The discipline
Built over thirty years of practice, and every idea in it implemented and running. What an organization values, what it is offered, what it can spend, what it chooses and who does the work form one system with economic roots. Frontnovo treats that system as a discipline: five decision objects that move through plan, manage and learn, worked on by five disciplines: economics, mathematics, behavior, governance and law, and automation. Each is assigned its part and wired to the others, and the power is in the joins. The IQ Suite is that discipline built into software.
Five ideas
01
Portfolio tools solve only the first half: which investments use scarce cash and people. The discipline also solves the second: how much capacity to produce, now and over years. Prices link them. Capacity is worth building until what a resource is worth meets what it costs to make.
02
Every tool optimizes over what it was handed. Here the givens move: a better pipeline adds candidates worth more, and analysis and negotiation move the limits. But which limit is worth moving? Inside a firm there is a market: teams compete for capacity, budget lines and mandates every cycle. But it trades in influence, not prices, so its prices stay invisible. The model makes them visible. It computes what each limit costs: its opportunity cost, what the portfolio gives up because the limit is there. It names the change in a number that would alter a decision, and asks that number's owner to agree.
03
We don't ask how uncertain you are. We show how uncertain you can afford to be.
Every decision carries a tolerance as well as a value, computed from the structure of the problem, with nothing to estimate. Wide margins run on autopilot; thin ones go to people, who answer the one question they are good at: could our numbers be that far off?
04
Ranking is greedy across projects; the annual portfolio is greedy across years. Decided one year at a time, a firm under-builds and under-maintains with a clean conscience. Planned as a path, with capacity as the means, it falls behind for two years and pulls ahead from the third.
05
Two things are set for every decision. Who decides: people decide, review or confirm. And what AI does for them: it draws out what they know, explains what the optimizer proposes, and records why. The margin from idea 03 sets both.
“A wealth of information creates a poverty of attention.”
Herbert Simon, 1971
Decide, review, confirm.
Elicit, explain, record.
The margin decides which.
Each does what it is best at. The optimizer computes the best answer and how far it is from flipping. AI curates: it elicits judgments people hold but have not written down, explains results in plain words, and keeps the record; it never optimizes and never commits. People commit every decision, with the attention its margin requires. Most tools give AI one role. Here its role follows the decision.
Most organizations set the split first, so much to run and so much to transform, and then fill the boxes. Here every investment is tagged once, as maintain, deliver, build or retire, and the split comes out of the optimization. Where leadership wants a floor, such as build at least 20%, it enters as a limit, and its price is shown: what the floor costs in value elsewhere. Three Horizons and pace layers name other dimensions, an investment's timing and its asset's pace, and become views of the same model rather than competing exercises.
Figures in the five ideas are illustrative.
The subject
Strategy, pipeline, budget, portfolio and resources are each decided, and each moves through three states. A plan is committed, managed against results, and learned from. Learning returns twice: to managing, to correct the course within a cycle, and to planning, to start the next cycle from better ground. Under uncertainty every state carries a tolerance as well as a value, and learning is what narrows it.
RootsWeingartner · Markowitz · Bower
Discipline 1 · The foundation
Every object has an economic principle under it. Strategy is preference, the budget is scarcity, the portfolio is allocation at the margin, and resources are production. The pipeline is the new piece. Each proposal is a demand for capital, and read from the other side the same proposals are the organization's supply of opportunities. Economics has always taken that supply as given. Here it is produced, priced and managed. Every price carries its tolerance too: how far it can move before the choice it supports changes.
RootsDean · Hirshleifer · Ramsey · Keynes
Discipline 2
Economics says what a price means; mathematics computes it. Projects are chosen whole, so the problem is an integer one, and every limit still carries a price, which in whole choices moves in steps rather than smoothly. The same mathematics runs backwards: from the decision to the variation it can absorb, and from the result to how far a coefficient would have to move to change it.
