ChatGPT answers, unedited

For an educational YouTube channel, you’ll usually want a tool stack that can answer questions like:

• “Do biology tutorials uploaded in 2025 retain viewers better than exam-prep videos?”
• “Which upload cohorts produce the most watch time after 30/90/180 days?”
• “Are newer viewers becoming returning viewers?”
• “Which categories create subscribers, not just views?”

Here are the best options depending on how advanced you want to get:

My recommended setup for an educational channel

1. Start with YouTube Studio Advanced Mode

• Create video groups such as:
  • Algebra
  • Physics
  • Study skills
  • Exam preparation
  • Beginner vs advanced lessons
• Compare:
  • Watch time
  • Average view duration
  • Returning viewers
  • Subscribers gained
  • Traffic source

YouTube Studio’s Advanced Mode supports comparisons, groups, breakdowns, metrics, filters, and exports. support.google.com

2. Build a cohort dashboard in Looker Studio
A useful schema:

Key charts:

• Watch hours by upload month
• Median watch time per video by category
• Subscriber conversion by topic
• Retention curve by category
• “Evergreen score” (watch time after 90+ days)

YouTube’s reporting APIs expose metrics such as estimated minutes watched, average view duration, views, and subscribers gained, which can feed custom dashboards. developers.google.com

3. Add a spreadsheet/database layer if you have many videos
For 100+ uploads:

• Export YouTube Analytics data weekly
• Store snapshots in:
  • Google Sheets (simple)
  • BigQuery (advanced)
  • PostgreSQL (technical teams)

Then analyze:

• “Videos uploaded in Q1 2026”
• “Beginner lessons vs advanced lessons”
• “First 30 days vs lifetime performance”

If you want the simplest good workflow:

YouTube Studio → Google Sheets → Looker Studio

That gives you most of what a creator needs without paying for enterprise software.

If your channel is large (100k+ subscribers)

I’d consider:

• BigQuery + Looker Studio for custom cohort modeling
• Tableau/Power BI if you have a data analyst
• A creator analytics platform if you need competitor benchmarking

One note: YouTube’s built-in Audience reports already separate viewer behavior into groups like new, casual, and regular viewers, which is useful for measuring whether educational content is building a repeat audience. support.google.com

If you tell me your channel size (e.g., 10k vs 500k subscribers) and whether you categorize videos manually (topics/playlists/tags), I can suggest a more specific stack.

For an educational YouTube channel, I’d use YouTube Analytics + Looker Studio as the default stack.

My recommendation

1. YouTube Studio Analytics — use this as your source of truth. Advanced Mode lets you compare videos/groups/time periods, change dimensions and metrics, filter audiences, and export data. Google Help Google Help

support.google.com

2. Looker Studio — use this for the actual cohort analysis/dashboard. Google provides a native YouTube Analytics connector, so you can connect your channel without building an API pipeline. Google Cloud Documentation Google Cloud Documentation

lookerstudio.google.com

3. Google Sheets or BigQuery — add one of these if you want to build a more sophisticated historical dataset. BigQuery becomes worthwhile once you're analyzing a large library or want to join YouTube data with things like email signups, course sales, or website behavior.

The analysis I'd build

For your specific question, I wouldn't just look at YouTube's standard "watch time by video." I'd create a video-upload cohort model:

Then add a viewer cohort dimension, such as:

• New viewers
• Casual viewers
• Regular viewers
• Subscriber vs. non-subscriber
• Ideally, the time since the video was published: 0–7 days, 8–30, 31–90, 91–180, 180+

That lets you answer much more useful questions like:

"Do my biology videos generate more long-term watch time than my math videos?"

rather than simply:

"Which videos got the most views?"

YouTube itself distinguishes new, casual, and regular viewers, and its Analytics data can be broken down by content, date, audience, and other dimensions. Google Help Google for Developers

One important caveat

If by "viewer cohorts by upload date" you mean "people who first watched my channel during the week/month a particular video was uploaded, and then how much they watched afterward," that's more sophisticated than the standard YouTube Studio reports.

In that case, I'd build a small YouTube Analytics API → BigQuery → Looker Studio pipeline rather than relying solely on the Studio interface. YouTube exposes time-period and video-level dimensions through its Analytics API, which makes this kind of longitudinal analysis much more feasible. Google for Developers

So my ranking:

• 🥇 YouTube Studio + Looker Studio — best starting point, inexpensive and flexible.
• 🥈 YouTube Analytics API + BigQuery + Looker Studio — best if you're serious about cohort analysis and have lots of videos.
• 🥉 Supermetrics/Coupler-style connector + Looker Studio — easier if you don't want to build the data pipeline yourself; third-party connectors can provide additional extraction capabilities. Coupler Blog

If you tell me roughly how many videos you have and what you mean by "viewer cohort" (e.g. new viewers in the month of upload vs. viewers who watched a particular upload), I can recommend the exact setup and the dimensions/metrics I'd use.

"Which videos got the most views?"

