Descriptive, predictive and prescriptive analytics: where a decision gets made
The essentials in 30 seconds
- Descriptive analytics asks what happened, predictive asks what may happen, and prescriptive asks what to do.
- Each phase uses the output of the phase below it.
- The decision gets made where a person or a system commits to an action.
- The EQOS Asia method rests on 3 pillars: Audit, Modelling and Prediction (EQOS Asia method page, as of October 2026).
- A team climbs the ladder one rung at a time.
Analytics work is often described as 3 phases: descriptive, predictive and prescriptive. Each phase answers a different question, and the decision itself gets made at the point where a person or a system turns an answer into an action. This guide explains the 3 phases, shows where the decision sits, and places a decision engine on the ladder. It builds on the comparison with a BI dashboard.
On this page
Compare the EQOS Asia offers and request a 30-minute pre-diagnostic.
What do descriptive, predictive and prescriptive analytics each answer?
Descriptive analytics asks what happened, predictive analytics asks what is likely to happen, and prescriptive analytics asks what to do about it. Wikipedia presents prescriptive analytics as the third and final phase, after the descriptive and predictive phases (Wikipedia, Prescriptive analytics, consulted October 2026). The order is a ladder: each rung uses the output of the rung below.
A retail chain in Southeast Asia can walk the ladder with one question about a product line. The descriptive phase reports last quarter’s sales by country. The predictive phase estimates next quarter’s demand. The prescriptive phase suggests how much stock to place in each warehouse. The table summarises the 3 questions and their outputs.
What does descriptive analytics tell a decision maker?
Descriptive analytics examines past performance by mining historical data for the reasons behind success or failure (Wikipedia, Prescriptive analytics, consulted October 2026). Its tools are reports, dashboards and summaries by period, product or region. A monthly report that shows revenue per channel is a descriptive product. It builds a shared view of the facts, and teams usually begin there.
The strength of this phase is clarity: everyone reads the same figures. Its limit is direction. A chart of last year does not rank the options for next year, so a manager adds judgement to move from the picture to the choice. The decision is made in the manager’s head, supported by the evidence on the screen.
What does predictive analytics add to the decision picture?
Predictive analytics estimates what is likely to happen by combining historical data with rules and algorithms, and sometimes external data (Wikipedia, Prescriptive analytics, consulted October 2026). A demand forecast, a churn estimate and a credit risk score are common products. The output is a probability or a projected value, which sets expectations for the coming weeks or months.
A forecast narrows uncertainty and still leaves the action open. Knowing that demand may rise in one market says little about how much budget to place there or which risk to accept. Teams therefore read the forecast, discuss the options, and decide. The decision moves one step closer to the data and remains a human step.
What does prescriptive analytics add as the third phase?
Prescriptive analytics builds on the two earlier phases by recommending actions that benefit from the predictions (Wikipedia, Prescriptive analytics, consulted October 2026). It uses optimisation, simulation and business rules to propose options and show the consequence of each. The decision moves into the tool: the system proposes, and a person approves or adjusts the proposal.
This phase demands a clear objective and well-stated constraints. When both exist, such as routing trucks or allocating a fixed budget, the recommendation can be strong. When the objective is contested, the phase needs an agreed weighting first, and reaching that agreement becomes part of the work.
Where does the decision get made across the 3 phases?
The decision gets made at the moment a person or a system commits to an action. In the descriptive phase the commitment sits entirely with the person. In the predictive phase the person still commits, helped by an estimate. In the prescriptive phase the tool proposes the action, and the person validates it. The share of work done by the tool grows with each rung.
Governance follows the same logic. As the tool takes a larger share, reviewers ask for the inputs, the parameters and the reasoning behind the proposal. Records of those items let colleagues explain any decision months later, and they turn a one-off call into a practice that a team can repeat and improve.
Where does a decision engine sit among the 3 phases of analytics?
A decision engine works across the phases. It reads descriptive inputs, such as internal finance, HR, operations and sales data, projects trajectories as a predictive step, and lists priority risks and corrective levers as a prescriptive step. The EQOS Asia method states 3 pillars, Audit, Modelling and Prediction (EQOS Asia method page, as of October 2026).
Its distinctive step is the structured measurement of a situation. The engine scores 9 dimensions, returns a φ(S) value between 0 and 1, and projects 3 trajectories labelled optimistic, median and adverse. A deterministic core computes the result, so a second run with the same inputs matches the first. The decision stays with the leadership team and rests on a reproducible reading.
How do you move a team from descriptive to prescriptive analytics?
Climb one rung at a time. First agree on the facts, with definitions that every report shares. Next add a forecast for the 1 or 2 metrics that drive the decision. Then write the objective and constraints in words before any model is built. Each rung produces something a colleague can read, so progress stays visible and each stage pays for itself.
A team that wants a structured first reading of a real choice can try the simulator, review the documented method, and book a free 30-minute pre-diagnostic by video, scheduled in SGT (EQOS Asia contact page, as of October 2026). This page is general information, not legal, tax or investment advice.
See also in this guide
Frequently asked questions
What is the difference between descriptive, predictive and prescriptive analytics?
Descriptive analytics reports what happened, predictive analytics estimates what is likely to happen, and prescriptive analytics recommends what to do. Each phase uses the output of the one before it (Wikipedia, Prescriptive analytics, consulted October 2026). The decision itself is made where a person commits to an action.
Which phase makes the decision?
Each phase supports the decision, and the commitment sits with the person who approves the action. In the descriptive phase the person carries the full weight of the judgement. In the prescriptive phase the tool proposes an action that the person validates or adjusts before anything moves.
Does every decision need all 3 phases?
Many decisions use a subset. A simple review may need only a descriptive report, and a budget allocation may move to prescriptive analytics once the facts are agreed. The ladder shows which kind of help exists, and the question decides which rung is worth the effort.
Where does the EQOS Asia engine fit?
It spans the phases: it reads descriptive inputs, projects 3 trajectories and lists corrective levers. Its method rests on 3 pillars, Audit, Modelling and Prediction, and computes a φ(S) score between 0 and 1 on 9 dimensions (EQOS Asia method page, as of October 2026).
Test your knowledge
1. Which phase asks what happened?
- Predictive
- Descriptive
- Prescriptive
Show the answer
Descriptive.
2. Which phase is described as the third and final phase?
- Prescriptive
- Descriptive
- Predictive
Show the answer
Prescriptive.
3. How many pillars does the EQOS Asia method rest on?
- 9
- 2
- 3
Show the answer
3.
4. Which labels name the three EQOS Asia trajectories?
- Optimistic, median and adverse
- Past, present and future
- Low, mid and high
Show the answer
Optimistic, median and adverse.
5. In the prescriptive phase, who validates the proposed action?
- The report
- A person
- The chart colour
Show the answer
A person.
About the author
Stéphane Jambu
President, EQOS Forecast Asia
Stéphane Jambu leads EQOS Forecast Asia and works as a consultant in e-commerce, search visibility, and AI. LinkedIn profile
Sources: EQOS Asia method page (eqos-asia.com/method.html), simulator page (eqos-asia.com/simulator.html) and contact page (eqos-asia.com/contact.html), en.wikipedia.org/wiki/Prescriptive_analytics, consulted on 22 October 2026. Page updated on 22 October 2026.