The Analytics Solution

For decades, business cases estimate one half of a decision with precision — the cost side is quantified, modeled, and defended — while the benefit side gets summarized in a sentence or reduced to a vague multiplier. The reason is structural: there was no standard method for putting a judgment call in the same place as a dollar figure, on the same scale, with the same ability to be challenged.

The previous articles in this series made the invisible elements of a decision visible: a shared outcome, a structured set of options, and a ranked list of priorities — debated, mapped, and owned by the people who will live with the decision.

The Analytics Problem identified four ways the signal fails to enter the analysis:

1. live signals from the people closest to the work with no way in;
2. qualitative signals ignored because they don't fit the mold of a "metric";
3. logic that wins by being opaque rather than right;
4. numbers that don't connect to one another, so the linkages stay invisible.

This article builds the process — the A in OPERA, Analytics — that gets each of them scored, on a common scale, in a single grid.

The Force Multiplier No Dashboard Measures

Sol Rashidi — the world's first Chief AI Officer — asserts that organizations have spent decades measuring efficiency and productivity, while effectiveness — how well human judgment actually amplifies outcomes — has never had a measurement framework, let alone a place in the model. As Rashidi puts it: "no algorithm can beat the collective strength of your workforce, your people, and their effectiveness in creating your greatest force multipliers."

The OPERA Analytics process places human judgment alongside quantitative data — scored, scaled, and seated in the same model — so it stops being an invisible input and starts becoming a force multiplier: visible, challengeable, and connected to the outcome.

Human voices don't just participate. They compete with the numbers. And have the power to change the outcome.

Decision Context

This article demonstrates the process through a previous project with the Navy to prioritize energy-efficiency investments for their data centers. The analytics process is grounded in Multi-Criteria Decision Analysis and the Analytic Hierarchy Process, adding what both omit in practice: a structured protocol for capturing qualitative judgment and live stakeholder signals directly — without converting them through mathematics that only an analyst can follow.

The real engagement data is proprietary, so every score, cost, and quantity that follows is representative — chosen to illustrate how OPERA's analytics process works, not to reflect the actual values modeled for the Navy. The mechanics are faithful; the numbers are not.

The OPERA outcome structure — verb + deliverable + impact — produced this outcome:

In Defining the Decision Space, we showed how a well-defined outcome can be broken down into candidate paths to achieve it.

The outcome fans out into four meaningfully different paths forward — each becomes a column in the modeling grid.

At this point in the OPERA process, the team has a defined shared outcome, a set of options, and ranked priorities. Each priority gets modeled, becoming an analytic — a number or score per option that can be debated, challenged, and connected to the rest of the decision.

That transformation — priority in, analytic out — is what this process produces.

The Analytics Process

From stakeholder priority to modeled analytic — each step narrows the scope, builds the model, and connects it to the decision.

Step 1: Scope

Each priority × option cell is an analytic to build. Priorities from the priorities process form the rows; options form the columns.

Every priority × option cell is an analytic to build.

Twenty cells is more than most teams can model well. Two successive filters narrow the grid to what's worth modeling — first the priorities, then the options.

Filter the priorities. Borrowing the same consensus-ranking mechanism from the priorities process, each stakeholder independently rates every priority on a common scale, and the team aggregates into a single ranked view. The priorities that cluster at the top carry genuine cross-stakeholder support; the ones near the bottom — championed by a single voice, or rated consistently low — fall below the line.

Stakeholders independently rate each priority, then the team aggregates. The top-ranked priorities advance to modeling; the rest drop below the line. Representative numbers, not from the actual engagement.

For the Navy decision, ROI, Carbon, and Energy Independence advance. Timeline and Community & Tenant Relations sit below — the team rated them lower.

Filter the options. Score each surviving priority against every option. Whichever option has the weakest column average — here, the Status Quo the outcome was designed to move away from — drops out. Status Quo wins one row (no capital outlay = best "return" in rough consensus) but loses the other two badly enough that its average falls below the cut. Winning a single priority isn't enough when the decision is weighted across the portfolio. The scoring is rough consensus, not modeled analytics; that's what the remaining five steps will produce.

Scoring the surviving priorities against the full option set reveals which column won’t earn its modeling cost. The struck-through column drops out, narrowing the grid from 20 initial cells to 9 analytics worth modeling. Representative numbers, not from the actual engagement.

The scoped grid that emerges — 3 priorities × 3 options = 9 cells — is the modeling assignment sheet.

Filter scores aren't analytic scores — and stopping here, treating the ranked grid as the decision, is the most common mistake. It skips the modeling, the debate, and the transparency that make a decision defensible.

The ranked grid tells you what to model. The next five steps build the analytics that inform the actual decision.

Step 2: Classify

For each priority in the scoped grid, determine whether it is qualitative or quantitative.

The priority determines the question, the classification determines the method.

• Qualitative: priorities where the answer is a judgment call — "how independent is the base from the grid under this option?" or "how resilient is our energy posture during a multi-day outage?"
• Quantitative: priorities where the answer is a number — dollars of capital and operating cost, tons of CO₂ avoided per year, months to commission

Step 3: Identify

For each priority, ask: where does the knowledge live?

Step 2's classification forks the priority. Step 3 sub-forks the qualitative side by where the knowledge lives. All three terminal paths produce a score per option.