YouTube itself distinguishes new, casual, and regular viewers, and its Analytics data can be broken down by content, date, audience, and other dimensions. Google Help developers.google.com

One important caveat

If by "viewer cohorts by upload date" you mean "people who first watched my channel during the week/month a particular video was uploaded, and then how much they watched afterward," that's more sophisticated than the standard YouTube Studio reports.

In that case, I'd build a small YouTube Analytics API → BigQuery → Looker Studio pipeline rather than relying solely on the Studio interface. YouTube exposes time-period and video-level dimensions through its Analytics API, which makes this kind of longitudinal analysis much more feasible.

So my ranking:

If your goal is specifically “How does watch time behave for cohorts of videos based on when they were uploaded, and how does that differ by educational category?”, I’d use a YouTube Analytics + Looker Studio setup rather than a creator-SEO tool.

My recommendation

1. Best overall: YouTube Analytics + Looker Studio

studio.youtube.com is the source of truth for your channel data. Its Advanced Mode lets you compare videos, demographics, dates, content, traffic sources, and other dimensions, and export the data. Google Help

Then use lookerstudio.google.com to build the cohort dashboard.

I'd structure your data around:

The important trick is age since upload. Instead of asking “How much watch time did videos uploaded in March get?”, ask:

“How much watch time did the March cohort accumulate during its first 7/30/90 days?”

That makes older and newer educational videos genuinely comparable.

YouTube's reporting APIs expose video-level and daily watch-time data, including watch_time_minutes, average view duration, percentage viewed, views, and subscriber gains. Google for Developers Google for Developers

2. If you want the most powerful setup: YouTube API → BigQuery → Looker Studio

For a serious channel, this is what I'd eventually build.

YouTube Analytics API → BigQuery → Looker Studio

YouTube's Reporting API can provide bulk video analytics with date and video_id dimensions, while the Analytics API supports targeted queries. Google for Developers Google for Developers

You'd maintain a video metadata table like:

video_id
title
upload_date
category
length_minutes
topic
format

And a daily performance table:

video_id
date
views
watch_time_minutes
avg_view_duration
avg_view_percentage
subscribers_gained

Then calculate:

days_since_upload = date - upload_date

That gives you extremely useful cohort analyses:

• 0–7 day watch time by category
• 0–30 day watch time by upload month
• 30–90 day watch time by category
• Watch-time curve by video age
• Median vs. average performance by cohort
• Evergreen vs. launch-dependent videos
• Category × video age
• Upload month × category
• Videos that continue accumulating watch time after 90 days

This is much more useful than simply ranking videos by lifetime watch time.

Where vidIQ fits

vidiq.com is useful if you also want topic research, competitor analysis, SEO, and channel-growth benchmarking. Its current research products, for example, analyze large datasets of YouTube channels and uploads. vidIQ vidIQ

But I wouldn't make vidIQ the primary tool for your cohort analysis. Your own YouTube Analytics data is more appropriate for answering questions about your audience's watch behavior.

What I'd choose

• Small/medium educational channel: YouTube Studio + Looker Studio
• Growing channel with lots of videos: YouTube Analytics API + BigQuery + Looker Studio
• Need SEO/topic/competitor intelligence too: Add vidIQ
• Want the simplest possible solution: Start with YouTube Studio Advanced Mode and export to Sheets

If you tell me roughly how many videos you have (e.g. 50, 500, 5,000) and what you mean by “category” (YouTube's category vs. your own subjects like math/history/programming), I can recommend the exact setup and the cohort dashboard I'd build for an educational channel.

I'd structure your data around:

The important trick is age since upload. Instead of asking “How much watch time did videos uploaded in March get?”, ask:

“How much watch time did the March cohort accumulate during its first 7/30/90 days?”

That makes older and newer educational videos genuinely comparable.

YouTube's reporting APIs expose video-level and daily watch-time data, including watch_time_minutes, average view duration, percentage viewed, views, and subscriber gains. Google for Developers Google for Developers

2. If you want the most powerful setup: YouTube API → BigQuery → Looker Studio

For a serious channel, this is what I'd eventually build.

YouTube Analytics API → BigQuery → Looker Studio

YouTube's Reporting API can provide bulk video analytics with date and video_id dimensions, while the Analytics API supports targeted queries. Google for Developers Google for Developers

You'd maintain a video metadata table like:

video_id
title
upload_date
category
length_minutes
topic
format

And a daily performance table:

video_id
date
views
watch_time_minutes
avg_view_duration
avg_view_percentage
subscribers_gained

Then calculate:

days_since_upload = date - upload_date

That gives you extremely useful cohort analyses:

• 0–7 day watch time by category
• 0–30 day watch time by upload month
• 30–90 day watch time by category
• Watch-time curve by video age
• Median vs. average performance by cohort
• Evergreen vs. launch-dependent videos
• Category × video age
• Upload month × category
• Videos that continue accumulating watch time after 90 days

This is much more useful than simply ranking videos by lifetime watch time.

