Research Fundamentals Part 5: Web Analytics Foundations for A/B Testers
Web Analytics Foundations for A/B Testers
Web Analytics Foundations for A/B Testers
⏱️ 2-min read
Analytics shows you where money is being lost. Every drop-off point in your funnel represents revenue walking away. Your job is finding the biggest leaks and testing fixes.
Here’s what to look at.
Funnel Analysis: The Foundation
Map your entire conversion funnel. For an e-commerce site, it might be:
Homepage → Category Page → Product Page → Cart → Checkout → Purchase
For a SaaS product:
Landing Page → Signup Page → Onboarding → Activation → Paid Conversion
Calculate the conversion rate at each step. If 1,000 people hit your landing page, 300 click to signup, 150 complete signup, 50 activate, and 10 convert to paid, your funnel looks like this:
Landing → Signup: 30%
Signup → Complete: 50%
Complete → Activate: 33%
Activate → Paid: 20%
The biggest drop is Landing to Signup. That’s where you’re losing 70% of potential customers. Start testing there.
Traffic Source Performance
Not all traffic converts equally. Break down conversion rates by source:
Organic search
Paid search
Social media
Email
Direct traffic
Referrals
If organic search converts at 8% but social media converts at 1%, either your social traffic is wrong (targeting problem) or your page doesn’t address social visitors’ needs (messaging problem).
Test different messaging for different sources. Create dedicated landing pages for paid traffic versus organic traffic if their intent differs.
Device and Browser Breakdown
Check conversion rates across:
Desktop versus mobile versus tablet
Chrome versus Safari versus Firefox
iOS versus Android
If mobile converts at 2% while desktop converts at 6%, you have a mobile experience problem. Maybe your forms are too long. Maybe buttons are too small. Maybe the page loads slowly on mobile.
If Safari users convert at half the rate of Chrome users, you might have a browser-specific bug. Test on Safari and see what’s broken.
High Exit Pages
Which pages do people leave from most often? If 40% of visitors exit from your pricing page, something on that page is causing concerns. Maybe the prices are unclear. Maybe the value isn’t justified. Maybe key information is missing.
High exit rates signal testing opportunities. Investigate those pages with heat maps, session replays, and surveys to understand why people leave.
Time on Page vs Conversion
For long-form sales pages or detailed product pages, time on page matters. If people who spend 3+ minutes on the page convert at 15% but people who spend under 1 minute convert at 2%, engagement correlates with conversion.
Test making critical content more engaging or moving important information higher so people see it within that first minute.
Form Analytics
If you have forms, track:
Field completion rate (which fields get filled versus which get abandoned)
Time per field (which fields take longest to complete)
Error rate per field (which fields trigger validation errors)
Abandonment point (which field do people leave on)
If 40% of people abandon at the “Company Name” field, maybe that field is unnecessary or unclear. Test removing it or making it optional.
Assisted Conversions
Some channels don’t convert directly but help. Someone might discover you through social media, research you via organic search, and finally convert through an email campaign.
Check assisted conversions in analytics. If social media has a 1% direct conversion rate but assists 20% of all conversions, it’s more valuable than it appears.
This affects testing strategy. If a test improves direct conversion but hurts brand awareness or assisted conversions, it might be net negative.
What Numbers Actually Mean
A 5% conversion rate isn’t inherently good or bad. It depends on your business, your traffic, and your industry.
What matters is: Where are your biggest drop-offs?
Which segments convert best and worst?
Where is the biggest opportunity for improvement?
Focus testing efforts on pages with: High traffic (more opportunity)
Low conversion rates (more room for improvement)
Critical funnel position (affects everything downstream)
Common Analytics Mistakes
Mistake: Looking at overall conversion rate instead of segment-specific rates. Mobile might be terrible while desktop is great. Averages hide problems.
Mistake: Ignoring low-traffic pages that have high value. Your enterprise pricing page might only get 200 visitors monthly but each conversion is worth $50,000. Test it.
Mistake: Not tracking micro-conversions. Signing up for a trial is the ultimate goal, but tracking email signups, demo requests, and content downloads shows the full journey.
━━━━━━━━━━━━━━━━━━━━━━
💡 QUICK WIN
Open your analytics right now. Map your conversion funnel. Calculate conversion rate at each step. Whichever step has the biggest drop-off is where you start testing.
━━━━━━━━━━━━━━━━━━━━━━
Coming up in Part 6:
Technical analysis: catching bugs before they kill your conversions.
Reply with questions anytime.
– Atticus
P.S. If your analytics setup is broken or you’re not tracking key events, fix that before running any tests. Bad data leads to bad decisions.