The expert, the survey respondents, and the model builder are the live signals — the people closest to the work now have a structured entry point into the grid. This is how OPERA closes the first gap from the Analytics Problem article: live invisible signals with no way in from those closest to the information.

Step 4: Model

This is where the analytic gets built — using the method identified in Step 3 to produce a single score per priority × option cell, on a consistent scale, ready for the grid.

Qualitative: What Makes the Scores Trustworthy

Filter scores and analytic scores are not the same thing. Step 1 used rough consensus to triage — judgment calls about what's worth modeling. This step produces the actual analytic: the number that enters the filled grid, gets debated, and informs the decision. Same 1–5 scale, different jobs, different rigor.

The method was already chosen in Step 3 — an expert conversation, a live consensus walk, or a distributed survey. What separates these from the surveys the Priorities Solution article warned about? Every question is derived from a named priority, not invented by whoever built the form. And every score enters the same grid as the quantitative analytics — same grid, same cell, same debate.

Maximize Energy Independence, scored:

Scored by the energy officer (the expert identified in Step 3) against a fixed rubric: 5 = full islanded operation, 1 = unchanged grid dependence.

Quantitative: Rapid Modeling

Same priorities, three options, three different models — cost and schedule modeled per option. Representative numbers, not from the actual engagement.

This is the part most teams already know how to do. Your finance team can build a cost model. Your PMO can build a schedule. The modeling skill is not the bottleneck — the bottleneck is giving each model a defined place in the decision.

Two quantitative priorities appear in nearly every decision workspace: cost and schedule. Cost breaks down into labor, capex, and opex. Schedule breaks down into phases, dependencies, and milestones. The specifics vary by decision, but the structure is the same.

For the Navy walkthrough we size every option at 1 MW of installed capacity so the three technologies can be compared on a consistent basis. Rapid modeling means decomposing each priority into its components with transparent arithmetic. Rather than an aggregate estimate — "Solar costs ~$2.4M" — break it down so every input is visible and every assumption is named:

Cost decomposed into CapEx, Labor and O&M - every input visible, every assumption named. Representative numbers, not from the actual engagement.

Solar draws on the DOE's Q1-2025 Solar PV System Cost Benchmarks.

This raises an obvious question: isn't higher-fidelity modeling sometimes warranted? Rapid modeling comes first because you don't yet know where to invest time into the fidelity. Once the grid is filled and the exchanges surface the trade-offs, you can see which cells actually drive the decision. If the decision hinges on the cost difference between two options, that's where you build the detailed model — not everywhere.

The rapid pass is fast enough to fill the whole grid, legible enough to debate, and precise enough to reveal where additional detail matters.

Step 5: Debate

This is the same pattern from the priorities process: brainstorm, refine, debate. The priorities process didn't end debate — it structured it. The analytics process does the same thing: collect the signal, make it visible, then debate the result. Debate applies to both qualitative and quantitative analytics.

The same structured debate pattern runs through both processes — surface, structure, challenge.

Step 6: Illuminate

Once every priority has been modeled and debated, the grid that started empty in Step 1 is now filled. Numbers have been chosen intentionally to illustrate the process - not factual.

The same grid from Step 1 — now filled. Each cell is a modeled, debated score. The tensions between rows are the raw material for exchanges. Representative numbers, not from the actual engagement.

No single option wins every priority. Solar leads on ROI — lowest capital, fastest payback at approximately 10 years. Hydro leads on carbon avoidance and energy independence — always-on baseload, highest carbon offset, strongest islanding score. Wind trails on every priority and is eliminated.

The decision is now Solar versus Hydro, and it cannot be resolved by the analytics alone. It depends on how much the Navy values the large amount of carbon avoidance relative to return on investment. That weighting is what the Exchanges process resolves.

The Decision Structure

Each element in OPERA connects to every other and can be shown across stakeholders. The outcome defines which options are worth considering. Options are scored against priorities. Priorities are modeled into analytics. Analytics feed exchanges where the trade-offs become visible. Risk applies to certain options and priorities. The graph structure allows the decision team to see the full picture in one place.

The OPERA decision structure, displayed as a graph of connected decision elements — outcome, options, priorities, and analytics. Risk and exchanges are covered in the next articles. Any numbers shown in node tooltips are representative, not from the actual engagement.

What's Next

The Analytics Problem article opened with four ways the invisible signal fails to enter the analysis. The analytics process closes each gap:

Competing priorities are inherent to any real decision — as the Priorities Solution article acknowledged. The analytics process illuminates them. Once every cell is modeled, the grid makes them all visible.

The Analytics Process makes the invisible signal visible so that human judgment can be measured and challenged on the same grid as quantitative data. Next, we look at how defining risk as a discrete event reveals the invisible uncertainty that enables debate and directs leaders towards action.

To explore how the OPERA framework applies to a decision you are working through right now, reach out at hello@operascale.com.

About the Authors

Laxmi Gandhi

Founder & President

1.5K followers

Versatile senior management consultant with over 20 years of experience leading transformational decision analytics programs across multiple industries, spanning startup, growth, and Fortune 500 companies.

Mehul Gandhi

Chief Technology Officer

1.2K followers

Technology and management leader with extensive experience guiding cross-functional teams in developing AI-powered consumer electronics, fintech platforms, healthcare, and defense.