Redfin Hot Homes Label: North Star, Forecast Accuracy and Guardrail Metrics

Redfin, the online real estate brokerage, shows a **Hot Home** label on some for-sale listings. Assume the label marks homes that Redfin's model expects to go under contract soon, based on demand signals such as views, saves and tour requests, and that buyers see it in search results and on the listing page in both the app and the website. You are the product manager who owns this feature. How would you measure whether it is a success? ```hint A forecast that buyers can see The label is both a prediction and a prompt to act, and showing it can change the very outcome it predicts. Work out what "it works" would look like for each of those roles before you pick a metric. ``` ### Clarifying Questions - What exactly does the label promise buyers, and over what time window: is it a forecast that the home will go under contract within a set number of days? - Has the label launched in every market, and can we still hide it from a small group of users or listings to measure its effect? - Which outcome matters most to the business: buyers acting in time, more tours and offers through Redfin agents, or engagement with the site in general? - Do sellers and their listing agents see when a listing carries the label? ### What a Strong Answer Covers - The feature's goal, tied to the buyer's problem of losing fast-selling homes and to how Redfin makes money - A buyer journey that shows the moments the label should change, from search results to tour request and offer - A north star with supporting metrics that cover both the accuracy of the forecast and the actions buyers take - Guardrails for label inflation, buyers pushed into bidding wars they lose, and attention taken from unlabeled listings - A measurement design that isolates the label's causal effect from selection and from the forecast fulfilling itself ### Follow-up Questions - Labeled homes go under contract much faster than unlabeled ones. Why is that not evidence that the feature works, and what would be? - The share of labeled homes that actually sell quickly falls in one metro after mortgage rates rise. How do you respond? - Would you let sellers see that their listing is labeled, and how would that change what you measure?