Where vidIQ fits

For what you’re describing, I’d use YouTube Analytics as the data source, then put the data into a BI tool rather than relying on a standalone YouTube analytics SaaS.

My picks

My recommendation: Looker Studio + YouTube Analytics/API.

YouTube Studio's Advanced Mode already lets you compare videos/groups/time periods, change dimensions and metrics, filter data, save reports, and export views. You can also create groups of up to 500 videos—which is particularly useful for categorizing educational content. support.google.com

The key, though, is to create your own upload cohorts.

For example:

• Upload cohort: Jan–Mar 2026
• Category: Algebra
• Videos: 18
• Days since upload: 0–7 / 8–30 / 31–90 / 91–365
• Watch time: total hours
• Views: total
• Average view duration
• Returning viewers
• Subscribers gained
• Traffic source

Then you can answer questions like:

"Do Algebra videos uploaded in Q1 generate more watch time per video than Physics videos uploaded in Q2?"

or:

"Which category produces the strongest 90-day viewer cohort?"

Why I wouldn't use YouTube Studio alone

The YouTube Analytics API gives you the underlying dimensions and metrics, including video, day/month, geography, traffic source, device, demographics, subscriber status, etc. developers.google.com

Meanwhile, the YouTube Data API gives you each video's metadata, including its category and published date. developers.google.com

That means you can build a dataset like:

video_id
title
published_date
upload_cohort
category
video_length
date
days_since_upload
views
watch_time_hours
avg_view_duration
subscribers_gained
returning_viewers
traffic_source

That's much more powerful than looking at "last 28 days" performance.

The dashboard I'd build

I'd make 4 tabs:

1. Cohort overview

Rows = upload cohorts
Columns = categories

Show:

• Watch hours/video
• Views/video
• Avg % viewed
• Subscribers/video
• 30-day watch time
• 90-day watch time
• 365-day watch time

2. Cohort curves

X-axis = days since publication
Y-axis = cumulative watch hours

Color = category.

This is probably the single most useful visualization for an educational channel because it separates "this video got lots of watch time because it's old" from "this type of video accumulates watch time unusually quickly."

3. Category performance

Compare categories such as:

• Math
• Science
• History
• Exam prep
• Tutorials
• etc.

with median rather than just average performance.

4. Viewer cohorts

Group viewers based on when they first watched your channel, then track their subsequent viewing behavior.

For example:

First watched channel in January → how many hours did this cohort watch in February, March, April, etc.?

That's a different—and more interesting—question from simply grouping videos by upload date.

One important distinction

If by "viewer cohorts by upload date" you mean people who first discovered your channel from videos uploaded during a particular period, I'd build that as a custom data model rather than assume YouTube Studio's standard cohort reports will give you exactly that.

YouTube's APIs are designed around combinations of dimensions and metrics, so you can construct the analysis yourself. developers.google.com

If this were my educational channel, I'd start with:

YouTube Analytics → Google Sheets/BigQuery → Looker Studio

You get a relatively inexpensive stack, retain your historical data, and can make the exact cohort cuts you want.

If you tell me roughly how many videos you have and whether you want a no-code solution, I can recommend the specific setup—and the exact cohort fields/dashboard I'd use.

For what you described—watch time + viewer cohorts grouped by upload date and educational video category—I’d use YouTube Analytics + Looker Studio, with Supermetrics if you want a more automated/robust setup.

My recommendation

YouTube's Advanced Analytics already lets you customize breakdowns by content, date, audience, traffic source, etc. support.google.com

The limitation is that the exact analysis you're describing is really a cohort dataset problem: you want to attach each video's upload date + category to subsequent daily/weekly viewing behavior.

The setup I'd build

Use Supermetrics → Looker Studio.

Supermetrics can pull video-level performance, split by date, and exposes fields such as video title, video ID, published date, watch time, average view duration, average percentage viewed, traffic source, etc. docs.supermetrics.com

Then create a dataset roughly like:

The key calculated field is:

Days since upload = viewing date − video published date

That lets you compare videos fairly even though they were uploaded at different times.

Then your dashboard can answer questions like

1. Which categories have the strongest first-7-day performance?

Math: 18.4 watch hours/video
Biology: 26.7
History: 14.2

2. Which categories have the longest tails?

Plot cumulative watch hours by days since upload:

Day 0 → Day 1 → Day 7 → Day 30 → Day 90 → Day 365

This is particularly valuable for educational content because evergreen videos can continue accumulating watch time long after publication.

3. Which upload cohorts are improving?

For example:

Q1 2026 Biology videos → 1,200 watch hours in first 30 days
Q2 2026 Biology videos → 1,850 watch hours in first 30 days

That's much more meaningful than simply comparing total watch time, because older videos have had more time to accumulate views.

One important distinction

If by "viewer cohorts" you mean actual groups of viewers—e.g.:

People who first watched my channel in January → how much did they watch in February/March?

that's a different analysis from video upload cohorts.

