YouTube Audience Overlap: Shared Commenters Between Channels

For brands, agencies and creator teams comparing YouTube channels before a sponsorship. About $0.03 per channel sample, with every pairwise comparison included. Two channels cost about $0.06; five cost about $0.15 and produce ten comparisons. No login is required.

Which public commenters appear in both channels' recent video samples? Get shared counts, directional percentages, Jaccard overlap and up to 50 shared handles. This measures commenter samples, not all viewers, subscribers or duplicated campaign reach.

What you get

Each row compares two channels, with a verdict first. The output also links a summary and an HTML report.

Field	Meaning
verdict	Plain-English answer with a whole-number percentage and shared-commenter count.
overlapLabel, overlapBasisPercent	Descriptive pilot label and the larger directional percentage used for that label.
recordType	channel_pair for a comparison, or note when no comparison is available.
channelA, channelB	Public handles identifying the compared channels.
channelAId, channelBId	Stable public channel IDs for the input creators.
channelAUrl, channelBUrl	Canonical public channel links.
sharedCommenters	Number of distinct commenter channel IDs found in both samples.
uniqueCommentersA, uniqueCommentersB	Distinct commenter counts after deduplication within each channel.
unionCommenters	Distinct commenters found in either sample, counting shared people once.
shareOfAPercent, shareOfBPercent	Shared commenters divided by the corresponding channel's sample, multiplied by 100.
jaccardPercent	Shared commenters divided by the union of both samples, multiplied by 100.
sharedCommenterHandles	Up to 50 public handles from the intersection. Comment text and contact details are not returned.
sharedHandlesTruncated	True when the handle list represents fewer people than the shared count, because of the 50-handle limit or missing handles.
videosSampledA, videosSampledB	Number of selected regular uploads checked for each channel.
commentsSampledA, commentsSampledB	Top-level comments collected before deduplicating their authors.
videoSamplesA, videoSamplesB	Source video links, comment counts, unique authors, pages read, exhaustion and availability notes.
sampledAt, notes	Collection time and interpretation limits.

SUMMARY lists channel outcomes, sample sizes, settings, comparisons and charges. REPORT.html presents comparison and coverage tables. Unavailable inputs produce free notes.

Useful jobs

Compare two tech reviewers before booking both, look for shared commenters across a five-creator fitness shortlist, or compare science channels whose topics appear similar.

Silent viewers are outside this measurement. Low commenter overlap can coexist with many shared viewers.

Run it in three steps

1. Enter two to ten YouTube handles or channel URLs. You can mix formats, for example @mkbhd and https://www.youtube.com/@LinusTechTips.
2. Choose the sample size. The default is five recent videos and up to 200 comments per video. The prefilled quick example uses three videos for each of two channels.
3. Start the Actor and open the comparison dataset or report. Check the sample counts and video notes before interpreting an overlap percentage.

The saved-task quick example completed in 35.4 seconds at 256 MB on 15 October 2026. A separate two-channel default sample of five videos and 200 comments per video completed in 38.6 seconds. Runtime varies with source latency, retries, sample size and availability.

Inputs

Input	Default	Accepted values and behavior
channels	Two prefilled example handles	Two to ten public handles, channel URLs or channel IDs. This is the preferred input.
videosPerChannel	5	Integer from 1 to 10. Selects recent public regular uploads from the channel's Videos tab. The quick prefill uses 3.
commentsPerVideo	200	Integer from 1 to 300. Maximum top-level comments collected from each selected video.
urls	Omitted	Optional channel-URL list when channels is omitted.
startUrls	Omitted	Optional list of URL objects when both channels and urls are omitted.

Example:

{
  "channels": ["@GamersNexus", "@Hardwareunboxed"],
  "videosPerChannel": 5,
  "commentsPerVideo": 200
}

Inputs resolve to stable channel IDs. Repeating a channel in different formats does not create another charge. Invalid references produce notes; only the first ten inputs are used. Invalid numerical settings fall back to documented defaults, with a summary note.

