DCT for the Boardroom

Ten questions for management

1. 1

   What state is the enterprise actually in?

   A good answer: A declared state with owners, sources and error bands, not a dashboard

   Enterprise state · Vol. I, Ch. 5

2. 2

   Is management's destination feasible, and can we stay there?

   A good answer: Kernel membership with its margin, not just feasibility today

   Feasible region; viability kernel · Vol. I, Ch. 14

3. 3

   Which constraints bind?

   A good answer: The binding constraints, each with its price

   Multipliers · Vol. II, Ch. 1–3

4. 4

   Which initiatives should come first?

   A good answer: A sequence justified by what must move before what

   Operator order; dependency graph · Vol. I, Ch. 6, 9

5. 5

   How much capital should we commit, and when?

   A good answer: A package priced against the board's appetite

   Capital architecture; tranches; committed delay · Vol. II, Ch. 1

6. 6

   What risks does the path itself create?

   A good answer: The chance of breaching a floor at any point on the path, not only at the end

   Path-dependent risk; floors · Vol. I, Ch. 8, 12

7. 7

   What assumptions would invalidate the plan?

   A good answer: The estimates the decision rests on, and the margin against each

   The sensitivity statement in the certificate chain · Vol. I, Ch. 14

8. 8

   When should management re-plan?

   A good answer: Named triggers, set before the plan is approved

   Triggers declared in advance · Vol. II, Ch. 16

9. 9

   What should the board monitor?

   A good answer: Distance to the floor, stability of the feedback core, value gap, model drift

   Certificate conditions · Vol. II, Ch. 15

10. 10

    Where may AI decide?

    A good answer: A state-dependent envelope with clear escalation

    The autonomy envelope · Vol. II, Ch. 13–15

Next: The certified decision →

The certified decision

The board receives three things, together or not at all: the policy; the certificate chain that establishes it is admissible and stable; and the sensitivity statement that says which estimates it depends on.

“a recommendation without it is an opinion with better notation”

Read it on the Map

Next: From → to →

From → to

• One enterprise-wide program that changes everything at once→Departments stepping on their own clocks, orchestrated around shared platforms
• An end-state and a finish date→A loop of step-changes that keeps climbing; the loop never closes
• A business case that is a single number→Plans as ranges, with a floor the board agrees in advance
• Readiness judged by the average score→Readiness capped by the weakest critical capability
• Benefits promised at launch→Benefits banked gate by gate, and counted only when they show in the accounts
• Technology as the transformation→Technology as a shared rung that serves every department

DCT Helix white paper, page 1 (verbatim)

Next: Steering committee →

Questions for the steering committee

1. Could each department’s next step run on its own clock, or does the plan force them all through one go-live?
2. Does the plan say what happens after the target is reached, and what will hold the gains?
3. Is the technology sequenced to the processes and data it must serve, or the reverse?
4. Has the board approved a range and a floor, or a single number?
5. What is the weakest critical capability, and does the plan build it before relying on it?
6. How much change can the people at the center absorb this quarter, and who is counting?
7. Which benefits are already visible in the accounts, and who reconciles them?

DCT Helix white paper, page 10 (verbatim)

Next: Question bank →

Board questions from 16 case studies

These organizations did not use DCT; each case reads a published decision in DCT's terms.

• Is every input of the model—here, distances, times, and promised windows—a declared quantity someone owns (Theorem 1.5)?

  UPS · Vol. II, Ch. 1

• Which business rules are written in as commitments, and which did the model leave out?

  UPS · Vol. II, Ch. 1

• Does the answer come with a certificate (Definition 1.7), and if not, how is its gain measured?

  UPS · Vol. II, Ch. 1

• Who explains the model to the people who must follow it, and may they override it?

  UPS · Vol. II, Ch. 1

• Is each commitment declared with its sense—kept on every path, or allowed to break at a price—and who set the price (Proposition 2.3)?

  Kellogg · Vol. II, Ch. 2

• When the plan breaks a priced commitment, who hears about it, and what does the price say about where to invest?

