Convergence pattern
Find connections between clusters the group treats as separate.
Purpose
Find arrows that cross from one cluster to another. Participants fill the part of the board they know, and connect it well. The connections between parts often belong to nobody, so there's a chance nobody will draw them.
Those arrows are often the most valuable on the board. They show how one group's problem becomes another group's problem.
When to use it
- At the end of Group, before the first arrow is drawn.
- During Connect, when the first pass of arrows is in and the board shows several clusters with nothing between them.
- When each function in the room has connected its own area and stopped.
Timing
Allow a few minutes for each pair of clusters. The pattern runs inside Group or Connect and uses their time.
Steps
-
Step back until the whole board is visible. On a digital board, zoom out.
-
Pick 2 clusters with no arrows between them. Favor clusters that different functions filled.
-
Point at both, and ask:
I see 2 separate clusters that might be related: this one about deals, and this one about releases. Does anyone see a connection between them?
-
Wait. Participants need time to read a cluster that is not theirs.
-
Draw the connection between 2 specific nodes, not between the clusters. If it runs through something that is not on the board, add that node first.
-
Ask who else sees it from their side.
Does anything else feed this?
-
Move to another pair of clusters, or go back to the step you were in.
Tips
- Send people to a cluster they did not fill. The person who sees the connection could be reading the other side for the first time.
- Look for a shared cause or a shared outcome. 2 clusters often rest on the same problem, or serve the same outcome, without either saying so.
- Let the group find it. If you already see a connection, ask the question and wait. Offer yours only afterwards, as a question. See the facilitator's role.
Traps
- Forcing a connection. Some clusters are separate. If nobody sees a connection after a fair pause, move on.
- Connecting clusters instead of nodes. A cluster frame is a reading aid and carries no meaning. An arrow between 2 frames says nothing anyone can act on.
- Connecting everything. A board where every cluster points at every other is as hard to read as one with no arrows. Draw the connections the group agrees matter. See runaway scale.
Variations
- During Group, without arrows. Nothing is connected yet, so the answer is not an arrow. Move the 2 clusters next to each other, and come back to them in Connect.
- In subgroups. In a large group, each subgroup connects its own area during Connect. Ask each one to bring back 1 connection from its area to another.
Example
The first pass of Connect in the renewals session left 5 well-connected clusters and almost nothing between them. Sales and product had drawn the arrows in deals and roadmap. Engineering had drawn the arrows in releases and reliability. The facilitator pointed at both and asked whether anyone saw a connection.
Step 1 of 3. Two clusters the group has treated as separate: deals and roadmap, and releases and reliability.
Step 2 of 3. Does anyone see a connection? Engineering: the reshuffled roadmap is why reliability work never makes it into a sprint.
Step 3 of 3. Support adds that escalations eat the same time. One question connected 3 parts of the board.
Diagram as text: The convergence pattern on the renewals board
Steps:
- Step 1: Two clusters the group has treated as separate: deals and roadmap, and releases and reliability.
- Step 2: Does anyone see a connection? Engineering: the reshuffled roadmap is why reliability work never makes it into a sprint. Changes: added "Reliability work never makes it into a sprint", "Reserve 20% of every sprint for reliability". new arrows: "The roadmap is reshuffled for every big deal" influences "Reliability work never makes it into a sprint"; "Reliability work never makes it into a sprint" influences "Reserve 20% of every sprint for reliability"; "Reserve 20% of every sprint for reliability" influences "Roll out to 5% of customers first".
- Step 3: Support adds that escalations eat the same time. One question connected 3 parts of the board. Changes: added "Most tickets are escalated to engineering". new arrows: "Most tickets are escalated to engineering" influences "Reliability work never makes it into a sprint".
After the last step:
Nodes:
- Root / Problem: "Sales promises custom features to close deals" (in "Deals and roadmap")
- Root / Problem: "The roadmap is reshuffled for every big deal" (in "Deals and roadmap")
- Root / Problem: "Releases ship in big monthly batches" (in "Releases and reliability")
- Root / Problem: "Most tickets are escalated to engineering"
- Root / Problem: "Reliability work never makes it into a sprint"
- Action / Intervention: "Lock the roadmap each quarter, with one slot for deals" (in "Deals and roadmap")
- Action / Intervention: "Release weekly behind feature flags" (in "Releases and reliability")
- Action / Intervention: "Roll out to 5% of customers first" (in "Releases and reliability")
- Action / Intervention: "Reserve 20% of every sprint for reliability"
- Desired State / Outcome: "The roadmap holds for a full quarter" (in "Deals and roadmap")
- Desired State / Outcome: "No outages during the morning rush" (in "Releases and reliability")
Arrows:
- "Sales promises custom features to close deals" influences "The roadmap is reshuffled for every big deal"
- "The roadmap is reshuffled for every big deal" influences "Lock the roadmap each quarter, with one slot for deals"
- "Lock the roadmap each quarter, with one slot for deals" influences "The roadmap holds for a full quarter"
- "Releases ship in big monthly batches" influences "Release weekly behind feature flags"
- "Release weekly behind feature flags" influences "Roll out to 5% of customers first"
- "Roll out to 5% of customers first" influences "No outages during the morning rush"
- "The roadmap is reshuffled for every big deal" influences "Reliability work never makes it into a sprint"
- "Most tickets are escalated to engineering" influences "Reliability work never makes it into a sprint"
- "Reliability work never makes it into a sprint" influences "Reserve 20% of every sprint for reliability"
- "Reserve 20% of every sprint for reliability" influences "Roll out to 5% of customers first"
An engineering lead answered first. Every time the roadmap is reshuffled for a big deal, the reliability work planned for that sprint is the first thing dropped. That became a new problem, "Reliability work never makes it into a sprint," with an arrow coming in from the roadmap cluster. A support lead added a second arrow: escalations eat the same engineering time.
The new node later got 5 votes. The path through it, from sales promises all the way to outages, was one of the 2 insights the group named during Close.
The facilitator had suspected this connection for months. She asked the question and stayed quiet. The connection meant more coming from the people who live with it.