Using Data Analysis to Choose the Right Advertising Strategy#

🎯 Data-Driven Decisions 🧭 Framing the Problem Lesson 001

Important

✨ AI-generated content. This page was written with the assistance of an AI language model and is provided as a learning aid. Despite careful review, it may still contain mistakes, omissions, or out-of-date information. Whether you are new to the topic, a team lead, or a senior practitioner, treat it as a starting point rather than an authoritative reference: read it critically and independently verify anything you act on (code, commands, figures, and factual claims) against official documentation and primary sources before relying on it.

A decision every business faces#

Where should the next advertising money go? Radio, social media, search ads, billboards, a sponsorship? Every organisation that markets faces this choice, and it is the canonical example of a prediction problem: use data about what past campaigns produced to make an informed decision about what a future one will produce. This section opens with it because it shows the entire decision-making arc of the coming lessons in one familiar case.

What the data can say#

Suppose the company has records of past campaigns: the medium, the location or audience targeted, the spend, and the number of new customers acquired in each campaign window. Analysis can then compare like with like:

cost per new customer, by medium — which channel acquires cheapest;
performance by location or audience — where each medium works best;
trend over time — whether a channel’s performance is rising or fading.

A simple table of cost per acquisition by channel and region is often the whole deliverable: it converts “which advertising do we believe in?” into “which row of this table is smallest, and do we trust why?”

What the data cannot say#

Past performance cannot guarantee future results — the honest phrasing is that analysis helps predict the best placement to reach the target audience, not that it certifies an outcome. Markets shift, competitors react, and a channel that worked at small spend may saturate at large spend. Two disciplines keep the prediction honest: state the assumption (next quarter resembles last), and instrument the decision — run the chosen strategy in a way that measures its result, so the next round of this same decision has fresher evidence. That is the decision loop from the foundations, operating.

Why this case opens the section#

Notice what the analysis required before any computation: someone had to frame the vague worry (“is our advertising working?”) as a specific, answerable, decision-attached question (“which channel acquires customers cheapest, for whom, and is that stable?”). The rest of this stage teaches that framing skill in general — the problem types that recur, and the questions that unlock them.

Hint