Baby & Maternity on TikTok Shop sells through demos and high-priced products in a shrinking category

1. Introduction

Brands and creators in Baby & Maternity must decide which products to push and which videos to make. Parents buy on trust and need, and often spend a lot per item, so the answer may differ from the wider marketplace. This paper asks how the category is moving, where its sales happen, what its top products cost and how they are rewarded, and which video formats are associated with more sales per view.

Prior work on short-video popularity finds that attention is long-tailed and that early popularity predicts later popularity [1]. Current prediction benchmarks find that creator features matter most and that content features add a smaller signal [2, 3]. Research on social commerce suggests live selling builds trust and engagement [4], and that affiliate commissions are set to balance incentives [5]. We add category-level evidence on sales rather than popularity.

All findings are associations in observational data, not effects of any choice a brand or creator makes.

2. Data

Our analysis covers Baby & Maternity on TikTok Shop from 2 September to 1 October 2026. We tracked 886 videos and analysed 72 in depth. The marketplace figures rest on the 100 top-selling products in the category, and the market comparison draws on all categories tracked, including 3,342 videos watched across them.

For each analysed video we recorded its hook, format, production style and other creative features. For each top product we recorded unit price, affiliate commission rate, launch date and sales channel. Category figures are weekly estimates, most recently as of 29 September 2026.

Not every analysed video has every attribute. The hook comparison covers 61 videos, the format comparison 66 and the production comparison 52, so group sizes are small throughout.

Daily rankings of top-selling TikTok Shop videos in the US were collected from several independent sources and merged into one record per video. Where sources overlap, their sales estimates are cross-checked against each other, and videos whose estimates disagree by more than half are flagged. An AI analyst watched each video and recorded its opening, format, angle, production style, use of AI, hook source, length, pacing and what appears in the first second, without seeing how the video sold. Sales per view is a video group’s revenue per view divided by the market’s, shrunk toward the average for small groups, with 90% ranges. 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 2026-09-02 to 2026-10-01: 886 videos tracked in Baby & Maternity, 72 of them watched and analysed; the market baseline uses all 3,342 watched videos; 100 top-selling products in Baby & Maternity. Predictions come from a logistic model retrained daily and tested on recent weeks it had not seen.

3. Methods

Sales per view is a group's revenue per view divided by the market's, so 1 is the market average. Small groups are shrunk toward the average by empirical Bayes [6], and we report 90% intervals on the log scale. The predictive model is a logistic model of top-selling videos, retrained daily and tested on recent weeks it had not seen. We report AUC [7] with a bootstrap range [8].

Momentum compares the category's estimated revenue over the latest 4 and 13 weeks with the same number of weeks before. Channels split the latest week's revenue into shoppable video, LIVE and the shop tab, with the share credited to affiliate creators. Prices, commission rates and product age use the top-selling products, with medians weighted by revenue and revenue per product expressed against the average product.

Competition is measured by shops, active products, average revenue per shop, the top ten products' share of top-product revenue and the Herfindahl–Hirschman index [9]. Seasonality compares each month with the category's average month over the past year, leaving out the partial first and last months. Separating season from trend follows [10].

4. Results: a smaller, video-led category

Baby & Maternity ranked 24th of 28 categories by revenue and took 0.8% of market revenue over the latest four weeks. Revenue over the latest 13 weeks was 16.2% below the previous 13, against a market decline of 11.1%. The latest four weeks were 5.9% above the previous four, so the recent trend is more positive than the quarter.

In the latest week, 77.7% of category revenue came through shoppable video, 8.1% through LIVE and 14.2% through the shop tab. The market figures are 64.9%, 17.1% and 18.0%. Affiliate creators were credited with 83.9% of the category's revenue, so creator video is the main route to sales here, and LIVE plays a smaller part than elsewhere.

Competition is thinner by shop count: 1,501 shops against a median of 4,461 across categories, with 9,544 active products. Average revenue per shop was $2,488 against a median of $2,401. The top ten products took 27% of top-product revenue, the same as the median category, and the HHI of 152 indicates low concentration among the tracked products.

