Introduction to Web Analytics - Zach Olsen

Main ideas & lessons

• Web analytics is part of a broader research/optimization toolbox

  • It fits in the quantitative + behavioral quadrant: it shows what users do on your site (navigation, clicks, time, conversions).
  • It doesn’t explain “why” users behave that way; to uncover reasons, teams often use qualitative methods like surveys, interviews, and usability testing.

• Web analytics answers “what,” not “why,” and should be combined with qualitative tools

  • Example gap: you might see low conversion (e.g., 2%) and know most people don’t buy—but not why.
  • Surveys can ask intent, e.g., “What is your purpose for coming to the website today?”
  • Usability testing can reveal where people get stuck or misunderstand tasks—sometimes the behavior is hard to interpret purely from analytics logs/reports.

• There are multiple categories of tools (qualitative + quantitative)

  • Surveys / attitudinal feedback tools: Foresee, Opinionlabs (feedback widgets; can support CSAT/NPS-style scoring).
  • Core analytics platforms (quantitative behavior):
    • Google Analytics (free) + Google Analytics Premium (enterprise cost mentioned).
    • Adobe Analytics (paid/enterprise; highly customizable; many large retailers use it).
  • Behavior visualization / heat-mapping tools: Clicktale, Crazyegg, Hotjar
    • Heatmaps for click concentration
    • Scroll-depth visualization
  • Usability testing platforms: UserZoom (remote usability; observe users performing tasks, watch facial/behavioral cues).
  • A/B & multivariate testing: split traffic between versions (A vs B) to measure which performs better in real time.
  • Personalization: use visitor/profile data (e.g., loyalty profile attributes) to dynamically tailor page content to improve conversions.

• Web analytics has two sides: analyst vs implementation

  • Analyst side
    • Build and interpret reports
    • Identify trends (traffic, conversion)
    • Recommend optimizations
  • Implementation side
    • Place tracking/tagging code on pages so events are captured
    • Requires technical skills (e.g., JavaScript familiarity—especially for Adobe)
    • Example: track clicks on a new landing page button by instrumenting the page so it shows up in reporting

Methodology / instruction-style workflow (detailed bullets)

• Start with business objectives and KPIs

  • Identify what the website is trying to achieve (e.g., sales, brand awareness, loyalty, customer satisfaction).
  • Determine the KPI(s) that directly align with those objectives.
    • E-commerce revenue objective → KPI: sales / goal completions
    • Brand awareness objective → KPI could be social share button clicks (and then assess downstream actions)
    • Loyalty objective → KPI could be return visits / repeat orders and email signups
    • Customer satisfaction objective → KPI could be Net Promoter Score (NPS) from surveys

• Use web analytics to look at user behavior through the KPI lens

  • Review traffic acquisition:
    • New sessions vs new users
  • Review behavior quality:
    • Bounce rate (users who leave after the first page)
  • Review outcomes:
    • Conversions / goal completions tied to business goals

• Generate and test improvement ideas (“turn the needle”)

  • Treat insights as a means to drive action on KPIs—not just reporting.
  • Example optimization logic:
    • If product detail pages have a low Add to Cart conversion rate (e.g., 15%), aim to raise it (e.g., to 20%).
    • Investigate which on-page interactions correlate with better conversion:
      • Tag and analyze clicks on reviews tab
      • Tag and analyze clicks on size/fit guide
      • Adjust UI elements (e.g., default tab, prominence/color of buttons like “Add to Cart”)

• Segment and drill down to determine “what’s different”

  • Compare channels (e.g., direct vs paid search vs organic/referral).
  • Ask diagnostic questions such as:
    • Why does direct traffic have higher average time on page but higher bounce rate?
  • Drill further:
    • Break direct traffic down by page
    • Use “page value” concepts (value attributed to pages that precede conversions)

• Choose appropriate channels and routes to high-value pages

  • Identify high-value pages (e.g., a specific “About”/development page that precedes purchases).
  • Decide how to drive more relevant visits to them:
    • SEO/organic ranking improvement
    • Email campaigns linking to that page
    • Analyze underperforming pages to see which traffic sources contribute poorly

• Build practice-based analytics skills

  • For beginners: install and use tools on a personal or test site first.
  • Suggested hands-on stack from the talk:
    • Install Google Analytics
    • Install Hotjar (heatmaps + scrolling)
    • Set up goals in Google Analytics aligned to KPIs
  • Emphasize that mastery comes from doing: tagging, setting up experiments, iterating.

When to switch from Google Analytics to Adobe (decision concept)

• Move to a paid/enterprise solution like Adobe Analytics when:
  • Google Analytics doesn’t provide the level of customization needed.
• For small to many mid-sized businesses:
  • Google Analytics is described as “brilliant” and sufficient.
• The talk emphasizes avoiding “magic-vendor” thinking: outcomes come from using the data to make changes.

Speakers / sources featured

People (speakers)

• Trevor Erickson (host; introduces the session)
• Zach Olsen (digital analytics manager; main presenter)

Tools / vendors mentioned as sources or examples

• Stukent (session platform/series)
• Foresee
• Opinionlabs
• Adobe Analytics (formerly Omniture)
• Google Analytics (and Google Analytics Premium)
• Clicktale
• Crazyegg
• Hotjar
• UserZoom
• Google Tag Manager
• NPS / Net Promoter Score (survey metric concept; survey not a specific vendor here)

Organizations/brands mentioned

• Columbia Sportswear (Zach’s employer)
• Other Adobe Analytics user examples: Best Buy, Guitar Center, Crocs, Nike, Patagonia

Websites referenced by URL

• ByDataBeDriven.com (Zach’s blog)
• ZachOlsen.net (Zach’s personal site)
• Indeed.com (used as an example for job browsing)