  Kellogg · Vol. II, Ch. 2

• Which data enter the model, who owns each, and how are they checked before each run?

  Kellogg · Vol. II, Ch. 2

• What does the model leave to people, and how do their changes come back into it?

  Kellogg · Vol. II, Ch. 2

• Which of our internal prices come from an optimization, and could someone outside the team check them (Theorem 3.4)?

  MISO · Vol. II, Ch. 3

• What does one more unit of each binding limit—a line, a budget, a floor—save, and who is told (Theorem 3.5)?

  MISO · Vol. II, Ch. 3

• Where our decisions are lumpy, how far from the best is the plan we accept, and is that gap reported (Proposition 3.2)?

  MISO · Vol. II, Ch. 3

• In the benefits we claim, which part did the model produce, and which came from other changes?

  MISO · Vol. II, Ch. 3

• Is our answer certified by a gap, or is it only the best of the plans we happened to find (Theorem 4.4)?

  NFL · Vol. II, Ch. 4

• Which rules are hard and which are preferences, and who decides when they conflict?

  NFL · Vol. II, Ch. 4

• If we fix part of the answer to make the rest solvable, what might the fixed part cost us?

  NFL · Vol. II, Ch. 4

• How will we know whether the chosen plan delivered what its score promised?

  NFL · Vol. II, Ch. 4

• When new data arrive, do we plan again from where we stand and commit only to the next step, or follow a schedule set months ago (Theorem 5.3)?

  Zara · Vol. II, Ch. 5

• What does our plan say whatever is left at its end is worth, and who checked that figure (Theorem 5.2)?

  Zara · Vol. II, Ch. 5

• Before we adopt a new decision method, how will we compare it with the current one, and what is the control?

  Zara · Vol. II, Ch. 5

• When the model and the managers disagree, who decides, and what evidence settles it?

  Zara · Vol. II, Ch. 5

• Which of our stocks—skills, customer goodwill, a brand, a natural resource—are we drawing down faster than they renew, and what is a unit of each worth left in place (Theorem 6.1)?

  Northern Prawn Fishery · Vol. II, Ch. 6

• Do we manage those stocks for the largest output or for the largest present value, and where do the two part?

  Northern Prawn Fishery · Vol. II, Ch. 6

• What floor will we never go below, and what signal will tell us to stop before we reach it (Proposition 6.3)?

  Northern Prawn Fishery · Vol. II, Ch. 6

• Where a stock is too variable to set a quota, what can we control instead?

  Northern Prawn Fishery · Vol. II, Ch. 6

• What is a unit of our capacity worth at each position, and does every decision to sell or hold it use that value (Proposition 7.1)?

  American Airlines · Vol. II, Ch. 7

• Do our limits respond to demand as it arrives, and what is that response worth against a plan fixed in advance (Proposition 7.2(ii))?

  American Airlines · Vol. II, Ch. 7

• Where a large problem has been split into parts, who checks that their answers fit together?

  American Airlines · Vol. II, Ch. 7

• Do customers understand the rules our prices impose, and what would complexity cost us in their trust?

  American Airlines · Vol. II, Ch. 7

• When we must exit a position—a business, a market, a hedge—what does speed cost us, and what does waiting risk (Theorem 8.3)?

  Institutional brokers · Vol. II, Ch. 8

• Have we measured those costs, or only assumed their shape?

  Institutional brokers · Vol. II, Ch. 8

• Does our schedule respond to what we learn while we carry it out?

  Institutional brokers · Vol. II, Ch. 8

• Who decides how much risk we will bear to save cost, and is that decision written down?

  Institutional brokers · Vol. II, Ch. 8

• For each stock we draw on—cash, capacity, inventory—what is a unit worth kept for later, and do today's decisions use that value?

  ONS · CCEE · Vol. II, Ch. 9

• Does our appetite count how deep a breach would go, or only how often it happens (Proposition 9.2)?

  ONS · CCEE · Vol. II, Ch. 9

• Who chooses how much weight the worst scenarios get, and who pays for that choice?

