Derrick Bradley home

Making decisions at work is hard. People don’t like doing it. They don’t seem to mind putting a prickly decision off, even when it slows their organization down. Usually that’s because people don’t want to have the finger pointed at them when things go badly. This is particularly true of large organizations. This decision-making paralysis slows everything down and can cannibalize people’s engagement.

One way we’re fighting that entropy at Undercurrent is by changing the way people meet. It’s deceptively simple, but has become one of the more effective ways to introduce responsiveness inside of an organization. It’s grounded in an existing decision-making process called Integrative Decision Making (IDM, Figure 5) that we’re continually adapting for us and our clients.

The basic format is this:

  1. Present Proposal. Proposer describes her tension and proposes a solution.
  2. Clarifying Questions. Participants can ask questions of the proposer to clarify their understanding. Proposer can respond, but no other discussion is allowed.
  3. Reactions. Participants can voice support, share their opinion, and suggest changes. Proposer cannot respond, Again, no other dialogue is allowed.
  4. Amend & Clarify. Proposer has the option to amend, clarify, or remove the proposal from the table. No discussion.
  5. Objections. If anyone objects to the proposal they are captured and tested for validity (more on this later). The proposal is adopted if none surface.
  6. Integration. The goal is to craft an amended proposal that satisfies any objections and the original Proposer’s tension. Each objection is processed one at a time. Once they are all integrated, go through another Objection Round.

Each session has a clear purpose with clear expectations of participants and each round is facilitated to keep everyone in process. If this sounds hard to practice and robotic in nature it’s because it is. Until it’s well practiced it feels awkward and frustrating.

People believe they are losing agency and losing voice because what they can say and when they can say it is structured – that not conversing is inhuman – but I firmly believe that this process, done well, actually materializes people’s agency most perfectly for the task at hand – thus it is highly ethical – but it often requires a deconstruction of some boundaries or overcoming of some fears to actually feel that - this is why trust is so important. - @Dara

This process has a bent towards action and enforces some important working principles:

What follows is a mix of observations and tips from our experience practicing this method with our clients. It’s broken up into the first three stages of the IDM process and will eventually include the others in a follow up post. There are many things we have yet to figure out, but hey, we’re learning.


Before kicking off an IDM session there are a couple of things worth paying some attention to and calling out to participants. Here’s what I find to be most important when setting the stage:

Establish clear roles. IDM works well when someone facilitates. The facilitator has the authority to enforce a few basic rules like no crosstalk (speaking in turn) and can callout participants for things like disguising reactions as questions. They keep everyone moving forward. The second and equally important role is that of a secretary. They record all of the decisions that are made in the meeting and help folks refine their proposals. These roles can quickly be elected at the beginning of an IDM session and held for any period of time.

Team size. The number of participants in the room is the single biggest lever we have over the efficiency of IDM. We try and follow a 7± 2 rule (two pizza rule stuff), but it’s often larger and the unsound reasoning for that goes something like “I can’t invite this person without inviting this other person so they should both come, you know?” The last client IDM meeting I attended had 13 people in the room +1 on the phone. I timed the length of each proposal, clarifying question, and reaction round: we averaged ~8min per round (8min proposals, 8min clarifying questions, 8min reactions). In a 90 minute minute we processed 4 proposals. In some of our more experienced groups (following the ±2 rule) I’ve seen teams process ~20 proposals in that same amount of time. Given, this was the first time this group was trying this process on (they did a great job), but no matter how familiar a team of that size is with the process, it just won’t scale – and isn’t meant to.

Table position. Where people sit in IDM matters a lot (just like in poker). When following a linear order for clarifying questions, reactions and objections we find people in an early position checking out halfway through the proposal to carefully craft their own reaction (and unintentionally anchor the room) and people in a late position not having much to add because their point has already been covered, diminishing their voice. We believe it’s best practice to randomize the order after every proposal. This method has worked well with some of our largest clients and helps keep everyone engaged in the meeting. Oh, and consider seating the leader in the room somewhere in the middle of the pack – you don’t have them having the first or last word.

