Metrics question

Choose one decision moment in Ramp, such as 60 days before renewal, during new vendor intake, or when spend suddenly expands. Design the in-product intervention: what should the user see and be able to do, what signals should trigger it, and what metrics would tell you it changed customer behavior rather than just generating clicks?

Practice this question out loud. An AI interviewer asks it, follows up like a real interviewer would, and scores your answer. Type or speak.

What this question tests

Whether you can design a trigger-based, in-product intervention and pick metrics that prove behavior change rather than vanity engagement.

Choose the moment: 60 days before renewal, since it gives a finance team enough runway to renegotiate or switch vendors.
Define the trigger signal: contract end date plus recent spend trend on that vendor, sourced from Ramp's existing card and subscription data.
Design what the user sees: a card showing current spend, the benchmark price range for comparable companies, and one clear action, for example 'start a renegotiation checklist' or 'flag for review'.
Keep the action low-friction: pre-fill a negotiation email or comparison doc rather than sending the user elsewhere.
Pick outcome metrics over click metrics: percent of triggered renewals where the user took the suggested action, and average savings or price change achieved on those renewals.
Set a guardrail metric, for example do not increase support tickets or renewal-cycle time as a side effect.

Distinguishes engagement metrics (click-through on the card) from behavior metrics (renegotiated price, switched vendor, or documented decision to hold), and prioritizes the latter.
Gives an illustrative target, for example assume 15% of triggered renewals result in a measurable price change versus a baseline close to 0% today.
Designs the intervention to reduce user effort, not just inform, for example auto-drafting a negotiation email using the benchmark data.
Notes a guardrail: false or stale benchmark data would erode trust fast, so freshness of the underlying price data is a precondition.

Measuring only impressions or clicks on the nudge instead of a real behavior or dollar outcome.
Picking a trigger with no reliable underlying data signal.
Ignoring the guardrail that a wrong benchmark number damages trust in the whole product.

How would you validate the trigger timing of 60 days is right instead of 30 or 90?
What would you do if usage of the nudge is high but savings do not materialize?
How would you extend this to the new-vendor-intake moment?