  ONS · CCEE · Vol. II, Ch. 9

• If the model is sound but the investments it assumes arrive late, who will tell us, and when?

  ONS · CCEE · Vol. II, Ch. 9

• Which suppliers, systems, or sites would stop us longest if they failed, whatever the cause (Definition 10.1)?

  Ford · Vol. II, Ch. 10

• Is our risk effort allocated by exposure or by spend?

  Ford · Vol. II, Ch. 10

• For each critical node, how long can we run without it, and how sure are we of the time it would take to recover?

  Ford · Vol. II, Ch. 10

• What set of events does our guarantee cover, and who audited it (Proposition 10.1)?

  Ford · Vol. II, Ch. 10

• Which law are we planning on, and what other laws would our records allow (Proposition 11.2)?

  Reclamation · Basin States · Vol. II, Ch. 11

• Have we judged our plan under every law in that family, or only under the one we declared (Theorem 11.1)?

  Reclamation · Basin States · Vol. II, Ch. 11

• What does our doubt cost: how far does the bill move from the kindest law to the harshest?

  Reclamation · Basin States · Vol. II, Ch. 11

• Which conditions—not which probabilities—would break our plan, and how would we know that they had arrived?

  Reclamation · Basin States · Vol. II, Ch. 11

• Which of our objectives conflict, and have we seen candidates that trade them, each scored on all of them (Theorem 12.1)?

  NS · Vol. II, Ch. 12

• Whose preference selects among them, and is it written down (Proposition 12.1)?

  NS · Vol. II, Ch. 12

• What do our weights stand for, and what would a different weighting choose (Theorem 12.2)?

  NS · Vol. II, Ch. 12

• When the losers of a trade-off object, who decides whether to move along the frontier, and at what price?

  NS · Vol. II, Ch. 12

• On what data, untouched by those who built it, has our model been judged (Theorem 13.3)?

  Netflix · Vol. II, Ch. 13

• Which decision does the model change, and what is the change worth, net of the cost of running it (Theorem 13.2)?

  Netflix · Vol. II, Ch. 13

• If our objective has moved since the model was specified, does its metric still measure value?

  Netflix · Vol. II, Ch. 13

• What could our data reveal about individuals, and who has priced that risk before we share it?

  Netflix · Vol. II, Ch. 13

• Which actions may the system never take, who declared them, and is every action checked against them before it runs (Definition 14.2)?

  Google · Vol. II, Ch. 14

• What judges its proposals before they act—a declared model, a simulator with a known error, or fresh trials—and do the records cover the actions it proposes (Theorem 14.3)?

  Google · Vol. II, Ch. 14

• Against what baseline are its savings measured, and were they tested by switching it on and off?

  Google · Vol. II, Ch. 14

• Who can switch it off, and on what signal does control return to people?

  Google · Vol. II, Ch. 14

• What do we read about our critical assets, how often, and how soon would we see a change (Theorem 15.3)?

  Rolls-Royce · Vol. II, Ch. 15

• Which decisions may our systems take alone, and which must they refer to a named owner (Definition 15.3)?

  Rolls-Royce · Vol. II, Ch. 15

• If we sell outcomes rather than products, which risks have we taken on, and are our contracts priced for them?

  Rolls-Royce · Vol. II, Ch. 15

• Who checks the model behind our monitoring when the world changes?

  Rolls-Royce · Vol. II, Ch. 15

• Which stocks do our business units share—cash, capacity, credit lines, reputation—and what does each unit pay for what it uses (Theorem 16.1)?

  Large banks · Vol. II, Ch. 16

• Is that price the marginal value of the stock to the whole firm, or a convenient average (Proposition 16.3)?

  Large banks · Vol. II, Ch. 16

• Who owns the shared buffer, and is its cost charged to the units that create the need for it?

  Large banks · Vol. II, Ch. 16

• Where could a unit get around the price, and what limit would stop it?

  Large banks · Vol. II, Ch. 16