Decision-making insight

Big bets, cross-cutting calls and delegated decisions: matching the process to the decision

24 October 2026 · 5 min read · DecixionIQ

Senior leaders spend a large share of their time on decisions, yet many say those decisions are neither quick nor consistently good. McKinsey's research suggests one reason: organisations often run every decision through the same process. A better approach is to sort decisions by kind, and give each kind the process it deserves.

Three kinds of decision

McKinsey groups organisational decisions into three kinds, based on how often they arise and how much is at stake.

Kind	What it looks like	What helps
Big bets	Rare and high-risk, such as an acquisition. Taken by top leaders and the board.	Structured debate: someone argues for, someone against. Board-level authority.
Cross-cutting decisions	Frequent and high-risk, such as pricing. Taken jointly across business units.	A clear, repeatable process with defined roles and forums. Committee or executive authority.
Delegated decisions	Frequent and lower-risk. Taken by a person or team close to the work.	Clear delegated authority, good tools and guidance. Individual authority within limits.

The point is not the labels. It is that a pricing change made every month should not wait for the same deliberation as an acquisition, and an acquisition should not be decided with the speed of a routine purchase.

Mapping the kinds to governance levels

In a governed organisation, these kinds line up naturally with levels of authority. Big bets sit with the board or a board committee, cross-cutting decisions with executive forums, and delegated decisions with named individuals within limits. The exact thresholds — amounts, risk levels, reserved matters — are set by each organisation in its delegation of authority. They are examples to configure, not a universal standard.

Delegation only works when it is real. People need clear authority, the information and tools to decide well, and guidance on when to escalate. Without that, decisions drift upwards and leaders become the bottleneck.

Five biases worth challenging

Whatever the kind of decision, judgement is vulnerable to bias. McKinsey highlights five common ones. Each can be countered with a plain question, asked and answered on the record:

Bias	A question that tests it
Confirmation bias	What evidence would show we are wrong, and have we looked for it?
Herd mentality	Are we backing this because others are, rather than on our own evidence?
Sunk-cost fallacy	Would we start this today if we weren't already invested?
Ignoring unpleasant information	What bad news or warning signs have we set aside, and why?
Halo effect	Are we trusting this because of who proposed it rather than the case itself?

Writing the question down matters. A recorded challenge, with its answer, shows later that the decision-makers actively looked for the weak points — not just that they agreed.

Where DecixionIQ fits

DecixionIQ routes each decision to the authority your organisation's rules require, from an individual to the board, and records who voted and why. The Challenge stage includes an optional bias check with the five questions above; any of them can be added as a challenge question in one click and answered like any other. AI can help draft and challenge, but only authorised people approve.

For the wider picture, start the decision-making series.

Sources and further reading

General information, not legal or professional advice.