When is push or pull appropriate for innovation orgs

Almost definitionally, innovation orgs depend heavily on “talent.” When ‘talent’ isn’t code for ‘specialized training’ it means the role or industry has not been systematized. However, there seem to be at least two distinct ways that different innovation orgs approach finding talent that are tightly coupled to the ideas of ‘locus of control’ and ‘idea inception.’ What are the different roles in an innovation org?

Both DARPA and Bell Labs took the “pull” approach. Bell Labs hired people before they had a role or expertise and would go out and just try to hire really smart people but then would give them a project to work on. DARPA is very much a “you don’t call us, we call you” system - sometimes it’s because someone is just doing good work in an area vaguely related to DARPA’s area of interest or because DARPA has a specific program for them to run.

Venture research on the other hand takes the push approach (ie. The person pushes themselves into the organization). They wait for people to come to them. (This is also the case for venture capital for the most part.)

It’s pretty clear that In an innovation org, the person bearing the risk needs to feel extreme ownership over an idea for it to succeed. On the one hand New ideas need to come out of a single mind, which might suggest that the push approach would have better results. On the other hand, people grow to feel ownership over ideas that didn’t start in their own heads all the time in grad school and throughout history. Additionally, both DARPA and Bell Labs gave people space to come up with ideas once inside the organization, which is a key difference from venture research.

There’s also a different valence between the two approaches. The push approach has the feeling of “you convince me.” Someone needs to want the thing to happen badly enough that they’re willing to work with you to explain the idea before committing to the organization. Venture research built trust by iterated meetings to refine precision around the idea. This approach seems like it might be more tenuous in the 2020’s than the 1980’s because of how the internet has warped engagement levels. The internet has warped our engagement levels. The pull approach has the feeling of “come join us” and requires a level of sales/convincing that joining the organization is the way to turn ideas into reality.

One wonders how much performance-affecting bias the pull system creates. By definition the pull systems will only recruit people who are inside their sphere of attention. You see DARPA primarily pulling from people they’ve worked with before and Bell Labs inordinately pulling from a few top-level universities where other lab members had gone to. At the same time, you have to consider the uncomfortable possibility that Bias can enable more effective action. In the case of finding invention and discovery folks, the more you know what you’re looking for, the less this bias matters.

In practice, the approach will probably end up being a mix between the two, especially for a small/new organization. One could make an argument that when you are unproven you will have a hard time convincing well-known A-players to join so leaning into the “push” mechanism is important. At the same time, the initial people you work with can shape a culture and trajectory, which makes an argument for the “pull” mechanism. On the one hand, you could say “we will front load creditability via <funding, well-known-supporters, thought leadership 🤮> so that relatively unknown people will push. On the other hand, that strategy feels cringy and incorrect for an organization that wants to be a bit under the radar. The compromise might be to figure out exactly the creditability you need to get the people you want. Of course, that becomes harder the less clear you are on the people you want to get. Which is a bit of a chicken and egg problem. One solution might be to track a gut feeling on where you fall on a spectrum of “I’m finding the right people but they won’t give me the time of day” to “I don’t even know who the right people are”

Perhaps another way of looking at it is to ask “how much knowledge does management have about the area.” At some point, if you’re sober and paying attention to a specific field, it becomes increasingly unlikely that there are high-quality people who have done enough work to actually know their shit but are not on your radar. So the question of push vs. pull might depend heavily on a recruiter’s specialization. In the case of Venture Research, they were looking for people far outside their realm of expertise. Bell Labs was under a much more finite umbrella, and DARPA is somewhere in between.

There’s a tricky question about idea inception and tension between goals at different levels of the organization. Research requires massive intrinsic motivation but at the same time you need that motivation to be aligned with organizational priorities unless you’re all the way on the venture research end of the spectrum. It is especially hard when you’re working towards a specific vision or endpoint. Somehow, J.C.R. Licklider managed to pull this off. Alan Kay talks about how old school ARPA gave them complete freedom, and yet we know that Licklider managed to coordinate several independent research groups to collectively produce The mother of all demos so obviously they didn’t actually have “complete freedom.” It’s wild speculation but I suspect that Licklider managed to “chunk up” the vision into boxes. He could then find people who already had the intrinsic motivation to explore within those boxes. This seems extremely hard. <need to finish The Dream Machine>

What are other well-known examples of “pull” organizations who get people in the door because they are smart and generally competent? Many startups claim to be this way, but whether that’s actually the case


Web URL for this note

Comment on this note