Decision Velocity: How Senior ICs Lead Teams Without a Title
For mid-career designers who lead by influence, not org charts.

Decision velocity = how quickly a team moves from problem to clear decision to action, without trading away quality. And you don’t need a title to impact it. In this piece you’ll get: a one-page decision map (RAPID-lite), a pre-read that prevents meeting rehash, a clean “consent vs. consensus” move, a lightweight decision log, and simple metrics to prove it works. (Hello, influence without authority and fewer stalled projects.)
Table of Contents
Why decision velocity is a quiet superpower
The 5 common blockers (with counter-moves)
Playbook: raise decision velocity in 7 days
Scripts for sticky moments
Make it visible to compound influence
Proving it works (tiny metrics, real signal)
Sources and further reading
Why decision velocity is a quiet superpower
Design work stalls less on drafting and more on decisions: unclear owners, vague asks, too many cooks, or no deadline for feedback. Research from McKinsey & Company during and after 2020 found that faster decisions can coexist with quality when roles are clear, debate is focused, and teams feel empowered — all levers you can pull without a title.
Definition you can use in a deck:
Decision velocity: the speed from problem through decision to first action, at a quality bar the team accepts.
The 5 common blockers (with counter-moves)
“Who decides?” is fuzzy → Map roles on one page (RAPID-lite).
Re-hashing the last meeting → Send a short pre-read 24 hours ahead.
Waiting for full agreement → Use consent when consensus isn’t needed.
Decisions evaporate after the call → Log the decision in 5 lines.
We can’t tell if this helps → Track cycle time and first actions.
Let’s walk through each one—designer to designer.
1) RAPID-lite: a one-page decision map
When debates loop, it’s usually because roles are unspoken. Borrow RAPID and shrink it to something you’ll actually use:
R — Recommend (one person drafts the decision)
A — Agree (must sign off, e.g., legal/brand)
P — Perform (who executes after the decision)
I — Input (whose info shapes the decision)
D — Decide (the decider; name the human)
Post this as a single table in your doc and pin it in the project channel. RAPID is a Bain & Company framework — use the roles, skip the ceremony.
Tip: If you’re choosing between RAPID, RACI, or DACI, pick RAPID when the real risk is “who gets to decide.” Use RACI/DACI when execution ownership is the pain.
2) Pre-read, not re-hash: the 3-3-3 one-pager
Your meeting shouldn’t be the first time people see the problem. Send a tight pre-read the day before:
Evidence points (up to 3) — plain language, with a screenshot or quote
Constraints (up to 3) — tech, brand, legal, timeline
Viable options (up to 3) — each with a one-line tradeoff
Open the meeting by reading the info together (2–3 quiet minutes works surprisingly well). Amazon’s long-standing memo culture uses silent reading to align fast; you can adopt a tiny version without the six-pager.
Pro move: Put “Decision we’re here to make:” at the top with a deadline. That single line turns a status chat into a decision call.
3) Consent beats consensus (most days)
Full consensus is slow and rare. For many product/design calls, consent (no significant objections) is enough to move, especially when decisions are reversible. Sociocracy’s consent model is built for progress: ship “good enough for now, safe enough to try,” then iterate. Pair it with the phrase “disagree and commit” for momentum when opinions diverge.
Script you can borrow: “We don’t need unanimity; we need no material objections. If none, let’s consent to Option B for two weeks and review outcomes. If you disagree, can you disagree and commit so we can learn?”
4) Keep the receipts: a 5-line decision log
Re-debating old choices is a time sink. Capture decisions where the team lives (notion, wiki, repo) using a tiny Decision Log:
What: The decision in one sentence
Why: 1–2 bullets (evidence, tradeoffs)
Owner: Name + date
Next step: The first action with a date
Links: Brief, mock, Product Requirements Document (PRD), test doc
If you already use Architecture Decision Records (ADRs) on the engineering side, mirror the structure for cross-functional calls. ADRs exist to record intent and rationale — perfect inspiration for product/design decisions too.
