Plato HQ hooked me up with Hal Eisen on a conference call AMA, where we talked about achieving consensus on a team.

Hal is a VP at VeriSIM, which produces machine learning to create better simulations for aiding pharmacological research.

Hal was expecting more folks than showed up – and my name wasn’t on his guest list – a couple of things for Plato to look at, I guess. He also only had two people at his previous AMA.

The audience was pretty diverse – another entrepreneur, a lead architect, a few new managers.

Couple of names that were easy to search for: Barrett, Yevgeniy.

Hal has been director of Engineering at He has volunteered with Plato for about six months.

Hal started off with a story about database administration that got progressively harder. When he approached his executive team for buy-in on improving the situation, and got it, his team produced a very divergent set of proposals. Because the goal was for the changes to be long-lasting, the decision had to be good.

He decided to use a consensus process – its pros were to get everyone to “feel included” in a team where the levels of experience were extremely varied but everyone wanted their opinions to be heard.

The other pro is that it will, with enough time, a close-to-optimal solutions – and perhaps even “win-win” solutions. Third, what happens to the team in a healthy process is the building of lots of connections between team members.

This naturally results in getting a lot of buy-in from team members.

One drawback is that the process can be slow, and it’s not a good fit when the solution is required very urgently. Another challenge is that not all groups are well-suited for it – it’s not a one-size-fits-all process. Some people have a hard time listening to others, some are shy, etc. The leader/facilitator has to account for this.

The last risk is the emergence of “group think” and the desire among some to stick with majority and not risk skepticism. The moderator has to watch out for these people who are “mailing it in” and properly re-engage them.

One way to think about it is to use this process when you have a “Type I decision” to be made, to use Jeff Bezos’ taxonomy.

The Important Steps

This is a process he has distilled from various experiences.

  • Set up ground rules
  • Have clear expectations
  • State the initial (business) problem clearly
  • Have exit criteria/metrics that are used to determine the final decision
  • Make sure to disagree, but commit. If the process doesn’t require unanimity, the minority needs to commit to the final decision.
  • Set up a Brain-storming Process, then set aside time for research that could even involve building proofs of concept and prototypes.
  • Make sure to celebrate at the end!


How to handle people who have a hard time participating in groups: Have them work on important ideas that can be done solo but coach them to learn how to work in a group

The role of moderating – it helps to have a neutral party – do you step away often? Managers who insert their own opinions can stifle feedback. Hal finds himself able to compartmentalize his brain, to activate the management/moderation piece over the part that wants to contribute. It might also help to recruit a specifically neutral moderator.

Given the two types of decisions, and the need to select the right people for the team – when the team is cross-functional, the latter is hard to achieve. What do you do then? Disruptive people can certainly be a monkey wrench. If you can identify a particular bad actor, start by talking to them 1×1 about what their perspectives are and understand their motivations. If they didn’t want to participate or don’t see self-gain, you have to sell the process, as the moderator. You can also choose to escalate first to the participant’s manager, if the participant is reluctant to hear you out – specifically, use a situation behavior impact summary, perhaps. You can then choose to escalate to your own chain of command.

When setting decision making criteria, how do you handle needing to have a solution in the short term, and have a backstop, so to speak? Make it clear to the group where they are on the percentage scale of time used. At about a third, step in and note that a decision has to be made to move the process along. Hopefully, when the group sees this happen at different points, they will automatically start to feel the pressure themselves.