From Raw Decision Process to Data‑Driven Management: A Step‑by‑Step Guide

Data‑driven decision making is a buzzword, but the article walks through the full workflow, starting from the most primitive decision process that often relies on vague slogans and no data.

1. The most original decision process – sometimes doing something without any data, summarized by the simple "just do it" mentality.

Such empty slogans quickly become ineffective because they lack clarity on what to do, how to do it, and what the outcome should be.

2. The embryonic scientific decision model – a three‑stage "three‑tap" process that adds quantification and begins to resemble scientific management.

This model became popular with the spread of contract systems in the late 80s and early 90s, leading many managers to adopt the "what, how, what result" three‑step phrasing.

3. From coarse to fine – introducing data measurement and analysis to achieve finer‑grained management.

Quantitative evaluation before, during, and after decisions includes:

Pre‑decision: assess current revenue, expenses, profit; evaluate market opportunities; identify natural trends.

During decision: quantify resource requirements, feasibility, and expected completion rates.

Post‑decision: monitor execution, analyze improvements, and review outcomes.

These steps correspond to the classic PDCA (Plan‑Do‑Check‑Act) cycle, which ensures continuous iteration and quality improvement.

Technology is the key enabler: data collection is difficult and requires robust technical support, such as OMS/CRM systems, app or mini‑program tracking, CDP/ECRM tools, and data pipelines.

With these tools, organizations can build richer user profiles, create predictive models, and push data directly to business execution.

The refined process adds:

Goal decomposition to departmental sub‑goals.

Clear metrics and evaluation standards.

Use of CDP and A/B testing for method selection.

Popular OSM methods further break down and quantify indicators to drive decision implementation.

Successful data‑driven decision making requires appropriate tools at each stage, not a single magical model, and faces technical challenges in data collection and organizational challenges in consensus building.

The article concludes with common reasons why data‑driven decisions are hard to perceive in practice: outdated people, systems, and processes; manipulation of data for appearances; blind faith in AI/Big Data; over‑emphasis on metrics without standards; and disconnect between data analysts and business processes.

Ultimately, data‑driven decision making demands tight integration of business workflows and data, leadership support, and realistic expectations rather than magical shortcuts.