Syncly Social Index
Syncly Social Index: Methodology
The Syncly Social Index (SSI) measures how beauty and wellness brands, products and trends perform in short-form video. We publish it every month from Syncly’s social video library.
Last updated: October 2026. Applies to the September 2026 edition onward.
At a glance
Platforms: TikTok videos, Instagram Reels and YouTube Shorts.
Geography: global. Videos in any language count when they mention the brand.
Period: calendar months (UTC). The first edition covers September 2026.
Cadence: monthly. Each edition is published after its month closes.
Categories: skincare, makeup, hair, fragrance, and wellness & functional drinks.
Panel: 10 brands per category, chosen by video mentions that month.
Score: 0 to 100 per brand, relative to the other brands in its panel.
Where the data comes from
Every number comes from Syncly’s library of public short-form videos. For each video we index the original post (caption, title, hashtags and @handles) and what the video itself shows and says: on-screen text, the speech transcript, and an AI-generated summary and keyword list.
The library doesn’t hold every video posted on each platform, and the amount we collect changes from month to month. Raw counts from different months can’t be compared, so every trend and month-over-month figure is reported as a share (see Normalization). We don’t publish the total size of the dataset.
How we count a mention
A brand, product or term counts as mentioned in a video when it appears anywhere in the post or in the video itself: caption, title, hashtags, @handles, on-screen text, transcript, or the video’s AI summary and keywords.
Spelling variants. “Beauty of Joseon”, “#beautyofjoseon” and “Beauty-of-Joseon” all count as the same brand. Where a brand is widely written in another script, we search for that too (for example 메디큐브 for Medicube).
Ambiguous names. Some brand names are also everyday words, places or other products. For these, a video only counts if it also contains category context. Dyson counts only in hair videos, so vacuum videos are left out, and K18, Amika and Celsius need haircare or drink context. Some searches also exclude specific phrases: Rhode leaves out “Rhode Island”, and Red Bull leaves out Formula 1 racing posts. Brand names of four letters or fewer (such as e.l.f., NYX and K18) only count as a standalone word, hashtag or handle, so “herself” or “shelf” never counts as e.l.f.
Checking the search. Before a search is used, we check a sample of the videos it returns and refine it until the results are about the right brand or product. When a phrase keeps matching the wrong thing, we count those videos separately and subtract them. Known limits are noted on the report they affect.
Videos that mention several brands. A video that mentions two panel brands counts once for each.
How brands are chosen
For each category we keep a list of candidate brands that are widely discussed in short-form video. Each month we count every candidate’s video mentions, and the 10 brands with the most mentions form that month’s panel. Panels are chosen again every month, so a brand can enter or drop out. A brand that is new to the panel is marked NEW.
Each brand is ranked in one category. Luxury houses with large makeup and fragrance lines (Dior, YSL, Chanel and Armani) are ranked in fragrance, using fragrance-only mentions.
A brand’s total includes all of its lines in that category. Beauty of Joseon’s and SKIN1004’s sunscreens count toward their skincare totals, and hair includes styling tools such as the Dyson Airwrap and Shark FlexStyle.
If two brands tie for the last panel place, the brand with more videos over 100,000 views gets it.
Whether a brand is a Syncly customer plays no part in choosing panels or calculating scores.
The score
SSI combines six components, each measured for every panel brand in the same month.
Component
What it measures
Weight
Video share of voice
The brand’s share of all mentions of the 10 panel brands
25
Untagged share
Share of videos where the brand is shown or said in the video, but its name never appears in the caption, title, hashtags or handles
20
Momentum
Change in the brand’s share of all beauty video (wellness video for the wellness panel) vs the previous month
15
Organic sentiment
Share of positive videos among videos that aren’t sponsored
15
High reach
Share of the brand’s videos with 100,000 views or more
15
Creator breadth
1 minus the share of the brand’s videos made by its 10 most active creators
10
From components to a score
Rank each component within the panel. For the four components measured on all matching videos, the top brand gets 100, the bottom brand gets 0, and the others are spaced evenly between them. Brands that tie share the same score.
Score sampled components in thirds. Untagged share and organic sentiment come from samples (see Sampling), so small gaps between brands are within the margin of error. Brands in the top third of the panel get 100, the middle third 50, and the bottom third 0.
Combine. SSI is the weighted average of the six component scores, using the weights above. Every component score is published next to the composite.
What the score means
SSI is a relative score. It shows how a brand did compared with the other nine brands in its panel that month.
Scores can’t be compared across categories. A 70 in skincare and a 70 in hair come from different panels.
A brand’s score can move when other brands in its panel move.
The brand with the most videos isn’t always first. In September 2026, Medicube had the most skincare videos (12,731, against 4,663 for Anua) and led on share of voice and momentum. Anua ranked first overall (74.5 vs 67.5). It was in the top third for untagged mentions (28.8% of sampled videos vs 18.2%), and it had a higher share of 100k+ view videos and a broader creator base.
Sampling
Untagged share and organic sentiment need a check of each video, so they come from a sample of about 66 videos per brand per month:
the 30 most viewed videos of the month, and
36 videos spread across the month: 6 from each of six five-day windows, taken alternately from the start and the end of each window.
Videos that turn out not to mention the brand are dropped.
