Operationalall signals operational

Snapshot Oct 11, 06:04 UTCfetched 2m ago

The numbers

Yeah reads YouTube’s running-review creators at scale, then distills the consensus into a graded catalog. Here’s the whole funnel — from raw reviews to buyable verdicts.

Updated Oct 11, 2026

1 / Scale

What we’ve watched

The whole corpus — every review ingested, across every creator we follow.

5.3K

Reviews analyzed

How this is computed

Videos analyzed

Formula: count of sources where source_type = 'youtube_video'
Source: sources table
Window: all_time
Missing data: zero_when_empty
Freshness SLA: 1d 2h

Code

len([r for r in select(Source.source_metadata).where(Source.source_type == 'youtube_video')])

1034h

Hours watched

How this is computed

Video hours measured

Formula: sum(source_metadata.duration_s for youtube_video) / 3600, rounded 2dp
Source: sources table (source_metadata.duration_s JSON)
Window: all_time
Missing data: zero_when_empty
Freshness SLA: 1d 2h

Code

round(sum(duration_s) / 3600.0, 2)

30.3K

Product mentions

How this is computed

Total mentions

Formula: count of all mentions rows
Source: mentions table
Window: all_time
Missing data: zero_when_empty
Freshness SLA: 1d 2h

Code

select(func.count()).select_from(Mention)

111

Channels watched

How this is computed

Creator channels tracked

Formula: count of all creator_channels rows
Source: creator_channels table
Window: all_time
Missing data: zero_when_empty
Freshness SLA: 1d 2h

Code

select(func.count()).select_from(CreatorChannel)

18.2M

Channel reach

subscribers across every channel watched (overlapping)

How this is computed

Total subscribers across channels

Formula: sum(creator_channels.subscriber_count), coalesced to 0
Source: creator_channels table
Window: all_time
Missing data: zero_when_empty
Freshness SLA: 1d 2h

Code

select(func.coalesce(func.sum(CreatorChannel.subscriber_count), 0))

4.6B

Lifetime views

across all creators

How this is computed

Total channel views

Formula: sum(creator_channels.view_count), coalesced to 0
Source: creator_channels table
Window: all_time
Missing data: zero_when_empty
Freshness SLA: 1d 2h

Code

select(func.coalesce(func.sum(CreatorChannel.view_count), 0))

2 / Graded

The published catalog

Only shoes with enough evidence earn a public verdict — scored, tiered, and written up.

353

Shoes tracked

How this is computed

Published products with a verdict

Formula: count of products in the PUBLISHED scope (published=true AND deleted_at IS NULL) that have a latest-per-product verdict
Source: verdict_snapshots latest-per-product, filtered to published_ids
Window: all_time
Missing data: zero_when_empty
Freshness SLA: 1d 2h

Code

sum(1 for pid in latest if pid in published_ids)

Brands

How this is computed

Distinct published brands

Formula: count(distinct brand) over products where published = true and brand is not null
Source: products table
Window: all_time
Missing data: zero_when_empty
Freshness SLA: 1d 2h

Code

select(func.count(func.distinct(Product.brand))).where(Product.published.is_(True)).where(Product.brand.is_not(None))

Average score

of 100

How this is computed

Average verdict score (published)

Formula: mean(latest verdict.score) over published-scope scored products; N/A when none
Source: verdict_snapshots latest-per-product, filtered to published_ids
Window: all_time
Missing data: na_when_denominator_zero
Freshness SLA: 1d 2h

Code

round(score_sum / n, 2) if n else None

25.5

Mentions per shoe

How this is computed

Average mentions per scored product (published)

Formula: mean(latest verdict.mention_count) over published scored products; N/A when none
Source: verdict_snapshots latest-per-product, filtered to published_ids
Window: all_time
Missing data: na_when_denominator_zero
Freshness SLA: 1d 2h

Code

round(mention_sum / n, 2) if n else None

4.1

Creators per shoe

How this is computed

Average creators per scored product (published)

Formula: mean(latest verdict.creator_count) over published scored products; N/A when none
Source: verdict_snapshots latest-per-product, filtered to published_ids
Window: all_time
Missing data: na_when_denominator_zero
Freshness SLA: 1d 2h

Code

round(creator_sum / n, 2) if n else None

S11
A39
B171
C95
D–F37

3 / Buyable

Price bridge

For graded shoes we track live prices across retailers, so a verdict comes with a place to buy.

352

Shoes priced

How this is computed

Products with a current price

Formula: count of product_current_prices rows
Source: product_current_prices table
Window: all_time
Missing data: zero_when_empty
Freshness SLA: 1d 2h

Code

select(func.count()).select_from(ProductCurrentPrice)

1.4K

Retailers tracked

How this is computed

Distinct priced retailers

Formula: count(distinct retailer) over price_snapshots
Source: price_snapshots table
Window: all_time
Missing data: zero_when_empty
Freshness SLA: 1d 2h

Code

select(func.count(func.distinct(PriceSnapshot.retailer)))

Signal

What reviewers actually said

30.3K sentiment-tagged mentions, split by stance.

Worth it44%
Mixed14%
Skip6%
Neutral35%

Pulse

Still running

The pipeline keeps ingesting — here’s the recent activity.

2.2K

New reviews

published last 90 days

How this is computed

Videos published in the last 90 days

Formula: count of youtube_video sources with published_at >= now - 90d
Source: sources table (source_metadata.published_at JSON)
Window: last_90d
Missing data: zero_when_empty
Freshness SLA: 1d 2h

Code

sum(1 for p in published_at if p >= now - timedelta(days=90))

173

Videos ingested

last 7 days

How this is computed

Videos ingested in the last 7 days

Formula: count of youtube_video sources where created_at >= now - 7d
Source: sources table
Window: last_7d
Missing data: zero_when_empty
Freshness SLA: 1d 2h

Code

select(func.count()).where(Source.source_type == 'youtube_video').where(Source.created_at >= now - timedelta(days=7))

976

Mentions extracted

last 7 days

How this is computed

Mentions in the last 7 days

Formula: count of mentions where created_at >= now - 7d
Source: mentions table
Window: last_7d
Missing data: zero_when_empty
Freshness SLA: 1d 2h

Code

select(func.count()).where(Mention.created_at >= now - timedelta(days=7))

191

Verdicts written

last 7 days

How this is computed

Editorials generated in the last 7 days

Formula: count of editorial_reviews where generated_at >= now - 7d
Source: editorial_reviews table
Window: last_7d
Missing data: zero_when_empty
Freshness SLA: 1d 2h

Code

select(func.count()).where(EditorialReview.generated_at >= now - timedelta(days=7))

Over time

Growing

Tracking since Jul 14, 2026; updates daily.

Reviews analyzed5.3K

+2.2K since tracking began

Product mentions30.3K

+9.8K since tracking began

Shoes tracked353

+57 since tracking began