5) Meeting choreography (so decisions actually happen)
Start with the pre-read. 2–3 minutes of quiet reading aligns context.
Name the roles. “I’m the R; Jamie is D; Legal is A.” (referencing RAPID)
Time-box debate. “We have 12 minutes for critique; then we call it.”
Default to consent. Consensus when necessary.
Close with the log. Read it aloud; confirm the next step and owner.
Playbook: raise decision velocity in 7 days
Day 1–2 — Pick one recurring decision that stalls (e.g., “Which variant do we promote to full rollout?”). Draft a RAPID-lite for that decision and socialize the one-pager.
Day 3 — Send a 3-3-3 pre-read for the next meeting. Make the decision question explicit.
Day 4 — Facilitate the call. Use consent if no material objections arise; ask for “disagree and commit” when opinions diverge.
Day 5 — Publish the decision log and link it in the project README / channel.
Day 6–7 — Measure: How long from pre-read sent to decision made? Did the first action start within 48 hours?
Repeat the pattern on another decision next week.
Scripts for sticky moments
When the owner is fuzzy:
“Before we debate, who’s the D on this? I’ll take R, and I need A from legal if we use this claim.”
When debate expands sideways:
“We’re deciding placement, not strategy. Two options on the table; we have 8 minutes to choose or set a deadline.”
When one team blocks due to preference:
“Is this a material objection or a preference? If it’s preference, can we consent to test for two weeks and revisit?”
When you can’t get unanimity:
“We’re past diminishing returns. I’m asking for disagree and commit so we can learn from real data.”
Make it visible to compound influence
Make a slide for weekly 1:1s. Screenshot the latest decision logs + “what happens next.”
Increase demo cadence. If the decision produces a change, demo the outcome — even if small.
Receipts doc. A living page of before/after, dates, outcomes. When review season hits, you’re not scrambling.
These small rituals create a “gravity field” — people start routing decisions to you because it’s faster when you’re involved. That’s leadership without a manager title in practice.
Proving it works (tiny metrics, real signal)
Track these for a month:
Decision cycle time: pre-read sent to final call.
First-action rate: % of decisions with an action started within 48 hours.
Re-litigation rate: # of times a closed decision gets reopened.
Assist metric: # of cross-team decisions you facilitated (shows influence).
If your boss wants external validation that speed ≠ sloppiness, point to mainstream management research since 2020 on faster and better decisions through clear roles and focused debate (links at the bottom).
Final thought
Influence isn’t telling people what to do — it’s making it easier for them to decide well and move. If you raise decision velocity on two or three recurring calls, you’ll feel the difference in a week. Keep the roles clear, put a real pre-read in front of people, move ahead with consent when you can, and log the call. Titles don’t matter like momentum.
Sources and further reading
Working Backwards: Insights from Inside Amazon - Colin Bryar, Bill Carr
RAPID Decision Making Framework - Bain & Company
What is decision making? - McKinsey & Company
Consent Decision-Making - Sociocracy 3.0
How to Make Great Decisions Quickly - Harvard Business Review
Documenting Architecture Decisions - Michael Nygard (Cognitect)
DACI: A Decision-Making Framework - Atlassian
Enjoyed this piece?
I write weekly articles for designers and design leaders who want to grow their impact, lead with clarity, and build careers that actually feel sustainable.
FAQs
What slows decision-making on cross-functional teams?
Unclear roles, no pre-read, and chasing consensus. Map RAPID roles, ship a one-pager pre-read, and use consent for reversible calls.
RAPID vs. RACI vs. DACI—what should designers use?
Use RAPID when “who decides?” is the bottleneck; use RACI/DACI when execution ownership is messy. Keep any model to one page so people use it.
What belongs in a decision log?
Context, options considered, the decision, the owner/date, and the next step — link to artifacts. ADRs are a solid reference pattern.
Is consent “lower quality” than consensus?
No. Consent allows movement when there are no material objections; you can revisit with data. Consensus demands full agreement, which is slower and rarely necessary.