AB testing allows you to earn more from the same amount of traffic.
Example:
Both stores receive 50,000 monthly visitors
Both have $50 Average Order Value (AOV)
Store A converts at 2.0% → $50,000 revenue
Store B converts at 2.2% → $55,000 revenue
A 10% lift in conversion rate = $5,000 more per month, without spending more on ads.
Quick Impact Checklist
Answer these questions for your store:
What is your current conversion rate?
What is your current monthly traffic?
What is your current AOV?
What does a 10% lift mean in monthly dollars?
Formulating a High-Quality Hypothesis
A strong hypothesis is the foundation of effective AB testing.
For example: Because we observed [problem], we believe that [change] will lead to [behavior shift], resulting in [metric lift].
Example template
Because the PDP has a low add-to-cart rate and session replays show hesitation in the buy box, we believe simplifying the buy box will reduce cognitive load, increasing the add-to-cart rate.
Steps to Create a High-Quality Hypothesis
Step 1 - Identify High-Value Pages
Focus where traffic is highest:
Most visited product pages (PDPs)
Top landing pages
Popular collection pages
These are most likely to drive meaningful results.
Step 2 - Identify the Problem
Use data to pinpoint what isn’t working:
Shopify or Google Analytics
Heatmaps
Scroll depth
User surveys
Session recordings
Common problems:
Low add-to-cart rate
High landing page bounce rate
Low discovery on collection pages
Low click-through to variants
Poor engagement with images or reviews
Step 3 - Identify the Friction Behind the Problem
Every bad metric has a reason.
Examples:
Value unclear → hero section is confusing
Users hesitate → buy box is overwhelming
Images too small → details not visible
Variant selector confusing
Your job: explain the why.
Step 4 - Connect Friction to the Change
Fix the root cause, not the symptom.
Examples:
If the hero messaging is clearer → more users read on
If reviews are higher on the page → more trust → more add-to-carts
If images are larger → more inspection → more purchases
Now plug it into the hypothesis formula.
High-Leverage vs. Low-Leverage Tests
Not all tests drive meaningful outcomes. Focus on big impact.
Low-Leverage (1–3% lift)
Minor tweaks:
Button colour
Small spacing changes
Tiny copy adjustments
Icon changes
Useful occasionally — but rarely move revenue.
High-Leverage (Up to 20% lift)
Big changes:
New PDP layout
New section above add-to-cart
Better image strategy (higher quality, lifestyle shots)
Simplified variant selector
Before/after visuals
Clear shipping info
Spend 80% of your time here.
Common Mistakes to Avoid
1. Calling Tests Too Early
Run at least 7 days. You need enough traffic to reach statistical confidence.
2. Testing Tiny Changes
Avoid “pixel-pushing”. Focus on high-impact ideas.
3. Testing Low-Traffic Pages
Only test pages with ≥ 1,000 sessions per month. If you have less, start with the highest-traffic page.
4. No Hypothesis
Never test without a clear hypothesis. You need:
Problem
Behavior change
Metric
Belief
5. Running Too Many Tests
Start with one high-impact test at a time unless you have very high traffic.
6. Ignoring Traffic Quality
Avoid testing during:
Ads spikes
Influencer boosts
Promotions
Ensure stable traffic for clean results.
Repeatable AB Testing Workflow
Use this monthly.
Identify top 5 pages by traffic
Find the biggest friction point on each
Write a hypothesis for each
Pick the top 1–2 to test
Build the variant (Instant = no code)
Run for 7–14 days
Roll out winning variants across similar pages
Repeat monthly
This cycle compounds results over time.