5. Results: formats and hooks associated with sales per view

Demo or tutorial videos were the clearest result: 32 videos, 43% of analysed sales and 1.5× the market's sales per view (90% range 1.16–2.01), with high confidence. Unboxing reached 1.6× (1.17–2.30) but rests on 5 videos. Review or testimonial videos sold at 0.8× (0.58–1.12) despite 13.1% of sales, and the Other group at 0.6× (0.44–0.94). Talking head (0.94×) and voice-over B-roll (1.27×) have ranges that include 1 and low confidence.

Among hooks, Price or deal had the highest figure at 1.7× (1.26–2.34), but from 8 videos and with low confidence. Relatable POV, the largest group with 22 videos and 45.7% of sales, sat near average at 1.07× (0.81–1.41). Problem call-outs were at 0.96× (0.63–1.46). Visual pattern-break was lowest at 0.6× (0.41–0.92) from 5 videos.

All 52 videos with a production label were filmed live action, at 0.99× (0.80–1.23), so we cannot compare production styles within this category. Our predictive model reached an AUC of 0.64 (90% range 0.57–0.70) on 2,667 training rows. It was not trained only on analyses that never saw sales (cleanOnly is false). That is a modest signal, and its leading features should not be read as recipes.

6. Results: prices, commission and product age

Top products had a revenue-weighted median price of $84, with the middle half between $30 and $179. Products at $100 and over were the largest group, with 35 products and 42.0% of revenue, and earned 1.20× the average product's revenue. The $60–100 band earned 1.11× from 19 products. Below $35 revenue per product was under average: 0.58× under $10 (3 products), 0.71× at $10–20 and 0.91× at $20–35.

The revenue-weighted median commission was 8%. The 5–10% band held 50 products and 56.9% of revenue. The 15–20% band had the highest revenue per product at 1.17×, but from 8 products, and the 20–30% band (2 products) was at 0.62×. Higher commission is therefore not clearly linked to higher revenue per product here. Video carried between 86.8% and 97.8% of revenue in every band.

Products launched in the last 90 days produced only 6.6% of top-product revenue. Products aged 1–2 years produced 48.3% and those over 2 years 14.3%, with the oldest band at 1.19× the average product. In this window, established products dominated the top sellers.

7. Discussion

For brands, the evidence points to a category where established, higher-priced products sell through affiliate video, and where launches rarely reach the top within months. For creators, demonstration videos are the best-supported format, which fits purchases where parents need to see a product working. The result is consistent with the idea that high-trust purchases benefit from visible proof [4]. The strong figure for price or deal hooks is intriguing, but it is too thinly based to act on alone.

The season matters for timing. Over the past ten months, March (1.26× a normal month), April (1.20×) and June (1.20×) were strong, while January was weakest at 0.70×. November (0.89×) and December (0.90×) of 2025 were below a normal month. Halloween is under way, Black Friday / Cyber Monday begins in 40 days and holiday gifting in 45. The category did not appear to benefit from these moments last year, though one year of data cannot separate season from trend.

Alternative explanations are plentiful. Larger creators and ad boosting may drive both views and sales, and observational data cannot isolate the returns to advertising [11]. Demo videos may simply be made for products that already sell well. The weak model performance suggests most of what separates top videos is not captured by our creative labels.

8. Limitations

Most videos in the sample are top-ranked, so findings mostly separate strong sellers from good ones rather than from all videos; typical and weak videos are being added. 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. 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. The window is 30 days and covers one category, so results may not hold in other periods. Only 72 videos were analysed in depth, and most hook and format groups have 5–12 videos and low confidence. Seasonality draws on about ten months, which cannot separate season from trend. Marketplace figures rest on the 100 top-selling products, which describes the best sellers, not the whole long tail of products [brynjolfsson2003]. All findings are associations.

9. Conclusion

In Baby & Maternity, sales are concentrated in shoppable, affiliate-led video, in established products and in higher price points. Demo or tutorial videos were associated with sales per view 1.5× the market average, the one video-level result with high confidence. The category contracted over the quarter, by 16.2% against 11.1% for the market, though the latest four weeks improved.

The evidence supports testing demonstration-led video on proven, higher-priced products and planning around the stronger spring months. It does not show that any format causes sales. Several hook and format results rest on very few videos and should be treated as hypotheses for further testing.