I’ve read a few veteran playbooks that stress justifying each assumption, but in practice I freeze when asked “why” mid-calculation. I want to turn vague prompts into short, credible narratives that link an assumption to a simple signal (device penetration, household count, comparable product uptake). I don’t have a tidy checklist yet — what phrasing or one-liners have you used to justify an anchor without sounding defensive?
you’ll get grilled on that ‘why’ every time. stop offering vague words like ‘reasonable’ or ‘typical’ and give them a signal: ‘i’m using country household count because the product is sold per household’ or ‘iphone penetration is 60% in this demo, so i use that as a device anchor’. one crisp reason beats ten wishy phrases. also, don’t invent fancy multipliers on the fly unless you can explain the mechanism.
- i say short lines like “based on x stat” and name the source quickly. works most times. still nervous tho
Interviewers are checking your chain of logic. For each major assumption, pair a quantitative anchor with a brief qualitative rationale: 1) cite the external signal (e.g., smartphone penetration, household size), 2) explain the behavioral link (why that signal affects adoption), and 3) quantify conservatively. Practice delivering these in one sentence: it demonstrates both rigor and humility. Over time you’ll internalize a set of reusable one-liners tailored to finance vs. tech roles. Which domain gives you more trouble: justifying a revenue-per-user or adoption rate?
- you can do this! rehearse short ‘because’ lines and they’ll become natural. you’ve got this!
I map each assumption to a verifiable proxy: penetration → device or subscription rates; addressable households → census; revenue → ARPU from comparable offerings. In mocks I keep a mental list of quick sources (national stats, telecom penetration, public filings) and a short template: ‘Assumption X anchored to proxy Y because Z (behavioral link).’ Using a 3-point sensitivity around that anchor also reduces pushback—interviewers rarely challenge a defensible range. Which proxies do you currently use?