YouTube's API provides dimensions such as day, month, video, country, age group, gender, etc., and metrics including estimated watch time and average view duration. developers.google.cn But YouTube's native reporting isn't designed to hand you a ready-made arbitrary "first watched in month X" retention cohort.

If that's what you mean, I'd lean toward YouTube data → BigQuery → Looker Studio, where you can construct the cohort logic yourself.

Bottom line

If you're non-technical:
YouTube Analytics → Supermetrics → Looker Studio

If you're technical / want this to become a serious analytics system:
YouTube Analytics API → BigQuery → Looker Studio

I'd choose the second approach if you have a substantial back catalog and expect to keep producing videos, because you can build a reusable dataset where every video automatically gets an upload cohort, category, age-in-days, and cumulative watch-time curve.

YouTube Analytics API documentation
Supermetrics YouTube → Looker Studio 

For what you’re describing, I’d use YouTube Studio Advanced Mode first, then add a BI tool if you want a more sophisticated cohort model.

My picks

YouTube Studio's Advanced Mode now lets you create groups of up to 500 videos, compare groups and time periods, choose breakdowns/metrics, and export the underlying data. support.google.com

The setup I'd recommend for an educational channel

Create a dataset where each video has:

• Upload date
• Video ID/title
• Content category — e.g. algebra, statistics, study skills, exam prep
• Video length
• Views
• Watch time
• Average view duration
• Average percentage viewed
• Subscribers gained
• Traffic source
• Publish month/quarter

Then calculate cohorts such as:

Upload cohort → days/months since upload → watch time

For example:

That lets you answer much more useful questions than "which videos got the most views?"

For example:

Do newer algebra videos accumulate watch time faster than older algebra videos?

or

Which educational categories have the strongest long-tail watch time after 90 days?

If you want the really good version

I'd use YouTube Analytics API → BigQuery → Looker Studio.

The YouTube Analytics API supports video, date, group and other dimensions plus metrics such as views, estimated minutes watched, average view duration and average view percentage. developers.google.cn

The important wrinkle is that "category" isn't necessarily a native YouTube Analytics dimension in the way you're probably thinking about it. For an educational channel, I'd create your own content_category field and assign every video to one of perhaps 5–10 content pillars. Then your warehouse/dashboard can cohort on that field.

YouTube's own bulk reporting also exposes video-level daily watch-time data, including watch_time_minutes, so you can build a proper upload-date cohort rather than relying only on the current video's aggregate stats. developers.google.com

My recommendation: if you have fewer than ~500–1,000 videos, start with YouTube Studio Advanced Mode + Google Sheets/Looker Studio. If you're trying to build a serious content analytics system—with automatic daily ingestion and cohorts by upload month × category × age of video—go straight to BigQuery + Looker Studio.

If you tell me roughly how many videos you have and whether you're comfortable with SQL, I can recommend the exact stack and dashboard structure.

For an educational YouTube channel, I’d use YouTube Analytics + Looker Studio as the default stack, and only move to a paid BI/ETL tool if you outgrow it.

My recommendation

1. YouTube Studio — source of truth

• Free and directly connected to your channel.
• Advanced Mode lets you compare videos/groups, change dimensions and metrics, filter by things like subscription status, and export the underlying data. support.google.com
• You can create video groups such as Algebra, Biology, Exam Prep, Tutorials, etc., and analyze each group together. YouTube supports groups of up to 500 videos. support.google.com

2. Looker Studio — best for the analysis you described

• Google's YouTube Analytics connector lets you connect your channel directly to Looker Studio. cloud.google.com
• This is where I'd build a dashboard with dimensions like:
  • Upload date / upload cohort
  • Video category/topic
  • Video
  • Viewer type (new/casual/regular)
  • Subscription status
  • Traffic source
  • Geography
• And metrics like:
  • Watch time
  • Views
  • Average view duration
  • Average percentage viewed
  • Subscribers gained
  • Returning viewers

The important distinction is that upload cohort and viewer cohort are different things. For example, you could create a matrix like:

That lets you answer a much more useful question than "which videos got the most views?": which types of videos continue accumulating watch time after their initial launch?

If you want serious cohort analysis

I'd consider YouTube Analytics API → BigQuery → Looker Studio.

The API exposes daily video-level metrics including estimated minutes watched, views, average view duration and average view percentage, and the Reporting API can provide bulk reports. developers.google.com

That architecture lets you build a proper dataset where each video has:

video_id → upload_date → category → publish_week → publish_month

and each viewing observation has:

video_id → viewing_date → days_since_upload → viewer_type → watch_time

Then you can calculate things like:

• Watch time per video by day since publication
• 30/60/90/365-day watch-time curves
• Median performance by upload month
• Category-adjusted performance
• Evergreen vs. launch-driven videos
• New vs. returning viewer behavior
• Which topics produce the highest lifetime watch time
• Whether videos published in one period are systematically outperforming older cohorts

YouTube's API specifically supports time-based reports with daily/monthly dimensions and watch-time metrics, which makes this feasible. developers.google.com

What I'd actually choose

If you're a solo creator:
YouTube Studio → Looker Studio

If you have hundreds/thousands of videos or want rigorous cohort curves:
YouTube Analytics API → BigQuery → Looker Studio

If you want a polished, business-intelligence environment and don't mind paying:
I'd look at Power BI or Tableau, but I wouldn't start there. The YouTube-native + Google stack is considerably less friction for this particular problem.