A real result

This abbreviated row came from the 15 October 2026 cloud test, using five videos and up to 200 comments per video:

{
  "verdict": "2% of @GamersNexus's sampled commenters also appear on @Hardwareunboxed (15 shared commenters; higher sampled overlap).",
  "recordType": "channel_pair",
  "overlapLabel": "higher sampled overlap",
  "channelA": "@GamersNexus",
  "channelB": "@Hardwareunboxed",
  "sharedCommenters": 15,
  "uniqueCommentersA": 964,
  "uniqueCommentersB": 950,
  "unionCommenters": 1899,
  "jaccardPercent": 0.7899,
  "shareOfAPercent": 1.556,
  "shareOfBPercent": 1.5789,
  "videosSampledA": 5,
  "videosSampledB": 5,
  "commentsSampledA": 1000,
  "commentsSampledB": 1000
}

The verdict rounds percentages to whole numbers for readability. Numeric dataset fields retain four decimal places, so a verdict displaying 0% can still have a small nonzero overlap. Later runs can differ as comments and recent uploads change.

How overlap is calculated

The Actor reads public channel and watch pages, then the signed-out continuation endpoint used for the watch page's comment section. It follows top-level comment pagination up to the requested cap, using YouTube's top-comment ordering. Replies are excluded.

Each person contributes one identity per channel, matched by public channel ID. Repeated comments are deduplicated. Identical display names are never merged; handles are labels, not matching keys.

For samples A and B, the shared count is their intersection. A's percentage is the shared count divided by A's unique commenters; B's percentage uses B's denominator. Jaccard uses all distinct commenters in either sample. Directional percentages can differ when the sample sizes differ. None of these figures is a demographic estimate or an industry benchmark.

The Actor checks recent regular uploads visible on the Videos tab. Visibly marked members-only, upcoming and live videos are excluded. Shorts and reply threads are outside the sample. If a selected video's comments are disabled or unavailable, its video-level note explains that and its count is zero. A network failure after retries makes that channel sample unavailable rather than silently presenting it as complete.

What the label means

The label uses the larger of the two directional percentages. Zero shared commenters means no observed overlap. Positive overlap below 1% is low sampled overlap; 1% or more is higher sampled overlap. If either channel has fewer than 200 distinct sampled commenters, a positive result is labeled small sample instead. The numeric percentages remain available in every comparison.

These are pilot rules, not industry norms. In a 15 October 2026 test of ten tech, fitness and science channels, using five videos and 200 comments per video, all 45 pairs returned data. Thirteen pairs had no observed overlap, 31 were positive but below 1%, and one reached 1.67%; the next highest was 0.72%. The 1% boundary separates that observed gap. The 200-commenter guard is a conservative display rule, not a statistical confidence threshold. Neither label measures whole-audience overlap or adjusts for channel size.

Pricing and spending limits

The list price is $0.03 per successful channel sample used in a comparison. Comparisons between those samples add no charge. The number of pairs is the channel count multiplied by one less than that count, divided by two.

Successful channels	Included pairs	Approximate sample charge
2	1	$0.06
3	3	$0.09
5	10	$0.15
10	45	$0.30

Apify's small Actor start charge is additional: $0.00005 per start unit, with one unit at the default memory setting. Channel-sample prices are $0.0285 on Bronze, $0.027 on Silver and $0.024 on Gold. The displayed Store price reflects your plan. There is no separate dataset-row charge.

Failed, empty, duplicate and unpaired channel samples are not charged. An observed zero-overlap comparison is useful data and its successful channel samples are chargeable. At least two distinct usable samples are needed. If the spending limit cannot fund two samples, the Actor returns a note without a channel-sample charge. If only one channel succeeds, its sample is also free. The start charge can still apply to those runs.

For larger lists, the Actor limits collection to the number of samples the remaining budget can cover. The summary identifies channels skipped because of the spending limit. Free-plan users receive the same available sample settings; no external paid data service or AI API is used.

