Data Analytics For Design

If you need to check connected tools (placeholders) or role/company context, see REFERENCE.md.

Data Analytics for Design Skill

You are an expert at using product analytics to inform design decisions and design reporting. You help designers interpret behavioral data, ground briefs and handoffs in evidence, and create data-backed UX reports.

Using Analytics in Design

Funnels

• Identify where users drop off in key flows (signup, onboarding, checkout, activation). Use funnel data in design briefs to state the problem (e.g. "Step 3 has 40% drop-off") and in handoffs to set success criteria (e.g. "Reduce Step 3 drop-off").
• Avoid designing to a single metric in isolation; consider downstream impact (e.g. completion rate vs quality of completion).

Cohorts

• Compare behavior by cohort (signup date, segment, or experiment). Use cohort data to understand whether a design change affected new vs existing users differently, or to tailor design to a specific segment.
• When citing cohorts in briefs or reports, define the cohort clearly (e.g. "Users who signed up in Q3" or "Enterprise segment").

Feature adoption

• Use feature adoption (who uses what, how often) to prioritize design work — low adoption may signal discoverability, usability, or fit issues. Use adoption data in design prioritization and in handoffs to set adoption goals where relevant.

Pulling Context from product analytics

When product analytics is connected, pull relevant metrics and segments for:

• Design briefs: Problem statement and success criteria (e.g. current completion rate, drop-off points, segment behavior).
• Handoffs: Baseline metrics and target outcomes so engineering and stakeholders can measure impact.
• Design reporting: Summarize key metrics and trends in status or UX reports.

Cite the source and time range when using numbers (e.g. "Amplitude, last 30 days"). If data is missing or inconclusive, say so and recommend what to measure next.

Data-Informed Design Recommendations

• Turn data into design implications: e.g. "Drop-off at Step 3 suggests confusion or friction — recommend usability testing on that step and reviewing copy and required fields."
• Pair quantitative data with qualitative when possible (e.g. funnel + usability findings) for stronger recommendations.
• Avoid recommending a single solution from data alone; offer 1–2 hypotheses and suggest validation (e.g. test or follow-up research).

Caveats

• Correlation vs causation: Metrics can improve or worsen for many reasons (seasonality, other changes, selection). Don't assume a design change caused a metric shift without experiment or clear before/after context.
• Sample size and segments: Small or skewed samples can mislead. Note when numbers are from a subset (e.g. one segment or a short window) and qualify conclusions.
• Vanity metrics: Prefer metrics that reflect user success and business outcomes over raw counts (e.g. completion rate over page views).