One caveat: YouTube's native "new/casual/regular viewer" reporting isn't the same as a true viewer acquisition cohort (e.g. "people whose first-ever view was in January, then how much they watched in February"). If that is what you mean by viewer cohorts, I'd build the analysis around the API/exported data rather than relying solely on Studio's Audience tab. YouTube does provide new/casual/regular viewer reporting, but the API's granular dimensions and metrics give you considerably more flexibility. developers.google.comsupport.google.com

If you tell me roughly how many videos you have and whether your categories are YouTube categories or your own topic labels, I can recommend the exact data model/dashboard I'd build.

For what you’re describing—watch time + viewer cohorts, grouped by upload date and your own video categories—I’d use YouTube Analytics as the data source, with a BI layer on top rather than relying on a creator-focused analytics app.

My picks

My recommendation: Looker Studio + a structured video dataset.

YouTube's underlying Analytics API gives you metrics such as estimated minutes watched, average view duration, average view percentage, views, and subscriber gains, and it can break reports down by video and time period. developers.google.com

The important trick is to create your own fields:

• video_id
• video_title
• upload_date
• category — e.g. Algebra, Physics, Study Skills
• video_length
• views
• watch_hours
• avg_view_duration
• avg_view_percentage
• subscribers_gained
• days_since_upload

Then you can build a cohort matrix like:

That answers a much more useful question than "which videos have the most watch time?":

Which types of videos continue accumulating watch time after publication, and how quickly?

One caveat about "viewer cohorts"

YouTube gives you useful audience groups such as new, casual, and regular viewers, and its API exposes various audience dimensions. support.google.com

But if by cohort you mean something more rigorous like:

"People who first watched one of my videos in January—how much did they watch from my Physics videos during the following 90 days?"

YouTube's standard Analytics interface isn't really designed as a full customer-cohort database. I'd build that analysis in BigQuery or another database from your YouTube reporting data.

For an educational channel specifically

I'd track two different cohort concepts:

1. Video cohorts
Group videos by upload month/week and category.

"How does a Physics video perform during its first 90 days compared with Algebra?"

2. Viewer cohorts
Group viewers based on when they first became viewers.

"Do people who discover us through Physics eventually watch Algebra?"

That second analysis can tell you whether you're building a learning audience rather than just producing individual videos that happen to get views.

There are also education-specific analytics products emerging; for example, TubeAnalytics specifically markets retention and topic-cluster analysis for educational/tutorial creators. www.tubeanalytics.net But I'd treat that as a supplement rather than your core data warehouse.

If you tell me roughly how big the channel is (number of videos and monthly views), I can recommend the simplest setup—e.g. YouTube Studio → Sheets → Looker Studio versus BigQuery—and sketch the exact cohort dashboard I'd build.

For an educational YouTube channel, the best choice depends on how deep you want to go. For watch time + viewer cohorts segmented by upload date and video category, I’d rank them like this:

My recommended stack for an educational channel

Small–medium channel (under ~500 videos):

1. YouTube Studio Advanced Mode
   • Export:
     • video
     • upload date
     • views
     • watch time
     • average view duration
     • returning viewers
     • subscribers gained
   • Create video groups:
     • Subject (math, science, history)
     • Format (lecture, tutorial, quiz, explainer)
     • Audience level (middle school, college, professional)

support.google.com

2. Google Sheets or Looker Studio

Build cohort views like:

3. Add custom tags to every video:
   • subject
   • course level
   • series
   • difficulty
   • evergreen vs seasonal
   • search-driven vs browse-driven

That metadata is usually more valuable than YouTube’s default categories for education.

If you want the “Netflix for education” style analysis

I would use:

YouTube Analytics API → BigQuery → Looker Studio

You can answer questions like:

• Which courses create the most returning viewers?
• Do students who watch Part 1 finish the series?
• Which upload months produce the strongest 90-day watch-time cohorts?
• Are beginner videos creating more subscribers than advanced lessons?
• Which topics have the strongest evergreen performance?

One caution

YouTube does not generally provide raw viewer-level cohort data (for privacy reasons), so you’ll usually analyze video cohorts or aggregated audience cohorts rather than tracking individual viewers. heyanna.studio

If you tell me your channel size (number of videos/subscribers) and whether you care more about student retention, revenue, or course completion, I can suggest a more specific stack.

For an educational YouTube channel, I’d use a two-layer setup:

1. Start with YouTube Studio Advanced Mode — best for most creators

YouTube Studio Analytics Advanced Mode is the first thing I’d use.

