When CDN provider Edgio filed for Chapter 11, more than 935 customers relied on it for video streaming and web security. Akamai acquired selected assets, including certain customer contracts, but not Edgio’s staff, technology or network, and the service shut down in January 2025. Contracts can be transferred. Knowledge of how your system actually runs can’t.

Most comparisons of OTT managed services with an in-house team stop at the monthly fee. Two other numbers decide whether the model pays off: what it costs to move responsibility to a partner safely, and how much of your platform genuinely needs 24/7 cover. This guide sets out a TCO model you can adapt and covers both.

Maintenance, managed services and staff augmentation are different models

These three terms get used interchangeably, and that’s where many bad contracts start. The difference is who directs the work and who is accountable for the outcome.

Model Who directs the work Who is accountable for the outcome Typical scope
Staff augmentation You You Engineers join your team and work on your backlog
Maintenance Partner, within agreed tasks Partner, for defined fixes and updates Bug fixing, OS and SDK updates, change requests on existing systems
Managed services Partner, against agreed service levels Partner, for the service meeting its SLA Infrastructure, monitoring, integrations, maintenance, support and ongoing development

OTT managed services cover far more than keeping apps alive. A managed service provider can run cloud infrastructure, content delivery, monitoring, third-party integrations, incident response and new feature development, as well as the apps on smart TV, mobile devices and web. In every model, you keep full control of product ownership: the roadmap, video content and its rights, vendor choices and subscriber data.

Why running OTT platforms keeps getting more expensive

Two forces push up the price of ongoing maintenance, regardless of who does the work.

Platforms keep changing the rules

In April 2026, Roku updated its certification criteria. From 1 October 2026, US apps averaging more than 5 million streaming hours a month must support Instant Resume, and apps outside the US above 1 million hours a month must take part in Continue Watching. Roku also removed the exemption that let live streams and replays, including live linear channels and live sports events, skip deep linking. TVOD, live linear and made-for-kids apps are excluded from the Continue Watching requirement. None of this is a feature users asked for. It’s the price of staying in the store, and our breakdown of what can block a Roku app certification shows how often it catches teams out.

There are more platforms to maintain

Omdia expects VIDAA, Titan OS and TiVo to grow from 21% of the European smart TV market in 2025 to 28% by 2030, with VIDAA overtaking LG’s webOS in 2026. For streaming services, each new operating system is a long-term maintenance commitment, not a one-off launch.

Vendor risk sits on top. The Edgio case shows what happens when a supplier disappears: you keep your contract, but nobody who understands your integration comes with it.

A TCO model: in-house vs managed services vs hybrid

The model below is illustrative, not a client case. Its job is to make every spending category visible before you compare totals.

Model assumptions

  • Like many OTT services, the model service runs on web, iOS, Android, Android TV, Apple TV, Samsung Tizen, LG webOS and Roku, and combines subscription video with ad-supported video.
  • Eight platforms sit on four codebases: a shared web codebase for browser, Tizen and webOS; native iOS and tvOS; native Android and Android TV; and a separate Roku app, because Roku uses its own BrightScript and SceneGraph stack.
  • Releases ship monthly, and live streaming of sport runs on most weekends.
  • Product ownership, content and subscriber data stay with the broadcaster in every scenario.

The second assumption matters. Capacity follows codebases and certification cycles, not the raw number of platforms, so you don’t need a separate specialist per platform. You do need someone who knows each codebase and each store’s rules well enough to ship under pressure.

Cost category In-house Managed services Hybrid
Engineering capacity Team sized to cover all four codebases, whatever the demand Service fee, often a fixed monthly base plus priced change requests Smaller core team plus a dedicated team at the partner for the layers handed over
Hiring and replacing specialists Recruitment time, onboarding and gaps while roles are open Carried by the managed service provider; check continuity terms for named people You hire for the core, the partner covers niche toolchains
Multi-platform app maintenance Effort per codebase for OS updates, SDK changes and store rules In scope; check which platforms and how many releases are covered Usually the first layer handed over
QA, device lab and certification Devices and their refresh, test tooling, certification cycles Shared device lab; recertification in scope or priced per release Partner runs the lab, you sign off releases
Monitoring and incident handling Tooling, on-call rota and allowances Set by SLA tier; 24/7 is priced higher than business hours Partner on first line, your team handles business escalation
Live event support Extra people on shift, often overtime Scoped add-on or part of a 24/7 tier Agreed against the event calendar
Onboarding, vendor management and change Low Transition project, knowledge transfer, your time managing the partner, exit cost Coordination between two teams

What the model shows

Spend on an internal team is mostly fixed: the team has to cover four codebases, QA and on-call even in a quiet month. A managed service spreads that coverage across the partner’s other clients, but adds transition and management effort, and a monthly retainer is largely fixed too. How flexible it really is depends on how change requests and event cover are priced, so don’t assume outsourcing turns a fixed budget line into a variable one.

The deciding factor is utilisation. If your engineers are busy year-round across every codebase, keeping them usually costs less. If the work is uneven and your coverage needs exceed your steady workload, shared capacity usually wins. Hybrid sits in between and is where many broadcasters end up.

When in-house still wins

Keep the work internal when:

  • Your app and its viewing experience are your main differentiator and change weekly, so the people building them need to sit next to the people deciding.
  • You have the scale to keep specialists busy all year.
  • Your platform set is narrow and stable, such as web and two mobile apps.
  • Regulation or policy requires direct operational control.

Managed services also can’t fix unclear product ownership. If nobody inside the company owns the roadmap, a partner just moves the confusion further from the decision-makers.

