Research YouTube Keyword Competition (2026 Guide)

Thirdwatch's YouTube Scraper makes keyword-competition research a structured workflow — top-10 view stats, creator dominance, view-velocity dispersion, opportunity scoring across hundreds of keywords. Built for content creators, video-marketing teams, creator-economy SaaS, and SEO-research platforms that need raw YouTube data.

Why research YouTube keyword competition

YouTube is the dominant educational and how-to discovery surface. According to Google's 2024 YouTube Watch report, more than 70% of YouTube views originate from search and recommendation surfaces driven by keyword relevance. For content creators and video-marketing teams, choosing keywords with real demand and non-dominant competition is the single highest-leverage strategy decision — getting it right unlocks discovery; getting it wrong wastes months of production effort.

The job-to-be-done is structured. A YouTube creator-strategy agency wants 200 client niches × 50 keywords each = 10K keywords competition-scored quarterly. A video-marketing team launching a new educational series wants per-keyword competitiveness benchmarks before content investment. A creator-economy SaaS builder powers keyword-research dashboards for their creator users. An SEO-research platform extends YouTube SERP analysis into broader video-keyword competitive intelligence. All reduce to keyword list + top-10 video pull + competition aggregation per keyword.

How does this compare to the alternatives?

Three options for YouTube keyword-competition data:

TubeBuddy and VidIQ are the SaaS market leaders for individual creators. YouTube Data API v3 is the official path but the daily quota (10K units, with each search costing 100 units = 100 searches/day) makes high-volume research impractical. The YouTube Scraper actor page gives you unlimited keyword research at the lowest unit cost.

How to research keyword competition in 4 steps

Step 1: How do I authenticate against Apify?

Sign in at apify.com (free tier, no credit card), open Settings → Integrations, and copy your personal API token. Every example below assumes the token is in APIFY_TOKEN:

export APIFY_TOKEN="apify_api_xxxxxxxxxxxxxxxx"

Step 2: How do I pull a keyword-watchlist batch?

Pass keyword queries with searchType: "videos" and maxResultsPerQuery: 10.

import os, requests, pandas as pd

ACTOR = "thirdwatch~youtube-scraper"
TOKEN = os.environ["APIFY_TOKEN"]

KEYWORDS = ["how to start a youtube channel",
            "passive income ideas 2026",
            "morning routine productive",
            "kettlebell workout for beginners",
            "react tutorial 2026",
            "tax tips for freelancers",
            "easy meal prep for week",
            "how to learn guitar fast",
            "rust programming language tutorial",
            "minimalist home tour"]

resp = requests.post(
    f"https://api.apify.com/v2/acts/{ACTOR}/run-sync-get-dataset-items",
    params={"token": TOKEN},
    json={"queries": KEYWORDS, "searchType": "videos",
          "maxResults": 100, "maxResultsPerQuery": 10},
    timeout=900,
)
df = pd.DataFrame(resp.json())
print(f"{len(df)} videos across {len(KEYWORDS)} keywords")

10 keywords × 10 top videos = up to 100 records — affordable for an ad-hoc keyword sweep.

Step 3: How do I compute per-keyword competition metrics?

Aggregate top-10 stats and compute HHI for creator dominance.

df["viewCount"] = pd.to_numeric(df.viewCount, errors="coerce")
df["createdAt"] = pd.to_datetime(df.createdAt, errors="coerce")
df["age_days"] = (pd.Timestamp.utcnow() - df.createdAt).dt.days.clip(lower=1)
df["views_per_day"] = df.viewCount / df.age_days

def channel_hhi(rows):
    shares = rows.channelName.value_counts(normalize=True)
    return float((shares ** 2).sum())

competition = (
    df.groupby("searchString")
    .agg(
        median_views=("viewCount", "median"),
        median_views_per_day=("views_per_day", "median"),
        recent_count=("age_days", lambda x: (x <= 365).sum()),
        creator_hhi=("channelName", lambda x: float((x.value_counts(normalize=True) ** 2).sum())),
    )
    .assign(
        opportunity=lambda d: (
            (d.median_views >= 100000)
            & (d.creator_hhi < 0.4)
            & (d.recent_count >= 3)
        ).astype(int)
    )
    .sort_values(["opportunity", "median_views"], ascending=[False, False])
)
print(competition)

The three-condition opportunity filter selects keywords with real demand, non-dominant creators, and an active recent-content base.