Proposal duration. We suggest proposals are adopted for a month before they’re revisited. This short cadence helps the proposer get someone’s consent to adopt their proposal and keeps things from becoming overly abstract. What’s going to break in the next 4 weeks if we try this? It is really going to do any harm to our business? It’s easy for people to dream up reasons a proposal might fail – those are affectionately referred to as future tensions. If participants can’t back their objection with data that suggests a proposal will cause harm in the next 4 weeks, it’s invalid and the proposal passes.


Format: The proposer describes the tension that underpins their proposal and suggests a solution to resolve it. Only the proposer speaks.

Proposals are hard to write. Undercurrent has been at this for well over a year and we still struggle with it. Often times someone will call a ‘timeout’ and ask for help crafting their proposal because they’re unsure of what shape the resolve to their tension should take. In Holacracy, the majority of proposals come in the form of a policy, accountability, or role. For all three, you can create something new, update something existing, or propose to remove a thing altogether (roughly following CRUD).

This process exposes so much about a culture – not just about communication of ideas, but the socialization of ideas (e.g. consensus building and loss of personal stake/viewpoint and the ability to represent one’s own roles). -@Dara

We don’t follow that same proposal structure (policy, accountability, role) with our clients, in fact we don’t have any proposal structure, but I think we need to find a version that works for them. The proposals I’ve seen from our clients tend to be obtuse and unstructured. Sometimes they’ll come packaged as a 4-5 page decklet, where each page or bullet point feels like its own proposal, making them hard for the rest of the room to react to. This approach works, but it is noticeably slower than ones that are structured. If a proposal is always made up of the same parts they’ll be easier to write, react to, and more importantly train (stickiness) and scale across a large organization.


Format: Anyone can ask questions to better understand the proposal. The proposer can answer or say “unspecified” if they don’t have an answer; people tend to embrace this option to help things move along. Anyone can ask as many questions as they need. It’s not a space for back and forth discussion.

When starting out people people are confused about how to participate in the question round “Can I say whatever I want… or what?” And in equal measure the proposer tries to answer every question as thoroughly as possible, even when they don’t have a great answer. What’s interesting is this: often times reactions will come disguised as questions when teams are being introduced to IDM. And if not that, they’ll focus on ‘the how’ of a proposal during the question round. There’s a natural tendency to jump to execution and skip over the root tension of the proposal, which feels like a legacy behavior. Everyone needs to put their stamp on a thing before it goes out the door sorta stuff. From our perspective ‘the how’ shouldn’t be a part of any proposal, it should be left up to the team executing to decide how to execute.


Format: One at a time, each person reacts to the proposal as they see fit. No response or interruption is allowed during someone’s reaction.

Not surprisingly, we see a lot of verbose reactions when we do this with clients. Everyone is eager to share all of their thoughts, but you don’t score any points for a long reaction, but that’s antithetical to what most are used to. Generally, those who have suggestions for an edit to the proposal take more time to react and those in support of a proposal are brief: ‘no reaction’ and ‘safe to try’ are common responses. There are no rules around this, but that’s the natural tendency.

The practice of responding ‘safe to try’ to a proposal is particularly interesting because it carries a lot of meaning (if you let it). Safe to try means different things to different clients, but generally you’re giving the proposer your consent to take action on their proposal and that you believe it will not cause immediate harm to the business. The not-causing-harm bit is a nice filter when you work in a transparent, small organization, but it gets harder as you grow and I don’t know that we’ve figured it out with our larger clients. How do we know it won’t cause harm? Do we set risk thresholds? Who sets them? If so, are they central or local? That all sounds overly complicated. I haven’t seen it yet, but my hope is that this challenge of determining what is ‘safe to try’ pushes proposals to get smaller in scope – make more small bets, not fewer big ones.

It’s really fun to watch clients get good at this. There’s a whole ecosystem of tools and working principles that work nicely with IDM and we’re slowly finding ways to introduce the full spectrum of becoming responsive to our clients.

I’m leaving the second half of the IDM process out of this post on purpose. As we continue to battle test our own version of IDM we’ve noticed that cutting the process in half is surprisingly effective with clients and so I’ll follow suit. Only looking to gather feedback? Consider using the first half of IDM (proposals, questions, reactions) to get the job done. Need to make a decision? Run IDM end-to-end.

Fork me on GitHub