An independent study reference written by Dr Phuc V. Nguyen. It is not official subject material — for assessment requirements always follow your subject outline and vUWS.
What business analytics is
Business analytics is the practice of turning data into better decisions. Three things have to be present: a decision that repeats or matters, evidence drawn from data, and a method connecting the two. The field covers the description of what happened, the prediction of what is likely, and the choice of what to do about it. The deliverable is not a report. The deliverable is a changed decision, and the value of the work is the gap between the decision made with the analysis and the decision that would have been made without it.
Why it matters
Picture a manager about to make a call she makes fifty times a year. She already has instincts, a spreadsheet and some memory of last time. Analytics is anything that makes her fiftieth call reliably better than her first. That might be as small as a clean count refreshed every seven days or as large as a trained model. If nothing about the call changes, nothing has been added, however impressive the chart looks.
A retailer builds an accurate seven-day demand forecast. Buyers keep ordering from their own spreadsheets. In business analytics terms, what has the retailer produced?
Formulas
Worked examples
A regional pharmacy chain has two years of transaction data and a new dashboard, but stock decisions are still made by each store manager on Monday morning. The board asks whether the analytics investment is working.
Ask which decision changed. If managers order the way they always did, the dashboard has produced description and no analytics value. The useful version starts from the decision. Each store places roughly fifty order lines in every ordering cycle, and on each line the chain loses margin when it runs out and loses stock when it over-orders. Work that shifts even a fifth of those lines to a better quantity is worth doing, and it can be verified, because the outcome of every order line is observed seven days later.
A university wants to know whether analytics can help with student withdrawal in the first year of a degree programme.
Withdrawal is not a decision. The decision is which students receive an outreach call in the first month, and the constraint is that the support team can make about three hundred calls. That reframing makes the problem analytical. There is a repeated decision, a limited resource, an outcome that becomes observable, and a data trail in enrolment and learning system records. The question turns into which three hundred names, rather than which factors are associated with withdrawal.
Common mistakes
- ✗Business analytics is a technology you buy. The software is the cheapest part of it. What makes analytics work is a defined decision, data that is fit for that decision, and a person with the authority to act on the answer.
- ✗Analytics means predicting the future. Most of the value in most organisations still comes from describing the present accurately, because a shared and correct picture of what is happening settles more arguments than any forecast.
- ✗More data automatically means better decisions. Data helps only when it bears on the specific decision at hand. A terabyte of clickstream tells you nothing useful about which supplier contract to renew.
- ✗If the analysis is correct, the decision will follow. Correct analysis that arrives after the decision window, or in a form nobody can act on, changes nothing. Delivery is part of the work rather than an afterthought.
Revision bullets
- •Analytics needs three things: a decision, data, and a method joining them
- •The deliverable is a changed decision, not a report
- •Value = decisions touched x lift in decision quality x stake per decision
- •Description, prediction and prescription all count as analytics
- •Software is the cheap part; the decision owner and fit data are scarce
Quick check
A retailer builds an accurate seven-day demand forecast. Buyers keep ordering from their own spreadsheets. In business analytics terms, what has the retailer produced?
Two proposals compete for one budget. Proposal A improves a decision made once a year with A$5 million at stake, lifting the chance of the right call by two percentage points. Proposal B improves 40,000 decisions a year worth A$50 each, lifting the right-call rate by five percentage points. On expected value alone, which is the stronger case?
Connected topics
More in What Business Analytics Is
Sources
- Davenport & Harris (2007)Davenport, T. H., & Harris, J. G. Competing on Analytics: The New Science of Winning. Harvard Business School Press, 2007.Argues that analytics is an organisational capability tied to decisions and processes, not a software purchase.
- Provost, F., & Fawcett, T. "Data Science and its Relationship to Big Data and Data-Driven Decision Making." Big Data, 1(1), 51-59, 2013.Frames data-driven decision making as the point of the exercise and separates it from the technology used to support it.