selling has been linked to trust and engagement in social commerce [3], but we observe only revenue shares, not why they differ.
| Category | Video | LIVE | Shop tab | Other | Through affiliates |
|---|---|---|---|---|---|
| Beauty & Personal Care | 68.3% | 16.6% | 15.1% | 0% | 80.2% |
| Womenswear & Underwear | 71.9% | 11.5% | 16.6% | 0% | 79.6% |
| Sports & Outdoor | 63.7% | 13.5% | 22.8% | 0% | 75.2% |
| Fashion Accessories | 55.6% | 10.9% | 33.5% | 0% | 60.5% |
| Phones & Electronics | 62.7% | 19% | 18.3% | 0% | 75.7% |
| Health | 64% | 23.8% | 12.3% | 0% | 81.7% |
| Home Supplies | 59.2% | 15.4% | 25.4% | 0% | 71.3% |
| Menswear & Underwear | 67.6% | 4.9% | 27.5% | 0% | 69.6% |
| Shoes | 71.2% | 12.2% | 16.5% | 0% | 80.2% |
| Furniture | 56.2% | 14.4% | 29.4% | 0% | 71.3% |
| Household Appliances | 64.9% | 18.6% | 16.5% | 0% | 82.4% |
| Collectibles | 5.3% | 38.7% | 56% | 0% | 7.4% |
| Automotive & Motorcycle | 66.7% | 8.1% | 25.2% | 0% | 70.9% |
| Tools & Hardware | 60.4% | 10.9% | 28.7% | 0% | 70.2% |
| Toys & Hobbies | 62.5% | 16% | 21.5% | 0% | 74.6% |
Note. Shares of each category’s estimated revenue in the latest week. Through affiliates is revenue credited to creators other than the selling shop.
7. Discussion
For brands, the evidence suggests that a proposal should be judged on how it handles paid promotion, attribution and category fit, because paid promotion came with most top-video revenue. Because measuring advertising returns from observational data is hard [1], agency-reported ROAS is not automatically comparable with the estimates here, which come from a different method and cover top sellers only. Alternative explanations for our patterns include selection (top sellers attract spend), product mix and margins we cannot see.
Recommendation, not a finding: due-diligence scorecard. (1) Category benchmark: does the agency state expected ROAS against the range for your category (e.g. a middle half of 1.60–2.32 in Jewelry Accessories & Derivatives, 2.39–3.28 in Beauty & Personal Care) and explain any gap? (2) Attribution: does reporting separate organic, paid and affiliate revenue, and define how ROAS is calculated? (3) Margin: is the break-even ROAS for your product stated, since profitability depends on margins? (4) Channel fit: does the plan reflect your category's video, LIVE and shop-tab mix? (5) Evidence quality: are claims backed by sample sizes, ranges and a comparison group, or only by selected successes?
Recommendation, not a finding: proposed test plan. Agree in writing, before spend, the metric, a break-even ROAS, a fixed budget and window, and a comparison (for example a holdout product or period) so that results are not judged only on total sales. Review the paid versus organic split at the end, and scale only if the pre-agreed threshold is met. For creators, the lower engagement per 1,000 views of paid-led videos suggests agreeing whether content is judged on organic reach or paid conversion.
8. What this means for brands and founders
- Plan for paid reach: in this market at least 66.4% of top-video revenue came with paid promotion, so ask any agency how it will plan, report and fund paid spend.
- Ask for category-specific ROAS targets: the median ROAS of paid top-selling videos was 2.37, but category medians ran from 1.94 to 2.91, so a single headline target may not fit your category.
- Set a break-even ROAS from your own margins before a test: 27.7% of paid video snapshots had an estimated ROAS under 2, and profitability depends on margins we do not observe.
- Match the plan to your channel mix: shoppable video carried 68.3% of Beauty & Personal Care revenue but 28.4% in Jewelry Accessories & Derivatives, so check the agency covers the channels your category sells through.
- Require paid and organic revenue reported separately: organic video snapshots were 21% of the total but took 11.9% of revenue, and total sales alone cannot show which is driving results.
9. Conclusion
In US TikTok Shop top-selling videos, paid promotion came with at least 66.4% of revenue, median estimated ROAS was 2.37, and both paid dependence and channel mix varied widely by category. A good agency proposal should therefore show category-specific benchmarks, transparent attribution and a pre-agreed test.
The scorecard above is a recommendation built on these benchmarks. This paper does not rank agencies, observe agency fees or results, or show that any approach causes sales; it gives a reference point for the questions to ask.
Methodology
Weekly figures for the top-selling TikTok Shop videos in the US were collected from commercial data providers: each video’s estimated revenue, views and engagement in each week it ranked, the share of its revenue and views that came from paid promotion (reported per video and top-coded at 80%), its estimated return on ad spend and its publish date. One row is one video on one observation date (a video snapshot). Weekly category figures for the US (estimated revenue, its split between video, LIVE and the shop tab, affiliate revenue, shops and active products, and a year of weekly revenue) and daily rankings of each category’s top-selling products (unit price, commission rate, launch date and revenue by channel) were collected from several independent sources. Medians of price and commission are weighted by revenue; momentum compares the latest weeks with the same number of weeks before. This paper covers 2026-09-19 to 2026-10-18: 51,843 videos tracked.
Daily seven-day video snapshots can overlap and repeat the same video. Counts are observations, not independent weeks. Aggregate shares and medians are snapshot-weighted; summed revenue is not unique sales over the publication window.
Limitations
Product figures cover the top-selling products we track in each category, not every listing, so they describe what sells best rather than everything on sale. Video figures cover top-selling videos only. Paid shares are reported per video and top-coded at 80%, so they are lower bounds; ad spend and margins are not observed, so a reported return on ad spend says nothing about profit. Results are associations, not causes. Paid promotion, creator audience size and product price are not fully controlled for. Revenue, views and sales are estimates, not figures reported by TikTok. These are observational data on top-selling videos, so selection into the sample cannot be separated from effects. Paid shares are top-coded at 80% and so are lower bounds, ROAS is estimated and margins are unobserved. In-depth creative analysis is not included in this paper. The data cover US TikTok Shop only, and we observed no agency data, fees or results, so nothing here compares or ranks agencies. Daily seven-day video snapshots can overlap and repeat the same video. Counts are observations, not independent weeks. Aggregate shares and medians are snapshot-weighted; summed revenue is not unique sales over the publication window.
References
- [1]Lewis, R. A., & Rao, J. M. (2015). The unfavorable economics of measuring the returns to advertising. The Quarterly Journal of Economics, 130(4), 1941–1973. academic.oup.com/qje/article-abstract/130/4/1941/1914592
- [2]Szabo, G., & Huberman, B. A. (2010). Predicting the popularity of online content. Communications of the ACM, 53(8), 80–88. cacm.acm.org/research/predicting-the-popularity-of-online-content
- [3]Wongkitrungrueng, A., & Assarut, N. (2020). The role of live streaming in building consumer trust and engagement with social commerce sellers. Journal of Business Research, 117, 543–556. www.sciencedirect.com/science/article/abs/pii/S0148296318304168
Cite as
MediaLabs Research, in collaboration with Coherence Limited (2026). How to choose a TikTok Shop agency. ContentIQ Working Paper CIQ-WP-2026-53, version 1. https://medialabs-co.com/research/tiktok-shop-agency-selection
Analysis and data: Coherence Research, Coherence ContentIQ. Published by MediaLabs in collaboration with Coherence Limited.
Version history
- Version 118 Oct 2026This version