One of the important aspect of my current role at Google is to understand what to invest in. Its probably one of the fundamental skill and traits of a PM when kicking off the new initiatives or pulling plug on ongoing initiatives.
As always, there is is a high road and low road of thinking about this. I like to take time, iterate and think critically backed up with data when making this decision
This management space is flooded with many frameworks and we have many thought leaders in this area. I’ve developed some of my own decision framework with observing what has worked and what might work.
First side note:– There is a fine line between “How to prioritize” and “How to make the list of things to prioritize”. I like to think this is the later bucket and more suitable for a portfolio approach of initiatives that are inter-related.
Second note:- “Success” of a research initiative is super subjective and is a broad spectrum. A success story for research initiative or product research may not be a the best approach or “Problem setting” for a product. As a broad stroke, one way to quantify a research success is the impact on the state of the art many times in academia.
I’ve tried to assemble few thoughts with respect to research initiatives and getting research into products. This is somewhat also applicable to product research but approach could be super different if we look at this from a pure product investment.
Quadrant 1 – High likelihood of success and less effort to build a technology
In research, its highly unlikely that something like this exists. In today’s software technology landscape- where most of the low hanging items are already explored and exploited, increasing amount of efforts are needed to make a dent in almost any vertical. However, if something like this appears its mostly the “enabled thing” because of some breakthrough research or new technological advances that made a hard problem easy. If this is the case- its a great investment.
For a product team, as lucrative as it might sounds to jump on these opportunities, these are usually the small “snack” projects and has a great hidden danger of getting into the trap of building things that result into tiny incremental changes. There is nothing wrong with this for a mature product or a product that is in it’s later stages where late majority is on-boarded. However, there is a high likely-hood of team loosing interest if all they are doing is incremental. If we find a team only doing this, its a great indicator to look inside for details.
Quadrant 2 – High probability of achieving what we set out for but more effort
This is tricky and usually needs top down push for motivating team (in bigger organization). Efforts in this area needs more stronger vision since stakes are high. For incumbents, I would argue this could be a great strategy – as startups can’t afford to invest in these initiatives or expensive research may not results into great ROI since they have poor product ecosystem. I’ve also observed most research in this area falls into niche of a company or a product line (typical pharmaceutical industry for e.g.) The “high possibility” bit is tricky. At the end, research is as good as one’s hypothesis – in this case, probably backed up by stronger data points. In Software, there are many projects in this category that graduated to products where outcome was “good enough” to be useful. Arguably, self driving tech might fall in this bucket.
Quadrant 3 – Low probability and high effort
This seems to be no brainier to start with. Why would you want to invest in something where outcome is less likely to be a success and efforts are high? This bleeds into strategy and philosophy question of the company. Many times, the goal is to go through the research process. There are many situations where hypothesis is negative, in other words “Let’s find out what doesn’t work”. Many times, few serendipitous moments leads to some of the best outcomes. At Google, these are typically moonshots and are possibly encouraged for breakthrough innovation. Again, bigger companies with deep pockets invest in this area. After all “Is there a thing as bad research?” 🙂
Quadrant 4 – Low probability, low effort research projects
Personally, I would avoid investing here. Many times it feels like typical individual projects that are coming up as a result of gut feel or extension of an older projects. Could be great candidates for someones’s side project. These initiatives have great potential to “Look team busy”.
Investment decisions are hard and should be a team work. What decision making framework to use depends on many parameters. Strategy and culture being the primary followed by target product. Company “A” may not invest in an idea if high level roadmap doesn’t align, and at the same time, competitor company “B” might double down in that direction.
Technological landscapes change fast, efforts that are relevant today may not be relevant next quarter next year. Its hard and needs awareness to overcome strong desires of investing in projects that are sunk cost.
I hope some of the thoughts here resonates with folks who are thinking about profiling their portfolio for research and product investment.