2.5 million registered users. A small number of fully redundant servers. That is RTS Planeta (Radio Television of Serbia), a platform More Screens has operated continuously since 2018. As the user base grew, the server footprint did not follow it upward. The growth was absorbed by efficient software, workload optimisation and carefully designed redundancy instead.

At DigitAlb, the same principle produced a different result. We reduced the required Catch-Up storage capacity by roughly 50% with no reduction in the retention window and no content removed. The subscriber sees exactly the same service. The economics underneath it are very different.

Neither saving came from finding a cheaper vendor. Both came from changing how the platform works.

After years of building and operating video platforms for telecom operators, broadcasters and content owners, one thing has become clear to us: the biggest savings are rarely found in one place, and almost never in the line item everyone is staring at.

The real TCO is bigger than the cloud bill

When people discuss streaming costs, CDN and cloud infrastructure dominate the conversation. But the real Total Cost of Ownership is much broader. A typical service combines middleware, transcoding, DRM, CDN, analytics, subscriber management, applications and billing from several vendors. Each component may look reasonably priced on its own. The hidden cost appears between them: integrations, maintenance, support contracts, duplicate infrastructure, engineering resources and operational complexity.

Reducing TCO therefore requires looking at the complete architecture and operation of the service, rather than optimising individual components in isolation.

This is not an argument for running a cheap platform. Cost reduction that degrades playback, increases buffering or slows down feature delivery is not a saving, it is deferred churn. The goal is a platform where growth in audience does not automatically produce proportional growth in cost, while quality of experience holds or improves.

Cloud is not automatically the cheapest answer

Cloud provides flexibility, rapid deployment and scalability. That does not mean every video workload belongs there. A broadcaster serving a predictable national audience has very different economics from an OTT service reaching diaspora viewers across 30 countries. Continuously transcoding 100+ linear channels is different again.

At More Screens we therefore combine our Media Cloud, customer infrastructure, local CDN nodes, dedicated video-processing hardware and public cloud resources. Sometimes cloud wins. Sometimes dedicated infrastructure wins. Increasingly, hybrid wins.

The architecture should follow the workload, not the technology fashion.

CDN cost grows with success

For globally distributed audiences, global CDN infrastructure makes sense. But operators with tens or hundreds of thousands of viewers concentrated inside their own network or region have another option.

We use local and hybrid CDN architectures, combining local delivery with regional or global CDN capacity where it is actually needed. For national and diaspora oriented services, this can mean lower cost per stream, better traffic control and improved playback performance.

The useful measure here is not the CDN price on the contract. It is the CDN cost per delivered viewing hour.

Catch-Up TV exposes a hidden scaling problem

Catch-Up storage capacity is primarily driven by channels, bitrate and retention time. But once a platform reaches 50,000+ subscribers actively consuming Catch-Up TV, a second problem becomes decisive: concurrent read performance.

Thousands of viewers simultaneously request different programmes, channels and positions inside the archive. The conventional response is to add faster storage and more storage nodes.

We approached it differently. More Screens developed its own file read/write system, optimised for video workloads and high concurrency, together with on-the-fly encryption during delivery. Instead of solving the problem by adding hardware, we changed how SPECTAR+ uses the underlying storage.

The result is lower storage requirements, reduced I/O pressure, and the ability to keep existing even older storage infrastructure productive for much longer. At DigitAlb, that meant roughly half the Catch-Up storage capacity for the same service.

Some of the most valuable streaming innovation is completely invisible to the viewer.

Millions of users do not have to mean hundreds of servers

RTS Planeta demonstrates the same principle at platform level rather than storage level.

More than 2.5 million registered users, served by a core platform running on a small number of fully redundant servers, under continuous More Screens operation since 2018.

Instead of increasing the server footprint as the user base grew, the focus went into efficient software, workload optimisation and carefully designed redundancy. A platform needs the capacity and redundancy to absorb traffic peaks. It does not need excessive infrastructure sitting permanently idle to prove it.

Scale should not automatically mean infrastructure sprawl.

Specialised hardware can still beat generic compute

For large numbers of continuously running linear channels, encoding density, power consumption, rack space and cost per channel matter.

This is why we developed MT-LIVE, combining high-density transcoding, CDN origin and DRM packaging in a compact appliance. For the right workload, specialised processing offers better economics than running the same job indefinitely on generic cloud compute.

But we do not insist on our own hardware. If the customer already has suitable encoders, storage or networking infrastructure, we reuse it. If another transcoder makes more sense, we integrate it. If public cloud is more efficient, we use cloud.

Optimise the customer’s architecture, not the vendor’s footprint inside it.

Optimise the workload before buying more infrastructure

When a streaming service starts struggling with growth, the natural response is more storage, more CDN, more compute, more servers. Sometimes that is genuinely necessary.

But there are three questions worth asking first:

Before adding infrastructure: can the workload itself be optimised? Before replacing infrastructure: why did it become a bottleneck? Before moving something to cloud: does cloud actually improve its economics, measured rather than assumed?

SPECTAR+ consolidates content management, subscribers, entitlements, multi-DRM, CDN management, analytics and customer operations, while remaining open to external billing, CRM, advertising, CDN and other specialised technologies. The objective is deliberate integration without unnecessary fragmentation.

Optimisation does not stop at launch

Architecture is not a one-time decision. Audiences change. Concurrency changes. Libraries grow. CDN pricing changes. Cloud consumption changes.

That is why, for many customers, More Screens continues to operate and optimise the environment after launch: monitoring, releases, incidents, CDN, storage, infrastructure, capacity and ongoing development.

The real benefit comes from treating optimisation as a continuous process rather than a migration project performed every five years.

The KPI worth adding

Streaming technology is becoming mature. Most serious platforms can deliver video to a television, a phone or a browser. The more interesting question is increasingly how efficiently they do it and whether that answer still holds at three times the current audience.

Operators already track uptime, latency and picture quality. The measure we would add alongside them is cost per viewing hour: total infrastructure and operational cost divided by hours actually watched. It travels well across very different services, it resists the flattery of subscriber counts, and it moves when the architecture moves.

Cost per stream, cost per concurrent viewer, cost per channel and storage cost per hour are all useful views into the same question. Cost per viewing hour is the one that belongs on the business review, not just the engineering dashboard.

So: what architecture allows our video business to grow efficiently for the next five years?

Our answer, after years of building and operating these platforms, is increasingly clear. Open. Hybrid. Workload-aware. Continuously optimised. Because your audience can grow without your infrastructure cost growing at the same rate.

We will be at IBC 2026, Stand 1.F11. Bring us your cost per viewing hour and we will tell you where it is leaking.

More Screens — Technology. Cloud. Operations. https://morescreens.com