Home About How it works Services Blog Contact
PT EN
Request a diagnosis →
Decision Diagnosis

Quantify the risk of deciding without data.

Five questions, two outputs, every coefficient anchored in peer-reviewed research. An illustrative reference, not a forecast.

5 inputs ~60 seconds Open sources

Describe the decision

The results panel updates in real time as you fill it in.

01 · Size of the decision

The approximate financial weight of the decision (the capital, revenue, or cost at stake).

02 · Reversibility

A two-way door (you can go back in weeks) or a one-way door (hard to undo)? Bezos framework.

03 · Data maturity

How the organization usually gets the information that shapes the decision.

04 · Time pressure

How much time you have before the decision needs to be made.

05 · Stakeholders

How many people need to align on the final decision. The literature shows a sweet spot: 2–5 deciders tend to outperform both the solo decider (single-person bias) and the large committee (coordination cost).

How we calculate

The model

Each coefficient below is pulled from a peer-reviewed source or a published sector study. The goal is a defensible range, not an exact forecast.

Quality = 37% (baseline) + adjustments for data maturity, time pressure, and stakeholders. Capped between 5% and 75%.

Exposure range = size of the decision × (1 − quality) × [5% .. 25%], multiplied by 1.5 for one-way decisions.

Sources

  1. McKinsey & Company, "Decision Making in the Age of Urgency", 2019: 37% of senior executives surveyed rated their typical decisions as high quality. This sets the baseline of our estimate.
  2. Brynjolfsson, McAfee & Hitt, MIT Sloan / HBR, "Strength in Numbers: How Does Data-Driven Decisionmaking Affect Firm Performance?", 2011: data-driven companies are 5 to 6% more productive and profitable than peers. We conservatively translate this into +12pp of quality for systematic data maturity.
  3. PMI, "Pulse of the Profession" (2020): organizations waste ~9.9% of every dollar invested to poor project performance. This grounds the lower bound of our exposure range (5%, conservative).
  4. Jeff Bezos, 1997 shareholder letter and subsequent writings: the "one-way doors vs. two-way doors" framework. Irreversible decisions carry a 1.5× multiplier on the downside in our model.
  5. Bain & Company, Blenko, Mankins, Rogers, "Decide & Deliver" (2010) + Surowiecki, "The Wisdom of Crowds" (2004) + Kahneman's "decision hygiene": decision quality is non-monotonic with respect to group size. A small group (2 to 5) outperforms both the solo decider (single-point bias) and the large committee (coordination cost, groupthink). We apply +3pp for groups of 2 to 5, −6pp for 6+, and 0pp for solo.
  6. Kahneman & Tversky, decades of literature on biases (synthesized in *Thinking, Fast and Slow*, 2011): time pressure amplifies overconfidence and anchoring. Urgent decision → −10pp of quality.
  7. IBM, "Cost of a Data Breach Report 2024": the global average cost of a breach is US$ 4.88M. A contextual reference when the decision involves newly exposed surfaces.
What this is not

This diagnosis is an illustrative reference based on external benchmarks. It is not a forecast, not a valuation, not a consulting opinion. Real results depend on your data, your context, and your execution. For a decision that matters, run the real analysis, or schedule a closed-scope diagnosis with D.Lab.

Apply this rigor to your real decision

The diagnosis is public because the method is what matters. When you need the same rigor applied to your specific context, we have a closed-scope engagement for that.

Schedule a diagnosis → See the blog