Web Analytics Interpretation for UX

Description

Read core metrics, chart shapes, page patterns, landing pages and segments to locate where behaviour departs from expectation and form hypotheses for other methods.

Analytics shows the 'smoke' that leads to the 'fire' of user goals. Interpretation is relative: compare a page with others of its type and with its own history, not with the site average. Core metrics are read together: sessions against users for loyalty, new versus returning share against site maturity, pageviews and time by whether the product is a task tool or an engagement product, and bounce bands (10-30% good, 70-99% poor, near zero a tracking error). Chart shapes are diagnostic: traffic graphs show trends, spikes with a cause and cycles; item or step graphs normally slide exponentially, so out-of-order bars suggest an IA mismatch, bumps suggest power users and cliffs mark barriers. Landing-page reports with referral context, and rule-based segments that isolate relevant visitors, sharpen each diagnosis.

Attributes

Loss type
Active
Effort
Days

Steps

  1. Filter out rarely viewed pages (for example under 100 views) and group the rest by template or type.
  2. Read core metrics in combination and against the site's own history: sessions per user, new versus returning share, pageviews, time, bounce and exit rate.
  3. Classify each chart as a traffic graph or an item/step graph and note trends, spikes, cycles, out-of-order bars, bumps and cliffs.
  4. Sort page metrics for highest, lowest and deviation from the type average and map each pattern to candidate explanations (for example pageviews 40-50% above unique pageviews suggests a hub or pogo-sticking).
  5. Rank landing pages by entrances, add bounce and conversion, and infer arrival intent from medium, keyword or the linking page.
  6. Build segments that separate relevant users (by page visited, keyword category, landing page or goal), testing that they capture about 80-90% of the intended visits.
  7. Turn each suspect page or pattern into a hypothesis and test it with usability tasks, expert review or an A/B test.

Sources

  • UX For Beginners (Marsh, 2016)
  • Practical Web Analytics for User Experience (Beasley, 2013)
  • A-Z of Digital Research Methods (Dawson, 2020)
  • Benchmarking the User Experience (Sauro, 2018)
  • Just Enough Research (Hall, 2013)
  • Customer Analytics for Dummies (Sauro, 2015)

Also known as

  • Reading analytics graph shapes
  • Interpreting core web analytics metrics
  • Page-level metric pattern diagnosis
  • Landing page analysis
  • Rule-based analytics segmentation
  • Advanced segments
  • Outlier page analysis

Categories

  • Analytics & Behaviour
  • Product Analytics & Strategy