Views by quarter
Growth Strategy & Modeling
We model each channel's audience, topics and search demand, then set targets and a programming plan the data can defend.
YouTube growth, engineered from data.
MediaComms runs YouTube channels for media companies as a data science discipline. We model what the platform rewards, test every decision, and operate your catalogue to the numbers.
Model card · breakout classifierv7.2
Average % viewed
Breakout · 8.6%
Click-through rate
- Channels under active optimization
- 200+
- Views managed per month
- 500M+
- Content processed per month
- 30TB+
The model, in production.
Every channel we run feeds one analytics layer: forecasts, live experiments and anomaly alerts. Your team sees the same view we operate from.
MediaComms / PlatformPortfolio MC-DEMOAs of 22 Sept 2026--:--:-- UTC
Portfolio views · actual and 3-period forecast
562.6M+22.3%
- Views, 30D
- 562.6M+2.3%
- Watch hours, 30D
- 23.5M+2.4%
- Net subscribers, 30D
- +2.03M+2.5%
- Average CTR
- 6.5%−0.1 pt
- Forecast error (MAPE)
- 4.0%+0.4 pt
| ID | Channel | Variable tested | Status | ||
|---|---|---|---|---|---|
| EXP-2291 | Markets Explained | Thumbnail: face vs chart | +18.4% | 99% | Shipped |
| EXP-2288 | Field Science | Title: question vs claim | +11.2% | 97% | Shipped |
| EXP-2279 | Build Log | Length: 12 vs 22 min | +9.7% | 95% | Shipped |
| EXP-2270 | Studio Sessions | End screen: playlist vs video | +7.3% | 96% | Shipped |
| EXP-2285 | Kitchen Standard | Upload time: 07:00 vs 17:00 | +6.9% | 91% | Running |
| EXP-2276 | Nightly Headlines | Shorts cut-downs: 3 vs 6 | +4.1% | 84% | Running |
| EXP-2261 | Little Atlas | Chapters on vs off | +2.6% | 88% | Running |
| EXP-2266 | Motor Desk | Intro: cold open vs recap | -3.2% | 93% | Rejected |
What drives views · feature importance
- Click-through rate0.31
- Average % viewed0.27
- Upload cadence0.14
- Returning viewers0.11
- Topic demand0.09
- Metadata quality0.08
Anomaly & policy monitor
- Anomalies flagged, 30D
- 12
- Anomalies resolved
- 12
- Open copyright claims
- 3
- Policy strikes
- 0
YouTube is a ranking system. We treat it like one.
Most channels still run on instinct: upload, hope, repeat. The platform doesn't guess. It ranks every video on signals it measures in real time: who clicks, who stays, and who comes back.
MediaComms works as an embedded data and operations team. We turn your analytics into models, the models into decisions, and the decisions into daily execution across every channel.
- Measured, not guessed
- Every recommendation starts from your first-party analytics and a stated baseline.
- Tested before scaled
- Changes ship as experiments. What doesn't beat the baseline is rolled back.
- Yours, always
- Channels, revenue and data stay yours. We operate on delegated, revocable permissions.
Four disciplines, one data layer.
Take one or all four. Each feeds the same analytics layer, so every decision on your channels is measured.
See what’s included for Growth Strategy & Modeling
A/B outcome distribution
Content Operations
An embedded team that ships at volume. Packaging, metadata, scheduling and publishing run on documented procedures, and every change is tested.
RPM by hour and weekday
Monetization & Revenue
We find where revenue is lost: ad placement, audience mix and dormant catalogue. Every revenue line is modeled and reconciled.
Daily views, anomalies flagged
Rights & Compliance
Automated monitoring for claims, policy risk and anomalies, so a single strike never takes a catalogue offline.
From audit to operating model in four weeks.
Every engagement follows the same sequence. Nothing ships before it is agreed in writing and measured against a baseline.
01Audit
Week 1–2
Read-only access to your analytics. We baseline every channel: performance, risk and upside.
02Model & plan
Week 3
We build the channel models and a written plan: what to grow, restructure or retire, with a target for each.
03Onboarding
Week 4
Our team embeds with yours. Permissions, procedures and reporting lines are agreed in writing.
04Execution
Ongoing
Daily operations run as experiments. Every change is measured against a baseline before it is scaled.
05Readout
Monthly
A report for every channel and for the portfolio: what moved, why it moved, and what we test next.
Readouts, not screenshots.
Every month you receive a readout for each channel and for the portfolio: what moved, why it moved, and what we test next. We walk through it with your team line by line.
- Cadence
- Monthly readout and review call. Live platform access in between.
- Scope
- Performance against forecast, experiments, revenue, claims and anomalies.
- Format
- PDF for the record, CSV for your analysts. The same numbers in both.
Monthly readout
PORTFOLIO MC-DEMO · 14 CHANNELS
AUGUST 2026
Issued 03 Sep 2026
| Metric | Value | Context |
|---|---|---|
| Views | 38,412,906 | +11.8% vs Jul · forecast +9.5% |
| Watch hours | 1,904,233 | +9.4% vs Jul |
| Net subscribers | +212,480 | +6.1% vs Jul |
| Uploads published | 286 | Long-form 94 · Shorts 192 |
| Experiments concluded | 38 | 22 shipped · 16 rejected |
| Forecast accuracy | 95.9% | MAPE 4.1% |
| Estimated revenue | $186,420 | RPM $4.85 |
| Policy strikes | 0 | All channels in good standing |
Next tests
- 1.Title framing on Field Science (question vs number)
- 2.Shorts-to-long-form handoff on Kitchen Standard
- 3.Upload window shift for EU audience on Build Log
Sources: YouTube Analytics & Reporting APIs, AdSense.
Measured outcomes.
Client names stay confidential. What we can show is the scope, the method and the result.
Education publisher
14 channels · 6 languages
Modeled topic demand across 22 overlapping channels, consolidated them into 14, and repackaged a 4,000-video back catalogue.
+212%views in 9 months
Broadcast news network
9 channels · daily output
Deployed automated claim and anomaly monitoring that clears clips before publishing and resolves disputes within 48 hours.
0strikes in 18 months
Entertainment studio
31 channels · music & film
Segmented audiences by value, rebuilt programming around the highest-yield segments and activated dormant titles.
+64%RPM in two quarters
Questions, answered.
What data do you work from?
Your first-party YouTube Analytics and revenue data through the official APIs, plus platform-wide search and trend signals. Your data is never shared with or pooled across other clients.
Who owns the channels and the data?
You do, always. We work through delegated, revocable permissions. Channels, revenue, content rights and data never transfer to MediaComms.
What access do you need?
Read-only analytics access for the audit. For operations, a manager role scoped to the channels in the engagement. Every permission is listed in writing.
How are engagements structured?
As monthly retainers scoped to the size of the portfolio and the services involved. Scope and fees are agreed after the audit, before onboarding.
What does reporting look like?
A monthly readout for each channel and for the portfolio: performance against forecast, experiments concluded, revenue and risk. We walk through it with your team every month, with live figures in between.
Which verticals do you work in?
Entertainment, music, education, news, finance, kids and family, lifestyle and more. The method depends on scale and data, not on the niche.
Do you work with individual creators?
Our model is built for media companies and networks running multiple channels. Large creators with an established team are considered case by case.
Start with a channel audit.
Tell us about your channels. Within two business days our team will come back with the scope for a baseline audit.
Direct line