Step 4: How do I track competitive shifts over time?

Snapshot weekly, compute median-views and HHI deltas to detect entry-exit dynamics.

import pathlib, datetime

ts = datetime.datetime.utcnow().strftime("%Y%m%d")
competition.to_csv(f"snapshots/youtube-keywords-{ts}.csv")

# Compare against last snapshot
prev = pd.read_csv("snapshots/youtube-keywords-20260420.csv", index_col=0)
combined = competition.merge(prev, left_index=True, right_index=True, suffixes=("", "_prev"))
combined["views_delta"] = combined.median_views - combined.median_views_prev
combined["hhi_delta"] = combined.creator_hhi - combined.creator_hhi_prev

# Newly-tractable keywords: HHI dropping (less concentrated) + views rising
emerging = combined[
    (combined.hhi_delta <= -0.05)
    & (combined.views_delta >= 50000)
]
print(f"{len(emerging)} keywords with newly-tractable competition")
print(emerging[["median_views", "creator_hhi", "hhi_delta", "views_delta"]])

A keyword where HHI is dropping while median views rise is a competitive opening — established creators are losing share to new entrants on growing demand.

Sample output

A single YouTube video record looks like this. Five rows weigh ~6 KB.

{
  "videoId": "abc123XYZ",
  "title": "How to Start a YouTube Channel in 2026 (Complete Guide)",
  "url": "https://www.youtube.com/watch?v=abc123XYZ",
  "channelName": "Roberto Blake",
  "channelUrl": "https://www.youtube.com/@robertoblake",
  "channelSubscribers": 720000,
  "viewCount": 245000,
  "likeCount": 18500,
  "commentCount": 1240,
  "duration": 845,
  "createdAt": "2026-02-15T14:00:00Z",
  "thumbnailUrl": "https://i.ytimg.com/vi/abc123XYZ/...",
  "description": "Complete guide to starting a YouTube channel..."
}

videoId is YouTube's globally unique video identifier — the canonical key for cross-snapshot dedup. viewCount and createdAt feed the views-per-day computation that's central to competition analysis. channelSubscribers provides the creator-tier context for normalized engagement-rate analysis. duration (seconds) lets you filter to long-form vs Shorts (<60s) since the two surfaces have separate algorithm dynamics.

Common pitfalls

Three things go wrong in YouTube keyword-research pipelines. Search-result personalization — YouTube personalizes results based on viewer history; the actor's IP sees a relatively neutral feed but for advertiser-relevant audience research, supplement with first-party YouTube Analytics from owned channels. Shorts vs long-form mixing — YouTube Shorts (under 60s) have separate distribution dynamics and engagement norms; for long-form keyword research, filter out duration < 60 rows before computing competition metrics. Recommendations-driven dilution — top-10 search results sometimes include videos that ranked due to recommendation-feed boost rather than search-keyword relevance; for keyword-pure analysis, cross-check that the keyword appears in the video title or description.

Thirdwatch's actor uses a lightweight HTTP path so you pay only for the data, not for proxy or compute overhead. Pair YouTube with TikTok Scraper for cross-platform short-form research and Google Search Scraper for keyword-funnel analysis (search SERP → YouTube SERP). A fourth subtle issue worth flagging: YouTube's algorithm increasingly preferences "session length" optimization over per-video views, which means high-views-per-day on individual videos doesn't always translate to channel-level growth — a creator's overall channel-velocity matters more for sustainable competition than any single hit. Cross-reference per-keyword rankings with channel-subscriber growth from snapshot deltas to identify creators who are systematically winning vs creators with one viral hit. A fifth pattern unique to YouTube keyword research: thumbnail and title A/B testing means the same video shows up with different titles across snapshots; for stable cross-snapshot analysis, dedupe on videoId rather than title and treat title-changes as a separate signal worth tracking. A sixth and final pitfall: certain keyword-niches have heavy regional dominance (Hindi tech tutorials, Spanish cooking content) where the top-10 reflects regional language rather than English-dominant competition; for cross-region competitive mapping, supplement keyword search with explicit language qualifiers (e.g., "react tutorial in Hindi") to capture region-specific opportunity surfaces.

Related use cases

• Scrape YouTube search and channel data
• Track YouTube channel growth and views
• Build YouTube content trend pipeline
• The complete guide to scraping social media
• All Thirdwatch use-case guides