Best Practices for A/B Testing

Documentation Index

Fetch the complete documentation index at: https://docs.mention-me.com/llms.txt

Use this file to discover all available pages before exploring further.

​

Planning your A/B test

​

Test one thing at a time

Focus on a single variable — imagery, copy, or incentive type — to ensure you can isolate the cause of any performance changes.

• 7-day vs. 14-day validity period
• Branded vs. lifestyle imagery
• Avoid changing both incentives and visuals at once

​

Prioritise what to test

• Impact: Will the change likely drive major results?
• Ease: Is it quick and simple to implement?
• Volume: Do you have enough traffic to support a test?

​

Always be testing

Treat A/B testing as an ongoing optimisation process. Continually test and iterate to find new insights.

​

Promote your referral programme widely

• Ensure both test variations get equal exposure
• Include referral promotions in emails, homepage, and post-purchase flows

​

Common pitfalls to avoid

• Avoid testing too many things at once
• Don’t stop the test too early
• Act only on results from completed tests
• Don’t cherry-pick data — look at the full picture

​

Test templates

The templates below are sorted by effort and impact level. Use them to identify what drives the greatest performance uplift.

​

Reduce share methods (lower effort)

Test name + native + link vs. offering all share methods. Result: +43% Share Rate

​

Remove Meta sharing (lower effort)

Remove Facebook and Messenger in favour of higher-performing options. Result: +14% Share Rate, +18% Conversion Rate

​

Add WhatsApp sharing (lower effort)

Add WhatsApp to the mobile share options. Result: +24.5% Share Rate

​

Reward framing: Give vs. Get (lower effort)

Test CTA: “Give 20% Off” vs. “Get 20% Off”. Result: +12% Share Rate when “Give” used

​

Add brand USPs (lower effort)

Highlight brand strengths to add motivation. Requires copy changes on referrer page, corner peel, and messaging.

​

Lifestyle vs. product imagery (higher impact)

Lifestyle (social scenes) vs. product-only images. Result: +21% Share Rate and +5% Conversion Rate with product focus

​

Lifestyle vs. reward focus imagery (higher impact)

Ideal for service-led brands.

​

Money vs. percentage off (higher impact)

Example: £20 off vs. 55% discount (min £30 spend). Result: +20% Share Rate and +100% Purchase Rate for % discount

​

Reducing reward value (higher impact)

Test reward value sensitivity to optimise ROI.

​

Adjusting minimum spend (lower effort)

Test with vs. without minimum spend requirement. Result: +45% Share Rate when removed

​

Shorten offer validity (lower effort)

Increase urgency by reducing offer duration. Result: 3-day expiry vs. 7-day — +28% Purchase Rate

​

Extend offer validity (lower effort)

Beneficial for high-consideration purchases.

​

Referral best practice (when A/B testing isn’t feasible)

Even without formal A/B testing, you can analyse and optimise your programme using funnel data, benchmarking, and engagement metrics.

​

Understand the referral funnel

1. Customer purchases and sees referral prompt
2. Refers a friend
3. Friend clicks the link and visits your referral page
4. Friend registers to redeem the offer
5. Friend completes a purchase
6. Referrer receives the reward

​

Optimise using benchmarking and Explore

Key metrics to monitor:

• Impressions per order
• Enrolment rate
• Share rate
• Shares per sharer
• Responses per share
• Incented per response
• Purchases per incented

Explore section analysis dimensions:

• Promotion point (e.g. homepage vs. post-purchase)
• Share method (WhatsApp, Name Share, etc.)
• Audience segment
• Device (mobile vs. desktop)

​

Practical optimisation tips

1. Maximise promotion points

• Use homepage banners, corner peels, and email placement to ensure visibility
• Email campaign examples:
  • +107% Share Rate from Solus email launch
  • +336% Referrals from Valentine’s Day promo
  • 50% Open Rate with targeted reminder email

2. Improve referrer experience

• Enable native share buttons
• Allow message personalisation
• Encourage multiple shares per referrer

3. Improve referee experience

• Use social proof (e.g. Trustpilot)
• Highlight key USPs at every touchpoint
• Offer exploration CTAs
• Test offer duration for urgency impact

4. Enhance tracking

• Use “situation” tags to track performance by placement
• Run cohort or device-type analysis
• Link referral journeys to customer value outcomes