Sufficiency
Jump to navigation
Jump to search
Sufficiency is a condition that occurs in the process of sense-making. Sufficiency occurs when a decision maker knows that she/he knows enough to decide or act. In ordered systems, establishing sufficiency is straightforward:
- In the clear domain sufficiency is established when you have all inputs required to apply the best practice. Rigid constraints inform which are the required inputs. For example, obtaining a drivers' license could be seen as a condition of sufficiency to know you can independently drive a car.
- In the complicated domain sufficiency is established when you collected and analyzed enough data to apply known good practices. Governing constraints describe the thresholds you need to meet to have enough information. For example, an auditor will sign of on the annual report of a company after it has audited them, confirming that report is accurate and the firm is a going concern. The auditor typically won't break down all financial statements, down to the lowest level and verify all of them. Instead the auditor performs spot checks on some accounts and might review samples of invoices, financial transactions, client orders, inventory. Past a certain threshold (the governing constraint) a reasonable certainty is established the company financial statements are accurate, and the auditor can act (sign off).
In non-ordered systems, sufficiency can't be established in a pre-determined way, it is an emergent property.
- In the complex domain all agents in the system are highly interconnected to such an extent that no qualitative predictive models can be applied. You can change some of the modulators of the system (enabling constraints), and attempt to detect patterns. When useful patterns emerge, you have sufficiency to act, to give more resources to the desirable patterns. In the complex domain you can't be confident the patterns will remain valid over time, so the sufficiency is only temporary in nature. For example, a political party might discover a pattern in social media to influence voters preferences and exploit that pattern to their benefit. But at some point, social media operators might react and shut things down. Or the pattern of influencing voters might be made public, influencing voters preferences in previously unexpected ways.