Ops Operating System
16 articles on ops operating system.
The Wiki Goes Stale Because Updating It Is Nobody's Job
A single source of truth works for what changes rarely. Split it: stable facts get a page, reasoning gets a dated entry, current state needs something else.
The Hard Part Is Finding the Person, Not the Document
Org charts answer reporting lines, which is not what you needed. Build a who knows what list from questions people actually asked, with a second name per row.
Search Is Not Sharing: Turning a Chat Answer Into Something Findable
Stop asking people to write things down in advance. Work backwards: answer in the channel, restate the question, say when it stops being true.
Most of That Searching Is Looking for a Person
Atlassian puts a quarter of the week into searching for information and answers. The answers half ends in asking a colleague, and then in waiting.
How to Know Where the Work Stands Without Asking for Status Updates
You want one answer about one thing, not a dashboard. Let the person author the record, then read it when you need it instead of asking.
How to Stop Being the Bottleneck in Your Own Company
You are the bottleneck because you are the fastest route to an answer. Count the questions you answer twice, then make the other route cheaper than asking you.
Blocked on a Person: Measuring the Waiting in Your Delivery
Blocked usually means asked someone, no reply yet. Split the status in two, add who and since when, and the invisible cost becomes a number you can act on.
Why Projects Slip on Distributed Teams: Waiting, Not Working
Work waits for people to be free, and across zones each wait has a floor of one night. Measure elapsed against active time on five items.
Project Status Reports Executives Read, and Stop Asking About
Executives read for exceptions and decisions. Five lines: confidence, what changed, one risk, the decision you need, and next.
Cross-Functional Alignment When the Functions Are in Different Countries
Sales promising what product never heard is a commitment made where the other function could not see it. Fix the commitment path, not the meeting cadence.
Reducing Escalations: Most of Them Are Questions That Had an Answer
An escalation is what a question becomes after 48 hours of silence. Classify a month of them and the information category dominates. Reduce latency, not triage.
Making Decisions Stick Across Time Zones
Across zones a decision is made in one place and received secondhand in another. The reasoning and the authority are what get lost.
The DRI Model in Distributed Teams: Who Answers When They Are Off
A single directly responsible individual works in one zone and bottlenecks across three. Separate decisions from context, and only the decisions wait for them.
Who Owns What: Ownership Clarity for Distributed Teams (RACI and After)
Two people thinking the other has it is a missing boundary. RACI answers roles, not questions. Replace the matrix with one line per area.
A Weekly Ops Review That Is Not a Status Meeting
Two thirds of a typical ops review is updates read aloud. Move them to a pre-read, refuse to repeat them, and limit the agenda to off-plan items and decisions.
The Operating Rhythm for a 200-Person Company in Three Countries
A cadence on paper describes a tenth of how a company coordinates. Design the meetings around decisions and the gaps between them around answers.
Your team’s week, answered from the work.
If this is your team’s week, see what a StandIn does with it.