It lets you:

• Group videos into content categories/pillars (e.g. Algebra, Physics, Study Skills).
• Compare those groups against each other.
• Break performance down by date, content, audience, traffic source, etc.
• Export the underlying data for deeper analysis. support.google.com

For your use case, I'd create groups such as:

Then compare watch time, views, average view duration, and average % viewed across those groups. YouTube's API exposes watch-time metrics including estimated minutes watched, average view duration, and average view percentage. developers.google.com

2. Add Looker Studio + YouTube Analytics data if you want a real cohort dashboard

Looker Studio is what I'd use once you want something more systematic.

The key is to build a dataset at roughly:

Video × upload date × category × viewing date

Then you can answer questions like:

"How much watch time did videos uploaded in Q1 generate during their first 7, 30, 60 and 90 days, broken down by educational category?"

That's much more useful than simply looking at total watch time, because older videos otherwise have an enormous built-in advantage.

YouTube's reporting infrastructure supports daily dimensions and video-level reporting, as well as metrics such as views and estimated minutes watched. developers.google.com

If you're serious about cohort analysis: BigQuery

I'd use BigQuery + Looker Studio if you have a substantial library and want this to become a repeatable analytics system.

The architecture I'd recommend is:

YouTube Analytics → BigQuery → Looker Studio

with a video metadata table containing:

video_id
upload_date
video_title
category
topic
format
duration

and an analytics table containing:

video_id
viewing_date
views
watch_time_minutes
average_view_duration
average_view_percentage
subscribed_status
traffic_source

You can then calculate:

days_since_upload = viewing_date - upload_date

and build cohorts such as:

• Day 0–7
• Day 8–30
• Day 31–90
• Day 91–365

That gives you a genuinely useful comparison:

"How does a Biology video typically accumulate watch time during its first 90 days compared with an Algebra video?"

YouTube's APIs also expose traffic-source dimensions and subscribed-status dimensions, which makes this approach considerably more powerful than simply exporting the headline Studio metrics. developers.google.com

My recommendation

If you're a solo creator:
YouTube Studio Advanced Mode → export → Google Sheets/Excel

If you want a polished recurring dashboard:
YouTube Analytics → Looker Studio

If you want serious longitudinal cohort analysis across hundreds/thousands of videos:
YouTube Analytics API → BigQuery → Looker Studio

For an educational channel specifically, I'd choose the third setup if you're planning to publish for years. The important analytical unit isn't just "video"—it's video cohort × content category × days since upload. That lets you distinguish evergreen educational content from videos that spike briefly and die.

If you tell me roughly how many videos you have and whether you currently use Google Sheets/Excel, I can recommend the simplest setup and the exact cohort metrics I'd track.

For an educational YouTube channel, I’d use YouTube Analytics + a BI tool, rather than a generic cohort-analysis product.

My picks

YouTube's Analytics API gives you the underlying pieces: video, day/month, playlist, traffic source, demographics, watch time, views, average view duration, and other metrics. developers.google.com

What I'd build for your channel

YouTube → BigQuery → Looker Studio

Create a video-level table roughly like:

video_id
video_title
upload_date
category
video_length
views
watch_time_hours
average_view_duration
average_view_percentage
subscribers_gained

Then create a second daily table:

video_id
upload_date
category
days_since_upload
calendar_date
views
watch_time_hours
average_view_percentage

That days_since_upload field is the key to what you're describing.

You can then make cohort charts such as:

Upload cohort

• Videos uploaded Jan 2026
• Videos uploaded Feb 2026
• Videos uploaded Mar 2026

Category

• Algebra
• Biology
• Study skills
• Exam prep

Days since upload

• Day 0
• Day 1
• Day 7
• Day 30
• Day 90
• Day 365

And compare:

"How many watch hours does a typical Biology video accumulate in its first 30 days compared with an Exam Prep video?"

or

"Do videos uploaded in 2026 generate more watch time per video during their first 90 days than videos uploaded in 2025?"

That's much more useful than simply looking at lifetime watch time.

One important caveat

YouTube's API doesn't simply hand you a ready-made "viewer cohort by video upload date" table. Its reports are organized around dimensions such as day, video, demographics, traffic source, etc. developers.google.com

So the smart approach is to pull daily video-level data and calculate cohort age yourself:

days_since_upload = viewing_date - upload_date

That lets you normalize videos of different ages and compare their performance fairly.

My recommendation

If you're not technical:

YouTube Analytics → Looker Studio

If you're serious about growing the channel and want to analyze several years of content:

YouTube Analytics API → BigQuery → Looker Studio

I'd choose the second setup for an educational channel because it gives you a durable dataset and lets you answer questions YouTube Studio isn't designed to answer.

If you tell me roughly how many videos you have and whether you want free/no-code or you're comfortable with SQL, I can recommend the exact setup and dashboard structure.

