問34
頻出度★★★
あるオンラインサービスでは, 新たに作成したデザインと従来のデザインの Web サイトを実験的に並行稼働し, どちらの Web サイトの利用者がより有料サービスの申込みに至りやすいかを比較, 検証した。このとき用いた手法として, 最も適切なものはどれか。
令和4年度 ITパスポート試験 問34
ストラテジ系経営戦略マネジメント › マーケティング
This question's explanation and your answers so far.
解説
問 34
- ア
正解 The correct answer is ア “A/B testing”. A/B testing is a method in which two or more versions, such as a new design (version A) and the existing design (version B), are run in parallel at the same time; users are split among the versions, and real data is used to compare and verify which version achieves a higher outcome metric (here, the sign-up rate for the paid service, i.e. the conversion rate). The statement in the question, that the new and old designs were run experimentally in parallel and it was compared and verified which side of users were more likely to sign up, matches the definition of A/B testing exactly. It is used to improve websites and apps by choosing the better version based on measured data rather than on intuition.
Incorrectイ
ABC analysis is a method that arranges items such as inventory items, products, or customers in descending order of sales or monetary value, classifies them into three groups A, B, and C by cumulative composition ratio, and focuses management on the high-importance group A (an application of Pareto analysis). It is an analysis for narrowing down the targets of priority management, not a method for running two design versions in parallel to compare which is better.
Incorrectウ
Cluster analysis is a multivariate analysis method that automatically groups a large amount of data into several groups (clusters) of highly similar items. It is used for tasks such as customer segmentation, but it does not compare and verify the outcomes of two versions prepared in advance.
Incorrectエ
Regression testing is a test that, after software is fixed or changed, checks whether existing functions still work correctly as before and whether any unexpected defects (degradation) have arisen. It is verification for maintaining software quality, not a way to compare which design version is better based on user reactions.
ここがポイント
- A/B testing: a verification method that runs two or more versions in parallel at the same time, randomly assigns users, compares outcome metrics such as conversion rate using real data, and selects the better version.
- Distinguishing confusable terms: ABC analysis = classification for priority management, cluster analysis = grouping of similar data, regression testing = checking for defects after changes; their purposes all differ.
用語