I spent the first decade of my software engineering life building technology stacks for digital TV businesses for the likes of DirecTV, Sky, Multichoice DStv, Liberty, US Cable, etc - working at that time, for what was the world's leading Digital TV Technology Services company, NDS. We were in the business of selling full-stack embedded software (like Android / iOS SDK) tailored for set-top boxes (STBs), along with the backend infrastructure services needed for digitising, encrypting and transmitting TV signals over-the-[air/cable/internet] to these consumer device STBs, so people can basically watch TV. We would sell the tech stack along with a suite of TV applications that could essentially be tailored for any type of customer need, including changing different look-and-feel frontend/user experiences, configurable features like live recording or basic watching without recording, on-demand or internet streaming -- all without having to run multiple versions of software codebase per customer. In modern software parlance, we built multi-tenanted technology stacks in a way.
In 2010, we embarked on a vision to harmonize in creating the Nextgen version of the platform - taking the best of all customer engineering projects and core platform enhancements, and creating the next-generation stack, to scale to as many video entertainment providers across the world. We called this initiative Project Sunrise, symbolizing a new dawn for the next-generation experience built on Fusion Mediahighway Advanced offering a fully customizable Snowflake Unity UI experience.
Here's a short clip from 2011 on Snowflake:
Back then in 2010, I'd just come off delivering what was the biggest migration program in the history of the company, launching a new service for Sky - and thereafter landed another client build for UPC Horizon Gateway STB. In addition, we had 5+ other customers all lined up for new tenants! It was going to be a busy next few years indeed.
Our approach to building this technology stack included foresight from the very beginning. We were intentional about using a single stack, configurable architecture end-to-end, including customizable applications for custom experiences - to avoid rework and duplication, but most importantly quick delivery turnaround times. Our customers also benefited from leveraging features and capabilities they didn't have to pay for, because some other customers would have already funded the development anyway :-)
Technical Program & Product Management - Visualising the Roadmap
As our company was primarily a technology engineering company and a high-growth start-up, resources were constrained such that people took on multiple roles. I led the Sunrise project covering technical product and program management. I was responsible for creating the roadmap, backlog and overall sequencing, coordinating with multiple customer-delivery streams, along with the main core platform engineering deliveries - building the next-generation stack. The engineering activities were a mix of software integrations and application customizations through configuration, building out the default flagship application, that would come "out-of-the-box" for selling to prospective customers. The sales team would close the sale by signing off on the profiling customizations and configurations - and the Sunrise factory would eventually produce a release for the customer. This pattern of template-driven, profile-based software configuration approach was not new to us, but the technologies we used had changed over time (see the slide deck at the end). Project Sunrise was never a fully funded initiative though, so we had to partner with customer project teams and core platform engineering, being scrappy and inventive - but still ensuring we have a reference stack available, at all times, for new sales.
How did I communicate the Roadmap then?
I had a ton of detail to manage, spanning multiple customer requirements backlogs - working with teams across the globe, managing a unified backlog, understanding the features and gaps, prioritizing features for the base profile & then owning a delivery plan (which I'll expand upon in future posts). The one mechanism that earned the trust of senior leadership was a visualization I produced, that showed on a single piece of paper how all the streams fit together. Once the executives saw the roadmap, they then had an easy mental model to understand the complex pieces and stages of convergence - that went into building out the NextGen Sunrise platform.
I decided to write this blog piece today, 13+ years later because, it so happens, I find myself now again responsible for building a Nextgen product (V2), with V1 (single tenant) that is currently supporting existing customers with an active roadmap - and V2 targeting multiple-tenants onboarding new customers with their own specific configs/capabilities - and my team are considering ways of communicating the plan!
Check this out:
The above roadmap accomplishes the goal of showing the interplay between Customer engineering deliveries (above the Sunrise line) and simultaneously showing the core platform engineering feature deliveries (below the Sunrise line) - all contributing to the holistic platform called Sunrise. So customers mutually benefit from the platform core and they themselves benefit from the internal platform development. All of these deliveries are contained within a timeline that serves as the roadmap.
As I revisited this picture, thirteen years later, I can still appreciate the value of a picture like this - a powerful visualization that will beat any detailed text narrative IMHO.
And here's our scrappy Project Kick-Off Charter
Oh, and here is the initial MVP we demoed on Sunrise:
No comments:
Post a Comment