Streaming Services in 2023

Blair Liikala
9 min readFeb 15, 2023

--

Today much of the streaming video technology stack is a commodity, and lots of options exist, but finding what you need can be time consuming.

Individual components of a video system can be provided by one service or many services. It all depends on what part of the technology stack you want to build, and what can be outsourced.

The problem with lots of options is trying to compare their differences and choose one. The Paradox of Choice. The wrong choice with video services can be an expensive mistake, and not all services are equal.

For over a decade running a media department that ran a streaming service, I continually tested and evaluated services to see if and when it made sense to buy into one. My goal was to find a service that integrated with our existing workflows, had good quality, and took responsibilities away I could delegate (read more). So I created product categories to help make decisions on the solutions I needed.

Video Technologies

Video files are gigantic and comes in many file types. Getting this to stream on a public highway called the Internet is tricky. To do this, video is encoded, a process that removes redundant or non-important visual data, then sliced into chunks a few seconds long and put on servers close to your house in order to load fast. The “stack” is all the software that makes this happen.

That’s the quick version. For an awesome overview of streaming, check out howvideoworks

A list of technologies may first look like this:

  • Deliverables:
    Storage
    for larger original and smaller encoded video files.
    Encoding to smaller files and in standard formats.
    Packaging of encoded files into streaming formats.
  • Distribution of video across the Internet.
  • Playback of video on all devices.
  • Management of your video library like adding, deleting, time-to-live, metadata…
  • Enrichment with captioning or tagging.
  • Integration into your content.

However if we plot services by their primary product we get a more relatable stack of technologies for choosing providers. From the fundamental video encoding libraries, to a complete service like Netflix. This should make it a little easier to reduce focus from dozen of companies to a group of 5–6 services that can fill your needs.

The question then is: what part of the stack to do you need?

Categories plotted with a handful of companies

Some companies focus on one category, while others span multiple. The darker purple shows (what I think) is the primary focus area of that company.

Disclaimers

To limit the scope here, these focus on more traditional live streaming and on-demand delivery, and less around real-time (apps like Zoom and Teams). Video analytics is also an important feature, but not considered. Cloud services from AWS, Google and Azure are minimally covered since they less specific to video, but are heavily involved in it. The list of examples is also likely to lean towards US-based companies, though there are some great international services out there.

Treat this as a general guide, not hard and fast rules.

Lets go over each lane:

Video Encoding

This is the product, the video itself, and is software that touches the video or audio directly such as encoding libraries, command line apps that start encoding, and local apps that can be run on a desktop.

This starts with encoding libraries like the most used video codec h.264, AVC (libx264), or a command line apps such as FFmpeg that manages converting video using libraries, to point-and-click apps like Handbrake. In this category, processing videos is manual, customized, and likely local. Costs are almost completely up-front in hardware, development time, and expertise.

Boxing Deliverables

What is it:

Services that provide raw resources or hardware for encoding and software to packaging and delivery of streaming files using the apps mentioned in the previous category. These apps and services encode, store, or package video for delivery.

This category also includes cloud services such as AWS Elemental/MediaLive or the Wowza Streaming Engine app running in the cloud. Other examples include Coconut or Qencode that are cloud-based, but still allow the customer to set the encoding settings. Increasingly there are more hardware encoding companies beginning to offer their product in the cloud.

Who is it for:
Streaming video engineers building for a client.
More ideal when there is more control over the source material when quality or speed are a priority.

What problem does it solve:
Needing to maintain servers, operating systems, compatibility, and updates. Better for scaling-up services.

What don’t you get:
Complete control over costs related to hardware, or core features with packaging. It requires more knowledge on how to best encode video.

Pricing:
Billing is usually more technical like time spent encoding, or the size of the files transferred. Costs increase with more complex codecs like AV1, bigger frame sizes like 4k and higher frame rates.

Examples

Video (APIs) as a Service

These services handle the complexity of video engineering, and turn videos into IDs that just work by straightforward API calls. They are great for developers who need to work with video, but are not video engineers. A developer does not have to stay up on the latest codecs, and encoding settings or worry about load-balancing an encoding pipeline like the previous category. This lane also introduces data analytics and open-source tools to make building with video easier, such as players.

What is it:
Video as a service; Storage, encoding and delivery as one.

Who is it for:
General developers building any kind of app that needs video.

What problem does it solve:
Puts the complexity of video encoding, storage and delivery into simple API calls.

What you don’t get:
Little to no control over quality, delivery, and cost control. No control over adding video-related features such as multichannel audio, automatic subtitles, and content recognition directly into the video platform (though some could be built outside the platform).

Pricing:
Based on usage, usually minutes encoded, stored and streamed.

