Three Rebuilds Show Operational Analytics Requires Management Logic, Not Just Data

After completing the third version of an operational analysis platform, the author contrasts the general manager's new reaction — asking for deeper margin analysis and inventory alerts — with earlier versions where flashy dashboards left leadership unable to answer why gross margin dropped two points.

Two Classic Pitfalls

Pitfall 1: Technical Self-Indulgence — Data Moving and Visualization

Many IT teams start by ingesting data sources, building KPI cards, and polishing large screens. They move numbers from Excel into BI, perfect color schemes and animations, and consider the project done. But the core pain point of enterprise operations is never "can't see the numbers"; it's "can't see the root cause." A pure-technology platform only shows a pretty result number without the structure, accountability, and action path behind it. When the boss drills one layer deeper, business teams spend days pulling data and rebuilding spreadsheets, turning the platform into a "looks good, works poorly" ornament.

Pitfall 2: Business Feeding — The Requirements Garbage Dump

Other teams over-respect business demands: sales wants customer reports, production wants capacity reports. The system accumulates hundreds of scattered reports with inconsistent definitions. In monthly operating meetings, sales claims 93% revenue achievement while finance says 91%, and half the meeting is spent reconciling numbers. Without a unified operating logic, all analysis stays at the department level — sales watches orders, production watches output, finance watches profit — and no one owns the overall operating result. The platform satisfies every department's request but cannot support company-level decisions.

Correct Approach: Management Logic as Skeleton, Data Capability as Flesh

The subject of an operational analysis platform is never "the platform" but "operations." First clarify and align the underlying operating logic, then use data capabilities to solidify it into a runnable mechanism.

First Wheel: Management Logic — The "Five Questions" Framework Unifies Corporate Operating Language

The team defined a core skeleton: Monthly Operating Five Questions — Revenue Gap, Gross Margin Decline, Cash Shortage, Delivery Instability, Organizational Delivery. This framework came from aligning top-tier enterprise review methodologies with the company's actual management actions and responsibility boundaries, solidifying a standard analysis path.

An "analysis path" means every question has a fixed decomposition: where to drill, how deep, and who is responsible — no vague language, no buck-passing.

Revenue Gap : Not just overall achievement rate; must decompose to orders, price, product mix, and recognition rhythm, down to specific customers and product lines.

Gross Margin Decline : Not just "cost increase"; must split price concessions, materials, labor, expenses, with every cent of variance mapped to a responsible department.

Cash Shortage : Not just "collection difficulty"; must calculate capital tied up in receivables, inventory, prepayments, and quantify losses at each link of the cash conversion cycle.

Delivery Instability : Not just on-time delivery rate; must attribute proportions to material shortages, plan changes, quality rework, and tackle the core contradiction.

Organizational Delivery : Not "team worked hard"; must verify completion rate of last month's committed actions and actual gains, using results to validate process.

The framework's greatest value is unifying the company's operating language. Previously meetings were cross-talk; now everyone follows the five questions: see result deviation, do structural decomposition, verify root cause, assign responsible actions. No more arguing "what numbers to look at"; only discussing "what problems to solve." Management logic injects soul into the system, turning scattered metrics into a decision-supporting operating system.

Second Wheel: Data Capability — Turning Management Logic from "Boss's Slogan" into "Organizational Mechanism"

Good management logic alone is insufficient. Previously the boss asked the same questions, but finance manually drilled in Excel for three days per deep dive; momentum faded and the logic remained the boss's personal habit, not an organizational capability. Data capability's core value is fixing the human-driven questioning logic into an automatic, continuously running mechanism.

Implementation spans the full operating meeting cycle:

1. Pre-Meeting: Auto Pre-Drill Anomalies — Turn "Reporting Meeting" into "Problem-Solving Meeting"

Before, finance, sales, and supply chain pulled all-nighters for three days pre-meeting, aligning calibers and building slides, with no energy left for deep analysis. Now the system locks data calibers at T-3, auto-identifies abnormal metrics per the Five Questions framework, and outputs a Top-3 problem list with alternative solutions. Which key customers churned, which products fell below margin red lines, which materials exceed age thresholds — the system computes and alerts in advance. The 90-minute meeting is spent entirely on decisions, zero minutes on number reconciliation.

2. In-Meeting: Layer-by-Layer Root Cause Drill-Down — Turn "Finger-Pointing Meeting" into "Accountability Meeting"

Before, a boss question like "why did gross margin drop" sent finance away for a week; the answer arrived at the next meeting, always one step behind. Now on the platform, clicking the gross margin indicator drills down: first split impact points of price concessions, product mix, materials, labor, manufacturing overhead; then click material cost to see exactly which raw material rose, which supplier, which batch. From operating result down to business documents, every layer has evidence, every item verifiable; root cause located on the spot, responsibility assigned on the spot. No more "roughly," "maybe," "basically" — every conclusion has data backing, every variance has an owner.

3. Post-Meeting: Closed-Loop Action Tracking — Turn "Talk Shop" into "Execution Meeting"

The critical step is sinking meeting resolutions directly into the system as an action ledger . Previously minutes were issued and action items sank; the same problems reappeared next month. Now every decision maps to a clear owner, deadline, verifiable acceptance criterion ; the system auto-tracks progress, overdue items auto-escalate. At the next monthly meeting, the first fixed agenda item is reviewing last month's commitments: how many completed on time, how many overdue without escalation, whether actual gains met targets.

Beyond that, stop-loss red lines are codified into the system:

Orders with gross margin below 18% auto-block approval flow, requiring GM sign-off.

Customers overdue >90 days with two broken payment promises auto-trigger shipment hold alerts.

Recurring issues with unclosed root causes auto-escalate to GM supervision.

Rules on paper always have elasticity; rules solidified in the system become rigid. Data capability makes good management logic run 7×24 without relying on the boss's memory or push.

Visible Changes

Meetings shrank from 2.5+ hours of inconclusive debate to fixed 90 minutes with 5+ verifiable decisions per session. Departments shifted from blaming each other to collaborating on the same logic and data. Business units that once saw IT systems as extra burden now proactively request features because the platform helps them find problems and get results. Supply chain used ABC-XYZ inventory analysis to stop predictive stocking for high-volatility materials, freeing ¥2M+ in one month. Sales used customer gross margin scorecards to deliberately shrink several low-margin large accounts; revenue dipped slightly but overall profit rose.

Core Philosophy

Digitalization is never "using technology to replace management"; it is "using technology to amplify management." Without sound management logic, data capability is aimless — however flashy, it's a facade. Without strong data capability, management logic is a castle in the air — reliant on human policing, it won't go far or fast. The first question for any enterprise digitalization should not be "what technology do I have, what features can I build" but "what operating problems must we solve, what management logic do we follow." This holds for operational analysis platforms and for all business digitalization. Technology is always the means; operations are the end. Projects that truly create value are never purely technology-driven; they twist management logic and data capability into a single rope — one sets direction, the other provides horsepower — so numbers become efficient actions and reports become scientific decisions.