RootsKantorovich · Dantzig · Gomory · Ben-Haim
Discipline 3
People decide with bounded attention, anchors and optimism, and they satisfice where theory says optimize. Proposals arrive inflated, weights are asserted rather than measured, and budget splits follow mental accounts. The discipline expects this and designs around it instead of assuming it away. It never asks people to estimate a probability, only whether a number could be off by as much as the decision can tolerate.
RootsSimon · Cyert & March · Kahneman & Lovallo · Thaler
Discipline 4
Rules, mandates, floors and authority shape every decision, and each one has a cost. Some are law or regulation, some are policy, some are habit. The discipline makes those costs visible, so an organization can tell a rule that protects it from one it only obeys. Every commitment leaves a record of who decided, on what, and with what margin, which is what a board or an auditor asks for. Review goes where tolerance is thin, not evenly across every project.
RootsBerle & Means · Jensen & Meckling · Stein · Scharfstein & Stein
Discipline 5
Three layers. Computation: the optimizer finds the best answer for what is given, with its prices and margins, exactly and repeatably. Curation: AI draws out what people know, explains the result, drafts the ask and flags what looks wrong, which is useful and fallible. And the allocation of the decision itself: who makes each one, at which level, from manual through supervised to autopilot.
That last layer is an economic question, so the first discipline governs the fifth: each decision goes to whoever has the advantage in it, measured by its margin, as idea 05 shows. AI may propose and explain, the optimizer computes, and only people commit.
RootsRicardo · Sheridan & Verplank · Parasuraman, Sheridan & Wickens
Why it is different
Portfolio management grew out of project management. It kept the vocabulary of investment and left the mathematics behind, so its tools score the candidates, sort them, and fund down the list until the money runs out. A portfolio is not chosen in isolation. It is chosen against a budget, against scarce skills and against itself, and the IQ Suite treats it as that one decision.
None of this starts from nothing. The economics of capital allocation was worked out decades ago; portfolio management never adopted it. Each element is marked as inherited, adapted, or what we believe is new.
Every binding constraint has a price, read at the margin.
The firm as activities competing for scarce factors, each factor priced by imputation. The anchor of the discipline.
Selection under a budget is one decision, not a ranking.
A portfolio is a composition chosen against an objective, not a list of good items.
Decide under friction. The optimum is a proposal for people to work with, not a verdict.
A firm's projects ranked by return form its demand for capital. Read from the other side, they are its supply of opportunities.
Proposals take shape bottom-up, long before the top sees them.
Divisions compete for funds and inflate their asks. The behavior is part of the system.
How much of today to spend on tomorrow. The question behind planning a portfolio as a path, not a year.
Some uncertainty cannot be given a probability. Decide in a way that does not depend on one.
From full human control to full machine control. Applied here decision by decision, with one test for where each decision belongs.
Ask how wrong the inputs can be before the decision fails. Applied here to every decision in the portfolio, and used to route each one to the machine or to people.
Priorities become an exchange rate between drivers, elicited with few comparisons and published with a range.
Applied to the pipeline: find what limits the whole before optimizing a part.
Design the options instead of only choosing among the ones offered.
Budget splits are set by people, so they are priced rather than obeyed.
What a stronger pipeline would buy at the same budget, and which priorities have no candidates at all.
The boundary ordered by how far the world must move before each decision changes.
Every decision in an internal investment system, typed, owned, and placed on one test for automation.
One engine for planning and custody, so there are never two answers to one question.
A method that only a specialist can run changes nothing. The IQ Suite keeps the economics underneath and shows the person questions, answers and prices in the units the decision is made in. Four design laws hold across every module.
A screen belongs either to the person or to the machine. The person always knows whether they are looking at their own judgment or at a computation.
The interface asks the question an executive would ask, such as "What does more money buy?", and answers in dollars, people and priorities rather than scores.
Every module opens on a live map of where the work stands, with one recommended next step, then carries the person through the decision.
Autopilot prepares the decision, Supervised proposes, and Manual shows everything. In autopilot the reading disappears, never the choosing.