IT Passport Examination (IP) | Strategy Questions on Data Analysis, Marketing, and DX Planning 06

Correlation shows that two variables tend to change together, but it does not prove that one causes the other. Here, temperature could affect both ice cream purchases and activity near water, producing a spurious correlation. Time order, other factors, comparison groups, and other evidence are needed before drawing a causal conclusion.

Even a large sample can produce biased estimates when its selection systematically differs from the target population. An in-app survey underrepresents customers who do not use the app or who have already left. To estimate all customers, define the population and consider sampling or adjustment across channels and customer attributes. Sample size alone does not eliminate selection bias.

Truncating the vertical axis of a bar chart can visually exaggerate a two-point difference because bar length represents magnitude. A zero baseline with actual values is clear. If an axis break is needed to show a small change, the break, scale, and purpose should be explicit. Hiding values or labels reduces verifiability rather than improving objectivity.

RFM classifies customers by how recently they purchased, how frequently they purchased, and how much they spent in total. The described customers have strong historical Frequency and Monetary values but weak Recency because they have not purchased lately, making them candidates for reactivation. Demographic attributes are not the three RFM measures.

To isolate the effect of wording, use comparable groups at the same time and keep conditions other than the wording consistent. Defining the metric and period in advance also reduces the risk of selecting only a favorable result. If month, price, and advertising all change, the source of any purchase-rate difference cannot be identified.

A PoC mainly checks whether a concept or technology is feasible. Since technical recognition accuracy has already been demonstrated, a PoV should test outcomes such as time saved, user demand, and value relative to cost. Even after value is demonstrated, production readiness still requires separate evaluation of operations, scalability, and commercial viability.

An API economy arises when functions and data are connected through APIs across organizations and third parties combine them to create new value. It is different from manually copying a file inside one organization. API exposure also requires design decisions about data scope, authentication and authorization, rate limits, contracts, personal data, and incident responsibilities.

BPR redesigns a process from its required outcomes and controls without assuming that current departmental boundaries and steps must remain. Here, it should examine the purpose of duplicate entry and approval, integrate or remove steps, and revise responsibilities. Automating the same wasteful procedure may make clicks faster but leaves the process problem intact.

Omnichannel integrates customer touchpoints such as stores, e-commerce, and apps so that customers receive a consistent experience across them. Shared inventory, identity, loyalty points, and in-store pickup are examples. Merely operating several channels while keeping their data and experiences isolated does not provide that integration.

Dynamic pricing changes prices according to conditions such as demand, supply, time, and inventory. Unlike skimming pricing, which may set a high price at product launch, this method continually reflects changing conditions in the price. Its operation should also consider explainability, fairness, legal requirements, and effects on customer experience.