Absolute:"Latency dropped from 450ms to 90ms" (good for context, not as a sole metric)
Trade-off:"… while memory usage remained unchanged" / "… at the cost of a 10% increase in CPU"
1 / 3
A developer writes in a PR description: "The new caching layer reduced average query time from 450ms to 90ms." Which sentence correctly describes this improvement in professional English?
Option C is the most complete and professionally precise description. It gives both: (1) the percentage improvement: (450-90)/450 = 400/450 ≈ 88.9% ≈ 80% (rounding down conservatively), and (2) the multiplier: 450/90 = 5× faster. Presenting both is common in engineering reports. Option A states only the absolute difference — "360ms" means little without context. Option B says "5 times" which is ambiguous — does that mean 5× faster (new = old/5) or 5 times as slow? In engineering, say "5× improvement" or "5× faster". Option D restates the absolute difference again. Vocabulary: "reduced by X%", "improved by a factor of X", "Xₓ faster", "cut in half", "halved", "tripled", "quadrupled".
2 / 3
During a sprint demo, your team delivers these benchmark results:
Image compression benchmark:
v1 (ffmpeg): avg 2,140ms per image
v2 (libjpeg): avg 310ms per image
Memory usage: unchanged (~45 MB)
Output quality: SSIM score 0.97 vs 0.96
Which sentence best summarises these results for a non-technical stakeholder?
Option B is ideal for a non-technical stakeholder. It: (1) leads with the key win ("85% faster"), (2) translates milliseconds into seconds for a lay audience, (3) addresses the two concerns stakeholders always have about a performance change — does it affect memory? does it affect quality? — and gives plain answers. Option A mentions SSIM (a technical image quality metric) without explaining it — a stakeholder can't evaluate this. Option C uses "6.9×" and "SSIM delta" — technical terms without context. Option D gives only the absolute time saving in raw milliseconds. For stakeholder communication: lead with the headline stat, translate units to human scale, address quality/cost tradeoffs.
3 / 3
A performance report states: "After deploying the new database indexes, p95 latency decreased by 67% while CPU utilisation increased from 12% to 18%." Is this a good trade-off? Which statement is the most professionally reasoned response?
Option C is the most professionally reasoned answer. It: (1) frames the latency reduction as user-facing impact, (2) correctly interprets the CPU change — noting that it is a 6 percentage point increase (not 50% of capacity reached), and (3) contextualises the remaining headroom. This is critical: going from 12% to 18% CPU means the system is still at 18% utilisation — far from saturation. Only at 70–80%+ sustained utilisation would CPU become a concern. Option A confuses absolute percentage point changes with catastrophic load. Option B ignores CPU cost entirely — in capacity planning, resource utilisation always matters. Option D jumps to "revert" without reasoning. Key distinction: "CPU utilisation increased by 50%" (relative) vs. "CPU increased from 12% to 18%" (absolute, 6 pp).