Quantify your impact
Turn vague results into concrete, credible numbers.
- Express results with metrics such as latency, cost, revenue, or time saved.
- Estimate honestly when exact numbers are unavailable.
- Explain why a result mattered to users or the business.
“The page got faster” is forgettable. “I cut the checkout page’s p95 load time from 4.1 s to 1.2 s, and conversion rose 6%” is evidence. Numbers make your impact concrete, show you understood what mattered, and give the interviewer something to write in their notes.
What to measure
- Performance: latency, throughput, error rate, uptime.
- Cost: infrastructure spend, engineering hours saved.
- Business: revenue, conversion, retention, support tickets.
- Team and process: deploy frequency, build time, onboarding time, incidents.
- Scale: users, requests, data volume, number of teams affected.
Pair each number with its so what: “Build time went from 25 to 6 minutes, so engineers deployed several times a day instead of batching changes.”
before, after = 25, 6 # build minutes
change = round((before - after) / before * 100)
print(f"Build time down {change}%: {before} -> {after} minutes")Build time down 76%: 25 -> 6 minutes
If you don’t have exact figures, estimate and say so: “roughly 30% fewer on-call pages, based on the incident dashboard.” Never invent numbers - interviewers ask follow-up questions, and precise-sounding figures you can’t explain destroy credibility. Relative changes (percentages) are often safer to share than confidential absolute values.
Key takeaways
Quantify outcomes: performance, cost, business, team, scale.
Always add the “so what”.
Estimate honestly and never fabricate.
Lesson quiz
5 questions · pass with 4 correct · up to 50 XP
Passing this quiz completes the lesson and keeps your streak going. Questions you miss come back in review sessions later.
Practice: sharpen your stories
Use short Python programs to check your story structure, time your answers, and quantify impact. These exercises run locally in your browser.
Express a change as a percentage
Read a metric name, a before value, and an after value on three lines. Compute the change as a whole-number percentage of the before value, round(abs(after - before) / before * 100). Print metric: before -> after (P% lower) or (P% higher), or (no change) if they are equal.
- Lower latency
- Higher conversion
- No change
Python runs in a sandboxed browser worker with a 60 second time limit. Its runtime loads from the Pyodide CDN; your code stays in this browser.
Questions about this lesson
Stuck? Ask. Figured something out? Share it. Explaining is one of the best ways to learn.
Loading posts…