Design prescription binding protocol

Description

A protocol that binds each design recommendation to its evidence, expected effect, success metric and validation plan, so prescriptions are accountable and testable.

Research recommendations often lose their link to evidence once handed over, and design changes ship without any check that they worked. The binding protocol formalises each prescription as a record: the problem, the evidence (with source and confidence), the specific design change, the mechanism expected, the metric that should move and by how much, guardrails, the validation method and the owner. Borrowing from evidence-based medicine's link between diagnosis, treatment and outcome measures, and from requirements traceability, it makes prescriptions falsifiable. Rigour comes from requiring every field before acceptance, quantified expected effects, pre-registered validation, and closing the loop by recording observed outcomes.

Attributes

Loss type
Community-proposed
Effort
Hours

Steps

  1. For each recommendation, record the problem statement and the evidence supporting it, with sources and confidence.
  2. Specify the design prescription precisely enough to be implemented and tested.
  3. State the expected mechanism and a quantified expected effect on a named metric (e.g. +3 points completion within 30 days).
  4. Define guardrail metrics and the validation method (A/B test, quasi-experiment, follow-up usability study).
  5. Assign an owner and review date, and have product and design leads sign off the binding.
  6. After launch, record observed results against the expected effect and update the evidence base.

Sources

  • Storytelling with Data (Cole Nussbaumer Knaflic)
  • Articulating Design Decisions (Tom Greever)
  • Envisioning Information (Edward Tufte)
  • Storytelling with Data (Knaflic)
  • Articulating Design Decisions (Greever)
  • Envisioning Information (Tufte)
  • The Inquiry Series, Ch 59 — Research Communication — From Evidence to Strategic Brief

Also known as

  • Evidence-to-design binding
  • Prescription traceability

Categories

  • Community-Proposed Frameworks