API examples

Use your own Apify API token and standard Bearer authentication. This example returns the dataset directly:

curl -X POST \
  'https://api.apify.com/v2/acts/vant47~youtube-audience-overlap/run-sync-get-dataset-items' \
  -H "Authorization: Bearer $APIFY_TOKEN" \
  -H 'Content-Type: application/json' \
  -d '{"channels":["@mkbhd","@LinusTechTips"],"videosPerChannel":3,"commentsPerVideo":200}'

Python:

import os
import requests

url = "https://api.apify.com/v2/acts/vant47~youtube-audience-overlap/run-sync-get-dataset-items"
response = requests.post(url,
    headers={"Authorization": "Bearer " + os.environ["APIFY_TOKEN"]},
    json={"channels": ["@mkbhd", "@LinusTechTips"],
          "videosPerChannel": 3, "commentsPerVideo": 200},
    timeout=300)
response.raise_for_status()
rows = response.json()

JavaScript:

const response = await fetch(
  "https://api.apify.com/v2/acts/vant47~youtube-audience-overlap/run-sync-get-dataset-items",
  {
    method: "POST",
    headers: {
      Authorization: `Bearer ${process.env.APIFY_TOKEN}`,
      "Content-Type": "application/json"
    },
    body: JSON.stringify({
      channels: ["@mkbhd", "@LinusTechTips"],
      videosPerChannel: 3,
      commentsPerVideo: 200
    })
  }
);
if (!response.ok) throw new Error(`Apify returned ${response.status}`);
const rows = await response.json();

For larger jobs, use Apify's asynchronous run endpoint, poll the run status and read the default dataset when it completes. The run's default key-value store contains SUMMARY and REPORT.html.

Use with AI agents (MCP)

Connect your agent to https://mcp.apify.com and request a YouTube audience overlap comparison based on public commenters. Give it the channel list, desired video and comment caps, and maximum spend. Ask it to include sample sizes and the limitation that commenters are not the whole audience.

Example request: Compare the public commenting audiences of these five YouTube channels before a sponsorship. Return shared-commenter counts, directional percentages and Jaccard overlap, using five videos and 200 comments per video. Include all pairs and explain the sampling limitations.

Agents should read recordType: a successful run can contain only a free note when inputs or budgets prevent a comparison.

Repeat runs and integrations

Save a task and use Apify schedules, webhooks, Make, Zapier, n8n or spreadsheets for repeated comparisons. No schedule is created automatically. Keep settings and source links when comparing dates. The Actor does not merge samples across runs.

Limits and FAQ

Is this total audience overlap? No. It measures the intersection of public commenter samples. It cannot see private viewing history, silent viewers, subscriber relationships or unique campaign reach.

Does zero mean the channels have different audiences? No. A limited sample may miss shared people. Consider the unique-commenter counts, sample settings and video selection alongside the percentage.

Are the comments random? No. YouTube's top ordering favors some comments. Creator moderation, language, video topic and publication timing also affect who appears. The Actor makes no claim of statistical representativeness.

Why are there fewer commenters than comments? People may leave several comments or appear on multiple sampled videos. Comment counts describe messages; unique-commenter counts describe distinct public channel IDs.

Why is a channel missing? It may not exist, may have no public regular uploads, may expose no comments, or may remain unavailable after retries. The summary gives the reason. No unavailable sample is charged.

Can I get every shared handle? The shared count uses the full collected intersection, but the output includes at most 50 available public handles. Increasing the comment cap does not lift that output limit.

Do I need a YouTube login, API key or residential proxy subscription? No. The Actor uses public signed-out reads with managed retries. It does not accept account cookies, login credentials or external service keys.

Can I use it for outreach? The output is intended for audience comparison. It provides no email addresses, telephone numbers, comment text or private identity enrichment. Use public data consistently with applicable rights, platform terms and your purpose.

Search phrases: YouTube audience overlap, shared commenters between channels, public commenter intersection, YouTube sponsorship audience comparison.