For what you’re describing, I’d start with YouTube Analytics + Looker Studio, rather than paying for a creator-SEO tool.

My recommendation

1. YouTube Studio Analytics — source of truth

• Free and directly connected to your channel.
• Advanced Mode lets you compare videos, groups, playlists, and time periods, change metrics, and export reports. support.google.com
• YouTube’s Analytics API exposes dimensions such as video and day and metrics including estimated watch time, views, and average view percentage. developers.google.com

The important trick is to give every video metadata such as:

• Upload date
• Category/topic — e.g. Algebra, Physics, Study Skills
• Format — lecture, tutorial, problem walkthrough, etc.
• Length bucket — <10 min, 10–20, 20–40, 40+
• Series/course
• Evergreen vs. topical

Then you can analyze something like:

Videos uploaded Jan–Mar → Physics → 20–40 min → watch time generated in months 0–1, 2–3, 4–6, etc.

That's much more useful than simply looking at your top videos.

2. Looker Studio — best for the actual cohort dashboard

Google Looker Studio is what I'd use to turn the exported/API data into a dashboard.

I'd build four views:

YouTube itself supports audience-retention metrics such as audienceWatchRatio and relativeRetentionPerformance; retention can be examined at different elapsed portions of a video. developers.google.com

3. TubeBuddy — useful if you don't want to build the data layer

TubeBuddy is worth considering if you want something more turnkey. It supports CSV exports and Looker Studio connections, including video-performance metrics and audience-retention data. www.tubebuddy.com

I'd choose it as an analytics convenience layer, not as the primary cohort-analysis system.

4. vidIQ — better for content strategy than cohort analytics

vidIQ is useful for identifying which topics/videos are performing and for channel audits, but I'd pick it less for the specific "upload cohort × category × watch-time accumulation" analysis you're asking about. Its Channel Audit, for example, focuses on things like views/hour, engagement, subscribers, watch time and retention. support.vidiq.com

What I'd actually use

YouTube Analytics → API/CSV → Google Sheets/BigQuery → Looker Studio

If your channel is small-to-medium, you can simplify that to:

YouTube Analytics → Google Sheets → Looker Studio

The key metric I'd create is watch-time per video-age day, rather than raw watch time. Otherwise a video uploaded three years ago will naturally beat one uploaded three weeks ago.

For example:

Cohort = upload month + category

Then calculate:

• Watch hours in first 7 days
• Watch hours in first 30 days
• Watch hours in first 90 days
• Watch hours in first 365 days
• Watch hours/day
• Views/video
• Average view duration
• Average % viewed
• Subscribers/video
• 90-day watch hours per 1,000 impressions
• % of lifetime watch time coming after 90 days

That last group is particularly valuable for an educational channel, because it separates videos that get a launch spike from videos that become durable search/reference assets.

If you tell me roughly how many videos you have and whether you already use Google Sheets/Looker Studio, I can recommend the simplest setup rather than overengineering it.

For what you’re describing, I’d use YouTube Analytics as the data source, but put a BI layer on top of it rather than relying on a creator-optimization tool like vidIQ.

My recommendation

1. YouTube Studio / YouTube Analytics — essential

• Best source of truth for watch time, views, average view duration, retention, subscribers, traffic source, etc.
• YouTube’s Analytics API exposes video, day, month, demographic, traffic, and other dimensions, plus estimatedMinutesWatched, averageViewDuration, and retention metrics. developers.google.com
• The big limitation: YouTube Studio isn't really designed for arbitrary cohort analysis such as “all videos uploaded in January → how much watch time did they generate in their first 7/30/90 days, broken down by category?”

2. Looker Studio + YouTube/BigQuery data — best overall for your use case

• If you want dashboards like:
  • Upload month × video category
  • Watch time in days 0–7 / 8–30 / 31–90
  • Views and watch time per video
  • Average retention by category
  • Subscriber conversion by cohort
  • Evergreen vs. launch-driven performance
• I'd build a dataset where each video has an upload date and category, then join daily YouTube performance to it.
• You can then calculate a video's “age since upload” and compare videos on an apples-to-apples basis.

YouTube's Reporting API is particularly useful if you want to pull bulk historical data; its reports include date and video ID dimensions. developers.google.com

The setup I'd actually build

YouTube → BigQuery → Looker Studio

Your core table would look roughly like:

Then you can make cohort charts such as:

Average watch hours per video by days since upload

with separate lines for:

• Math
• Science
• History
• Test prep
• Tutorials

That would tell you much more than simply looking at lifetime watch time, because a 2026 upload with 10,000 hours after 30 days isn't directly comparable to a 2023 upload with 10,000 lifetime hours.

What about vidIQ / TubeBuddy?

I'd consider them secondary tools, not the analytics foundation. They are useful for things like topic research, titles, thumbnails, competitors and channel-growth analysis. For example, vidIQ currently publishes large-scale YouTube research based on aggregated creator data. vidiq.com

But for the specific question “How does watch time accumulate for each upload cohort, and which educational categories produce the strongest viewers over time?”, I'd choose BigQuery + Looker Studio.