Examples

Video Content Management System (CMS)

This lane starts to look like what most developers expect when thinking about streaming. The goal is to make video management as easily and straightforward as possible, usually through a web portal. Need to upload a few videos and post the links on a website? This is your lane.

What is it:
A service where managing all video content is done through their web portal. Uploading, metadata such as title and descriptions, and analytics. Also publishing controls such as scheduling when a video is public. Not to be confused with a normal content management system that handles all content like blog posts, contact us page, user comments…etc.

Who is it for:
Content owners and creators who focus on premium video content itself and running their business with a web development team.

What problem does it solve:
Easy point-and-click interfaces for managing video. One source for videos, captions, and metadata.

What you don’t get:
Flexibility to build customized publishing workflows, and usually quality integrations into other systems. For example, creating a unique approval process, or triggering emails when updated. Also while services in this category have an API, they tend to be treated like afterthoughts. Building video management into another content management system becomes clunky.

Pricing
Typically fixed pricing monthly or annually with a set amount of usage.

A few Examples

Destination or OTT

Summary:
Get a complete solution to handle video management, subscriptions, payments, and have a good looking website ready to go. The additional value here is a destination website for viewers to visit.

What is it:
“Netflix” as a service. A complete solution.

Who is it for:
Content creators and owners, broadcasters, and operators who have a budget, but lack or do not want a web development team. Providing both the CMS and customer website as a complete solutions, channels to smartTVs, or streaming to socials.

What problem does it solve:
Meeting fast time to market deadlines and reduces most technical complexity.

What you don’t get:
Usually higher costs and cookie-cutter websites. Little to no customizations in layout and design beyond templates. Lack the ability to incorporate more value-adding services unless the platform builds it.

Pricing
Typically fixed pricing annually with a set amount of usage. Usually a contract per-customer. Some premium features may be locked in higher tiers.

Examples:
Vimeo OTT used for sites like Detroit Symphony and Dallas Symphony

Value Adders

These are services that solve a specific workflow headache. More generally, they add value to the video. This could include using AI to process the video in some way, or making a multi-step process into fewer steps. Many of these solutions use companies from the API category to host their video.

There are a lot of companies in this category, even more when we talk about AI. Even though these tend to not be thought of a core parts of a video stack, I believe with growing user feature expectations they need to be covered.

I would also include advertisement systems in this category since they (quite literally) add value to the primary video. Though because ad insertion modifies the video either server-side or client-side, ads could fit into packaging.

What is it:
Specific services centered around video, usually designed to solve a problem such as casting to social networks, editing video, easy capturing for reactions…etc.

Who is it for:
Content creators, usually smaller or single producers.

What problem does it solve:
Improving a specific workflow. For example, creating a simple live stream with graphics through a browser instead of needing to buy graphic hardware, or automatically generating chapters from a podcast instead of manually entering timecode.

What you don’t get
Reliability. Many in this category are startups, and risk being reliable in the long-term.
All-in-one solution. Being a-la-cart means possibly stringing several services together.

Pricing
Can vary depending on the solution.

A “Few” Examples

Streaming Studios

Netflix, Hulu, Disney…etc. These are what your parents think of when we talk about streaming services. If you made a movie and want to stream it, find a partner in this category. The difference here is exclusivity. I can not sign up and create my own Netflix site like I could with an OTT or CMS platform.

Many in this category likely use services mentioned here, especially in the API grouping. Others like Disney or Netflix with more expertise have developed more of their technology stack further up the list.

Other lesser-known examples that fully own their streaming product:

Socials; Twitter, Facebook, Twitch…

These are somewhere between OTT and streaming services, but since the video is a small part of a larger social networking experience I will omit them from this analysis.

There are some companies I did not include, but the idea is providing a framework for narrowing down your search for a video provider.

Commentary

The diversity of services around the technology stack is amazing. The diversity of video-enriching services is equally amazing. So much technology is at the click of a signup button. So much more accessible than it historically has been since the start of TV.

However it can be time-consuming to choose, and the business changes rapidly.

Being on the production end where quality is painstakingly preserved, I have always felt that the more you hand over to services, the more the encoded video quality decreases, until we get to streaming studios. It becomes a one size fits all model. A possible explanation with this quote, a summary of another quote:

Customers that can’t assess the quality of products pay the wrong amount for them, creating a disincentive for high-quality products to emerge and working against their success when they do.

If image quality is important, learn more about the tools that throw out more than 90% of your video data. Lean towards core tools. However if encoding speed, time-to-market or lower technical debt are priorities then using more managed services may be more ideal.

Next

What are the differences between companies and how do you pick one? I’ll save that for the next post, but in the mean time if you need help in this space I am available.

--

--

Blair Liikala

Solutions architect, previously manage a music recording department. Audio, video gear, web streaming and web development.