Untagged is decided by fixed rules. After the same spelling normalization used for search, we check whether the brand name appears in the caption, title, hashtags or handles. If it doesn’t, and the brand appears in the on-screen text, transcript, summary or keywords, the video is untagged.
Sentiment is labelled positive, neutral or negative by an AI model, from an excerpt of each video’s caption, on-screen text and transcript.
Sponsored videos are left out of organic sentiment. A video counts as sponsored when the platform marks it as a paid partnership (TikTok reports this), or when the post discloses it with tags such as #ad, #sponsored, #partner or #gifted.
With about 66 videos, a sampled share has a margin of error of roughly ±12 percentage points, which is why these two components are scored in thirds.
The samples on trend, product and comparison pages (how a term is framed, a creator’s stance, which side a creator prefers) are also labelled by an AI model from each video’s caption, on-screen text and transcript.
Untagged mentions
An untagged mention is a video where a brand is visible or spoken about, but the creator never names it in the caption, hashtags or handles. Listening tools that only read captions and tags miss these videos. A high untagged share means a large part of the brand’s video exposure doesn’t show up in caption or hashtag searches.
Normalization and month-over-month change
Monthly changes are measured as shares, so a change in how much we collected doesn’t show up as growth or decline.
Beauty categories: the baseline is all short-form beauty videos that month, meaning videos that mention makeup, skincare, haircare, perfume or sunscreen.
Wellness & functional drinks: the baseline is all short-form wellness videos that month (supplements, vitamins, energy drinks, electrolytes, protein, gut health).
Momentum, the month-over-month % shown on brand and product pages, and trend lines all use these shares. Rank changes (▲ ▼) compare a brand’s rank with its rank in the previous month’s panel.
Trend reports
Category reports open with what videos are saying: the ingredients, claims and techniques that grew fastest that month. We find them in two ways:
the most frequent keywords and hashtags in each category’s videos, compared with the previous month, after removing generic words such as “skin”, “routine” or “fyp”; and
a tracked list of ingredients, claims and techniques that we count every month. The list grows as new terms show up.
A term must pass a minimum volume before it can be called rising. When a term stands out, it gets its own deep-dive report. A deep-dive shows the term’s monthly share of beauty (or wellness) video over at least four months, the brands and products mentioned with it, how often it appears only inside the video, how creators frame it, and the top videos. Deep-dive samples are about 75 videos: the 45 most viewed plus 30 spread across the month.
Product reports
Product pages cover products from panel brands that people search for on Google and that appear in enough short-form video to analyze. Each product search combines the brand name with the product name. Before a product page is published, we check a sample of its videos, and at least 80% must be about that product. The page’s analysis uses about 50 videos (the 30 most viewed plus 20 spread across the month) and notes the creator’s stance, the claims and uses mentioned, any products named as alternatives, and any complaints.
Comparison reports
Comparison pages cover pairs that people search for as “A vs B”. They put both sides’ measures next to each other and add a sample of up to 40 videos that mention both (the 25 most viewed plus 15 spread across the month), labelled with which side the creator prefers, if any, and why. These pages report what creators say. We don’t rate the products ourselves.
What’s open and what needs an email
Rankings, scores, component scores and analysis are open to everyone. Some lists of example videos show the first few and ask for an email address to see the rest.
Limits
SSI measures attention and reaction in short-form video. It doesn’t measure product quality, sales or market share.
The library doesn’t include every public video, and it has no private videos or videos from other platforms.
View counts are recorded when we collect the data, so videos posted late in the month have had less time to gather views.
Sentiment and sponsorship detection depend on what’s said and disclosed. A partnership that isn’t disclosed counts as organic.
Searches can over-count or under-count a name. We don’t search for MAC on its own because of MacBook posts, for example, so MAC is likely under-counted. We note limits like this on the affected report.
Independence and corrections
Syncly plans and publishes the index on its own. Brands can’t pay to be included, change their scores or remove results. Every claim about a specific video links to that video so it can be checked. If we find an error, we fix the page and note the correction on it.
How to cite
When you cite the index, name the measure (video mentions, share of voice or SSI score) and the period, for example: Source: Syncly Social Index, September 2026 (syncly.app/social-index).
The most-mentioned brand (video count) and the top-ranked brand (SSI score) can be different brands.
Changes to this methodology
October 2026: This page was expanded with panel selection, scoring steps, sampling, and how trend, product and comparison reports are built. Definitions and weights didn’t change.
FAQ
How is the Syncly Social Index calculated?
Six measures of a brand’s short-form video presence are each ranked within a 10-brand category panel and combined with fixed weights: video share of voice 25, untagged share 20, momentum 15, organic sentiment 15, high reach 15 and creator breadth 10. The result is a score from 0 to 100.
How are the brands in each ranking chosen?
Each month, the 10 brands from our candidate list with the most short-form video mentions in a category form its panel. Panels are chosen again every month.
Is the brand with the most videos always ranked first?
No. Video volume is one of six components. In September 2026 Medicube had the most skincare videos, while Anua ranked first overall.
Can brands pay to be included or ranked higher?
No. Panels and scores come only from video data.
What is an untagged mention?
A video where a brand is shown or spoken about but never named in the caption, hashtags or handles.
Can I compare SSI scores across categories?
No. Each score is relative to the other brands in the same category panel that month.
How often is the index updated?
Monthly. Past editions stay available in the edition archive.
