Roadmap to the future

This is Facebook’s roadmap from their F8 Developer Conference in April 2016. It’s cool to see that Zuckerberg and the Facebook team were publicly preparing for a long-term journey of pushing the frontier in AI, drones, satellites - which, 8 years later, are today’s most important technologies. It’s also impressive to see their continuous dedication to VR/AR. Despite having poured $63B into the sector with not much to show for it (yet), this commitment to their vision of the future is a testament to what a founder-led company can pursue.

Building roadmaps at the birth of the mobile era

I started my career at Samsung as a Product Manager on the mobile processor System-on-Chips (SoC) that went into the first generation of smartphones in the late 2000’s. One of the PM team’s primary responsibilities was to develop and manage the product roadmap. We tried to craft a vision of the future and worked a broad spectrum of teams: internally at Samsung, with customers, and with partners building other parts of the technology stack. Some of our key responsibilities in driving the roadmap included:

• Imagining the future of mobile: Will we need to support HD quality video? Will consumers play premium 3D games? These were very theoretical questions in 2007. We often fielded questions like “Why will people watch HD video on their phone when they can watch it on their laptop??”
• Drive the ecosystem: We needed alignment between the capabilities of our SoC and the capabilities of the ecosystem: the operating system, wireless connectivity, GPS, etc. Alignment with the ecosystem allowed us to push the envelope on key technologies that were needed to enable a new paradigm of computing in our pockets.
• Drive internal R&D teams: In a rapidly evolving environment product launches always seemed delayed. Customers and/or the ecosystem would require modifications to product specifications, and as customers rushed to introduce new devices ahead of their own competitors, product launch timelines were a moving target. Ensuring the R&D teams were aligned with these timelines, often causing intense sprints and engineer all-nighters, required crisp yet transparent communication, with the roadmap serving as the north star.
• Manage feedback from customers: Rapid iteration in the world of semiconductors has very different timelines than the world of software. It was the Product Management team’s job to manage the expectations of the sales teams who wanted everything the customers asked for, ASAP, with the meticulous prudence of the R&D teams who (rightfully) had no intention of rushing products out of the door. The roadmap, serving as a quasi-source of truth, played an important role in aligning a very large organization with teams scrambling in a highly dynamic environment.

As I wrote about in May, there are many parallels between the mobile shift and today’s AI shift. Rapid iteration is paramount to provide value to customers in a way that wasn’t possible before. At the same time, having a clear roadmap (that will certainly go through its own rapid iterations) is important to align the team while marching towards a grandiose vision.

The role of roadmaps in evaluating startups

So how do I think about roadmaps as I work with early stage startups? Two questions I always ask founders are:

• What is the most important engineering priority for the next 12 months?
• How about for the next 5-10 years? What does your product/service look like then?

Every founder has a very clear idea of what they need to accomplish in the next 12 months. Almost all of them have a vision for the next 5 years. But the most impressive ones have an incredible sense of excitement in their long-term vision and what they can accomplish in 5-10 years. I always ask these two questions as they help me to understand what has been built already, what needs to be built today, and what can be built if everything goes right for the next 5-10 years. The long term roadmap discussion also helps me imagine how big the company can be and how ambitious the founder is.

I love roadmaps as they are an embodiment of a company’s vision. Sometimes we look back and a roadmap did an incredible job predicting the future like Meta did in 2016. But in most cases, the roadmap will have countless iterations as both external & internal factors change. I’m sure this will be the case for the vast majority of companies building around AI, and that’s completely fine. Building the future is a really hard endeavor!