What managed services look like in practice: ITV

In 2022 we built an ESAM application for ITV that manages SCTE-35 markers at encoder level, normalising, correcting, replacing, removing and creating markers to protect broadcast signal integrity. We’ve maintained it since. Application maintenance, change requests and regular infrastructure maintenance run during business hours. Alongside that, ITV has 24/7 infrastructure support with direct telephone access to our DevOps engineers for urgent issues.

This is where much of the money in managed services sits. Round-the-clock cover is the most expensive tier in any SLA, and it’s easy to buy it for every component by default. ITV’s setup applies it where a failure would affect live broadcast and uses business-hours cover for the rest. Before comparing quotes, list your components and mark which ones genuinely need 24/7. The full ITV case study also covers moving metadata services off a legacy Oracle-based system run by a third-party supplier, which is the kind of exit most teams postpone too long.

How to take over an existing platform safely

Most managed services contracts start with a platform built during an earlier OTT app development project, often paid as a one-time fee and delivered by someone else. Before a new managed service provider takes responsibility, check five areas. The last two are easy to miss, because they sit in accounts and credentials rather than code.

Five checks before handover

  1. Code and builds. Can the partner build and release every app from source, without the outgoing team’s laptops? If a build only works on one machine, you’ve found your first risk.
  2. Documentation and runbooks. What exists, how current it is, and what lives only in people’s heads.
  3. Vendor dependencies and access. Every SDK, licence, API key and contract the platform relies on, with owners and renewal dates, plus cloud consoles and monitoring tools, ideally in your name rather than the outgoing OTT provider’s.
  4. Store accounts. Moving an app between developer accounts is a formal process with its own rules. On the App Store, the Account Holders of both accounts must initiate and accept the transfer, app data is removed from the original account once it completes, and push notifications stop working until the recipient creates a new APNs certificate in its own account. If your apps sit under the outgoing supplier’s account, plan the transfer as a project, not an afternoon task.
  5. DRM credentials and certificates. Apple only approves FairPlay Streaming production credentials when the Account Holder of the content owner’s or licensee’s own developer account requests them, and it doesn’t approve third-party accounts acting on their behalf. Confirm where your DRM credentials actually sit, then build an inventory of every certificate and key the platform depends on, with expiry dates and owners. A certificate nobody is tracking is one of the most avoidable causes of a playback outage.

Staged knowledge transfer

Run knowledge transfer in stages. The new team shadows the incumbent, then the incumbent shadows the new team, and only then does responsibility move. Keep the outgoing team available for a short overlap, freeze non-essential releases during the handover, and never cut over during a major live event or a product launch. If the move involves replacing the platform itself rather than just its maintainer, read what breaks in an OTT platform migration first.

SLA terms and how to measure maintenance quality

Many SLAs promise a fast response and very little else.

Three times to define

Separate these three, because users feel each one differently:

  • Response time: how quickly the partner acknowledges and starts work on an incident.
  • Time to restore: how quickly users can watch again, even through a workaround.
  • Time to resolve: how quickly the root cause is fixed for good.

Define all three by priority, with stricter targets during live event windows.

Metrics that show maintenance quality

Report a short set of metrics monthly: service availability, playback failure rate, video quality against agreed thresholds, mean time to restore (MTTR), and on-time delivery of releases and certifications.

Contract protections

Put four protections in the contract: code, build pipelines and store accounts owned by you; secure source code access or escrow; a named core team with notice before key people change; and an exit plan with a defined knowledge transfer period. Our checklist for choosing an OTT tech partner covers the rest of the evaluation.

Which parts should you keep, and which should you hand over?

That’s the real decision, and it rarely has a whole-platform answer. Keep internal what defines your product and changes weekly. Hand over what needs broad coverage, deep platform knowledge or round-the-clock cover but doesn’t differentiate you. For most OTT services, the usual candidates are multi-platform maintenance, certification, monitoring and infrastructure support. And before you compare any quotes, decide which components genuinely need 24/7 cover.

If you’d like to work through that split for your own platform, we can start with an assessment of what you run today and the capabilities you want to keep, a proposed division of responsibility and the cooperation model that fits your business goals. Any savings come after that analysis, not before. Talk to our experts about OTT managed services.

FAQ

Staff augmentation adds engineers to your team, and you direct their work and own the outcome. Maintenance covers defined fixes, updates and change requests on existing systems. OTT managed services go further: the partner is accountable for the service meeting agreed levels, and scope can include infrastructure, monitoring, integrations, support and ongoing development. Product ownership stays with you in all three.

Not automatically. An internal team usually costs less when it’s busy year-round across every codebase. A managed service tends to cost less when work is uneven and you need broad platform coverage, device testing and on-call support that a small team can’t sustain. Compare full total cost of ownership, including transition, vendor management and exit costs, rather than day rates.

Separate response time, time to restore service and time to resolve the root cause, each defined by incident priority with stricter targets during live events. Report monthly on service availability, playback failure rate, mean time to restore and on-time delivery of releases and certifications. Match service levels to criticality rather than buying 24/7 cover for everything.

It depends on what you own. When Edgio collapsed, Akamai acquired selected assets including customer contracts, but not the staff, technology or network, and customers had to move quickly. If your code, build pipelines and store accounts sit with a vendor, recovery is slow. Source code escrow, current runbooks and accounts in your own name shorten it considerably.

Start with an audit of code and builds, documentation, vendor dependencies, store accounts and DRM credentials. Then run knowledge transfer in stages: the new team shadows the incumbent, the incumbent shadows the new team, and responsibility moves last. Keep the outgoing team available for a short overlap, freeze non-essential releases and avoid cutting over during major live events.

arrow_circle_right Table of contents

Damian Maicher

Business Researcher

Related articles