### Overview Replacing one product with another is rarely all upside. Alienation Analysis answers the question behind that worry: of the participants who like what you have now, how many would reject its replacement? It compares a **Control** against a **Test** — the change you're considering — using your **Overall Liking** question, and returns a verdict of **Low Risk**, **Moderate Risk** or **High Risk**. Reformulating a drink? It tells you what share of your Baseline Formula fans wouldn't buy Reformulation B. Throughout this article, **Products** are the labels you see on a physical product study. On a concept study they read **Concepts**; on a survey-only study, **Surveys**. ### Where to Find Alienation Analysis Go to **Insights**, open the **Analysis** menu, and choose **Alienation Analysis**. The tab is built for alienation studies and needs an **Overall Liking** question. Don't see it? Ask your account manager. You don't have to wait for fielding to finish — the analysis runs on whatever has arrived, and the verdict can move as more responses land. ### Choosing Your Control and Test Both are set in the **Settings** sidebar, and you need both before anything calculates. - **Control** — your current or original product, the baseline. - **Test** — the new or modified product you're evaluating. Each dropdown hides whatever the other has selected, so the same product can never be both, and the analysis recalculates on its own — there's no Apply button. Because it's one product against one other, [Product Groups](/content/help/product-groups/index.html) aren't offered here. ### The Four Groups Each participant lands in one of four rows, based on how they rated the two products. A rating above the midpoint of the scale counts as liking a product; at or below it counts as disliking. On a 9-point scale, 6 and up is liking and 5 and down is disliking — hover the **Response Category** column header to see the split for your own scale. - **Liked** — liked both. - **Disliked** — disliked both. - **Alienated** — liked the Control, disliked the Test. These are the people you'd lose. - **Noise** — disliked the Control, liked the Test. Movement in the opposite direction. The **Total** row is how many participants were matched across both products, and it's the denominator for all four percentages. Everyone has to have rated both — an alienation study is recruited that way on purpose — so anyone who answered **Overall Liking** on only one product drops out. ### Why Noise Matters A chunk of people going from liking the Control to disliking the Test sounds alarming — until you notice a similar chunk moved the other way. That symmetry isn't loyalty broken by the Test. It's people whose preferences in this category aren't firm to begin with. So the verdict weighs **Alienated** against **Noise**: - When **Alienated** is higher than **Noise**, the alienation figure sets the tier — below 10% is **Low Risk**, 10% to 19% is **Moderate Risk**, 20% or more is **High Risk**. - When **Alienated** is at or below **Noise**, the noise figure sets it instead — 10% or below is **Low Risk**, above 10% is **Moderate Risk**. **High Risk** can't be reached on this branch. ### Your Verdict, and What to Do About It - **Low Risk** (green) — the change is relatively safe. Proceed, and watch your post-launch metrics. - **Moderate Risk** (yellow) — some potential for losing people. Do more research first: [Penalty Analysis](/content/help/penalty-analysis/index.html) shows which attributes are dragging Overall Liking down. - **High Risk** (red) — a real risk of losing loyal customers. Refine the product or test alternatives first. **Tip**: **Moderate Risk** means two different things. Driven by **Alienated**, a genuine minority would defect. Driven by **Noise**, preferences are unstable in both directions — a signal about the category or the sample, not about your Test. Check which percentage is larger before you act. ### Looking at One Segment The **Filter** dropdown in the **Settings** sidebar narrows the whole analysis to one segment. A project-wide **Low Risk** can hide a group that would defect heavily, and running the analysis segment by segment is the only way to find it. Only segments that are Ready appear in the list; build them at **Insights** > **Tools** > **Segments**. **Filter** isn't the only control that rescopes the analysis. **Full Completes Only** and **Read Position**, further down the sidebar, do too — and **Full Completes Only** is project-wide, so a colleague can change it and change your verdict. ### Exporting the Analysis There's no download button on this tab. To get the analysis out of Insights, generate an [Instant Report](/content/help/instant-reports/index.html) — the deck carries an **Alienation Analysis** slide with the same table and verdict. Set the **Control** and **Test** on the **Reports** tab first, or the slide is left out silently.