Decision Rights Matrix: Who Decides What in a Small Company
A decision rights matrix is a table that lists your company's recurring decisions and names, for each one, who decides, who must be consulted and who only needs to be told. Its point is to end two failures: decisions that stall because nobody owns them, and decisions made twice because two people each thought they had the final word.
You don't need a consultant or a heavy framework to make one. You need a list of the decisions that actually cause friction, one name per decision and a rule for how long a decision can sit. Here's how to build the matrix and keep it from becoming shelfware.
Vendors Covered in this Article
Disclosure: We may earn a commission if you buy through some links on this page. It doesn't change what we recommend.
Which decisions belong in the matrix?
Start with friction, not completeness. Ask your leadership team to write down the last ten decisions that took too long, got reversed or caused an argument about whose call it was. Group them into a short list of recurring decision types.
Typical entries for a small company:
- Hiring and firing: opening a role, choosing a finalist, setting an offer.
- Spending: approving a new vendor, a purchase above a set amount, a budget change.
- Pricing and discounts: changing list price, approving a discount outside the normal range.
- Customers: accepting a nonstandard contract, giving a refund or credit, firing a customer.
- Product and delivery: changing scope, delaying a launch, choosing a technology.
- Policy: changing a company rule or a benefit.
Aim for fifteen to twenty rows. If you list a hundred, nobody will read it. Anything rare can be handled by asking.
How do you assign roles for each decision?
Use a simple set of roles, and insist on exactly one person in the deciding role. If two people share the decision, then neither owns it.
- Decider: the single person who makes the call and answers for it.
- Recommender: the person who does the analysis and proposes an answer, often the one closest to the work.
- Consulted: people who must be asked before the decision, because they have information or will be affected. Their advice is heard, but they don't hold a veto.
- Informed: people told afterward, so they aren't surprised.
If you've used a RACI chart, this is similar but tighter: the accountable person and the person who does the work are separated from the person who decides, and there's a strict limit of one decider. The RACI chart template shows the difference in more detail. Bain's RAPID model (Recommend, Agree, Perform, Input, Decide) is another variant, and it adds an "Agree" role for people with a real veto, which you should use sparingly.
Worked example: an operations team with four recurring decisions
Suppose your operations lead, finance lead, sales lead and CEO keep colliding over the same issues. You put four rows in the matrix.
For a new vendor under a spend limit the operations lead sets, say $5,000 a year, the operations lead decides, finance is consulted and the CEO is informed. Above that limit, finance decides after operations recommends, and the CEO is informed. For a customer discount beyond the standard range, the sales lead recommends, finance is consulted on margin and the CEO decides. For a delivery date change that affects a customer commitment, operations decides and sales is informed.
Notice what the tiers do. Tying the decider to a spend or risk threshold means most decisions are made at the lowest level that can bear the consequences, and only the unusual ones climb.
How long may a decision sit before it escalates?
A matrix without a clock still stalls. Attach a time-box to each decision type: for example, routine spend gets two business days, a customer exception gets one day and a hiring offer gets three. If the decider hasn't decided when the time is up, the decision goes to the next person up the chain, who either decides or gives the original decider a firm extension.
Write the rule where everyone can see it, and apply it to yourself too. Founders are the most common bottleneck, so consider a rule that a decision waiting on the CEO for more than a set number of days is approved by default when the risk is low and the recommender has documented the reasoning. That transfers responsibility to where the information is.
Time-boxes also make disagreement cheaper: people can commit and revisit if the facts change, instead of debating indefinitely.
How do you publish the matrix and keep it current?
Publish it in a place people already use: your wiki, a shared doc or a work management tool, and link to it from the channels where decisions get made. Pair it with a simple decision log, one line per significant decision with the date, the decider and the reasoning, so people can see precedent instead of relitigating.
Review the matrix quarterly. Ask three questions: which decisions caused conflict this quarter, which rows nobody used and which thresholds have become too low or too high as the company grew. Update it, date the change and tell people. Notion is a common place to hold the matrix and the log, and Asana can carry approval tasks that route to the right decider. For choosing a wiki, Notion vs Slite vs Confluence is a good comparison, and Asana vs Monday vs ClickUp covers the task side.
What Good Looks Like
A good decision rights matrix names exactly one decider for each recurring decision, sets thresholds and time limits, and lives where people make decisions instead of in a forgotten slide.
Building The Capability (5-Stage Skill Ladder)
How to Get Started
Disclosure: We may earn a commission if you buy through some links on this page. It doesn't change what we recommend.
Frequently Asked Questions
What is a decision rights matrix?
It's a table listing recurring business decisions and assigning each person a role: who decides, who recommends, who is consulted and who is informed. It gives every decision one owner and makes it clear who can act without asking permission.
What is the difference between a RACI chart and a decision rights matrix?
RACI assigns roles for tasks and deliverables, while a decision rights matrix assigns roles for decisions. The key difference is that a decision matrix names exactly one decider, so a call doesn't stall between several people who are each partly accountable.
How many decisions should the matrix cover?
Start with fifteen to twenty recurring decisions that cause the most friction. A longer list becomes hard to read and won't be used. Add rows only when a new type of decision repeatedly stalls or causes conflict, and remove rows nobody references.
About the numbers
This guide doesn't quote a sourced benchmark. Figures in it are estimates or general guidance, so check them against your own numbers.
Related Guides
How to Build a RACI Chart for a Small Business
Build a RACI chart in five steps, with worked examples, the one-accountable-person rule, and the mistakes that make responsibility charts useless.
Notion vs Slite vs Confluence: Company Wiki Comparison
Compare Notion, Slite, and Confluence for company knowledge bases and asynchronous team wikis. Evaluate search, document structure, and AI search tools.
Asana vs Monday.com vs ClickUp: Best Project Management Software
Compare Asana, Monday.com, and ClickUp for operational project management: cross-functional dependencies, workload planning, and tool fatigue.
Outsourcing vs Hiring In-House: How to Decide
A decision framework for choosing between outsourcing an operational function and building it internally, based on strategic value and control needs.
A Five-Stage Hiring Process Small Businesses Can Actually Run
Run hiring in five stages with an owner, output and time limit for each, so roles fill faster and every candidate is judged against the same standard.
Leadership Meeting Notes: A Format That Records Decisions
A leadership meeting notes format built around decisions, owners and dates, with a filled-in example, a scribe routine and tips for keeping notes findable.