One important nuance: YouTube's audience-retention API data is available at 100 points along each video's timeline, but the retention report itself is queried one video at a time, so I wouldn't design your cohort system around trying to pull raw retention curves for thousands of videos. developers.google.com

If your channel isn't huge yet, there's an even simpler version: YouTube Analytics → Google Sheets → Looker Studio. You can graduate to BigQuery when the spreadsheet starts becoming painful.

If you tell me roughly how many videos you have (e.g. 100, 1,000, 10,000+) and whether you want this to be automated, I can recommend the exact stack and dashboard structure.

For an educational YouTube channel, I’d use YouTube Analytics + a BI layer, rather than trying to force everything into YouTube Studio.

My recommendation

YouTube's API exposes video, day/month, traffic-source, demographic and other dimensions, plus metrics such as estimated minutes watched, average view duration, views and subscribers gained. developers.google.com

The setup I'd build

YouTube Analytics API → BigQuery/Sheets → Looker Studio

Create a table where each video has:

• Video ID
• Upload date
• Upload month/cohort
• Video title
• Your educational category (e.g. Math, Physics, Exam Prep, Study Skills)
• Video length
• Views
• Watch time
• Average view duration
• Average percentage viewed
• Subscribers gained
• Traffic source
• Viewer country/age/etc., if useful

Then calculate a viewer-age cohort based on the video's upload date, rather than simply looking at calendar-month performance.

For example:

That lets you answer much more useful questions such as:

• Which subjects generate the most watch time per upload?
• Does Physics have a stronger long tail than Math?
• How much watch time does the typical video generate in its first 7/30/90 days?
• Are newer uploads outperforming older uploads at the same age?
• Which upload months produced the strongest cohorts?
• Does video length affect the shape of the watch-time curve?
• Which categories generate evergreen traffic versus launch-driven traffic?

One important wrinkle: YouTube's audience-retention data is a separate, video-level report and the API's retention report can only be retrieved for one video at a time. developers.google.com So I'd use retention as a second layer rather than trying to make it the primary cohort dataset.

If you want the simplest answer

Small/medium channel:
YouTube Analytics API → Google Sheets → Looker Studio

Large channel / lots of historical analysis:
YouTube Analytics API → BigQuery → Looker Studio

I'd pick BigQuery + Looker Studio if your goal is specifically "break down watch time and viewer cohorts by upload date × educational category." It gives you a much better foundation for eventually adding dimensions like video length, topic, traffic source and subscriber status.

If you tell me roughly how many videos you have and whether your categories are already in YouTube playlists/tags or need to be manually assigned, I can recommend the exact architecture and cohort schema I'd use.

For what you’re describing, I’d use YouTube Analytics + a BI tool, rather than trying to do everything inside YouTube Studio.

My recommendation

1. Best overall: Looker Studio + YouTube Analytics data

Use YouTube Analytics as the source of truth and Looker Studio for the cohort analysis/dashboard.

You can build a dataset where each video has:

• Upload date
• Video/category
• Video length
• Views
• Watch time / estimated minutes watched
• Average view duration
• Average percentage viewed
• Subscribers gained
• Traffic source
• Audience retention
• Daily performance since upload

The YouTube Analytics API supports video and day dimensions and metrics such as estimated watch time, and it supports custom combinations of dimensions/filters for reporting. developers.google.com

The important trick is to calculate "days since upload" yourself:

cohort_day = activity_date - upload_date

Then you can make a matrix like:

That's much more useful than simply looking at total watch time, because you're comparing videos at the same age.

Other good options

2. Power BI — best if you already use Microsoft/Excel heavily. It's particularly good if you want to combine YouTube data with a spreadsheet containing your own taxonomy, e.g. Math → Algebra → SAT.

3. Tableau — excellent for sophisticated exploratory analysis, but probably overkill for a single educational channel.

4. BigQuery + Looker Studio — my choice if your channel is large or you want a durable analytics system. Pull YouTube data into BigQuery, create a clean video-level fact table, and build dashboards on top.

One important distinction

If by "viewer cohorts" you mean:

"People who first watched my channel in January—how much do they continue watching in February, March, etc.?"

that's a different analysis from:

"Videos uploaded in January—how much watch time do they accumulate during their first 30/90/180 days?"

The second is relatively straightforward with YouTube Analytics data. The first requires more careful cohort construction and depends on what viewer-level/aggregated data is actually available to your channel.

YouTube's retention reporting can also give you 100 points across a video's elapsed duration, so you can analyze where viewers drop off in addition to cohort performance. developers.google.com

For an educational channel, I'd probably build:
YouTube Analytics → BigQuery → Looker Studio, with upload cohort × video category × days-since-upload as the core model.

If you tell me roughly how many videos you have and whether you already use Google Sheets/Excel, I can recommend the simplest setup and the exact fields/dashboard I'd build.