Buy vs. Build

BVBWhen creating new services and products, organizations always face a challenge whether to buy or build key underlying components and functionality. As WebRTC attracts an increasing degree of interest, we regularly hear from customers that they are considering the trade-off around the decision to buy or build. Many go so far as to try and build their own real time video or audio solution before they turn to a hosted platform like OpenTok. Not surprisingly given the business we are in, we come down pretty strongly on the side of leveraging a hosted service.

Based on the use cases we have seen and some of the questions we have answered, we wanted to share some of the common contemplations and drivers we are hearing from our customers  when making the buy vs build decision.

1. Your core business- Unless you are trying to build the next Skype (and yes, there are plenty of people  trying to do that), the delivery of video and audio is probably not your main business. Your expertise lies in running a teaching platform, or a banking website, or an online medical practice – and you simply want to add real time communications to your online and/or mobile experience. Every hour and dollar that you spend trying to figure out how to do voice and video is time and money not spent on your core business. It’s a much better investment of resources to leave that to the experts. In some ways voice seems seductively simple – plug in a click to call, or bolt on an SMS. But when contemplating real time voice and video, the technical challenges are complex in the extreme.

2. Future-proofing (and it’s overlooked cousin, past-proofing). Let’s say you could justify the initial build and go ahead with an inhouse developed solution. Going forward you will be faced with the challenge of ensuring  your RTC implementation stays current, and that it keeps working with those old browser versions and mobile operating systems customers insist on using, as well as the new versions which are released seemingly on a quarterly basis. The WebRTC standard is relatively new and it is evolving rapidly. New mobile devices and browser versions are always coming along, and competitors will always be adding new features and experiences to their websites and mobile apps. Keeping up with that will be challenging to take on, even if you have the spare engineering cycles and expertise to do it. Platform providers, like TokBox, have dedicated teams of engineers, product managers and QA guys focused solely on making sure this stuff just works, regardless of evolving protocols and emerging technologies.

3. The power of multi-tenancy. Behind our APIs and the OpenTok platform lies our global infrastructure. Think of this as multiple regional data centers running the core capabilities you need to power your RTC features: signaling, firewall traversal, media routers for multi-party calls, databases, web servers, and archival systems. These are monitored and maintained by our ops team and service provider partners 24×7. The fact that these systems are shared by many customers gives a hosted platform provider like TokBox a huge advantage in economies of scale. A company trying to do a single point implementation for RTC could never do it as reliably, predictably, or as cost effectively as a hosted provider like TokBox.

There is one more consideration we do occasionally hear, which is the perceived enhanced privacy and security of an internally hosted solution. If this post on build vs buy has been helpful to you, you might want to stay tuned for our upcoming post on security, identity management, and privacy.

  • Olivier

    This is a good question…I think that we must compare a service solution like Opentok based on usage fee versus an in-house solution built on top of existing librairies that can be baught to suppliers that give maintenance and support. When you already have a scalable, redundant infrastructure on top of AMAZON for instance, the questions needs to be adressed based on volumes of minutes.
    I do think that Opentol should have a look on a pricing based on connexion instead of minutes (like Bistri for instance) to be really competitive against in-house solution. And I’m an Opentok customer not a Bistri one ..at the moment :)