In marketing, intuition is often wrong. A headline you think is "better" may convert worse. A/B tests (A/B testing or split testing) give you the answer with real data: you show two or more variants to different users and keep the one that performs best.
What are A/B tests?
An A/B test is a controlled experiment: variant A is shown to one part of your audience, variant B to another, and you measure which one achieves the better result (conversions, clicks, recoveries). Decisions are based on evidence rather than guesswork.
Where you can run A/B tests in iSend
iSend has A/B testing built into three areas:
1. Popups (contact collection)
Test multiple variants of the same popup at the same time: headline, colors, CTA button, offer or success message. Features:
- Unlimited variants per popup
- Weighted traffic allocation (adjustable weights from 1 to 100%)
- Impressions, submissions and conversion rate in real time, per variant
- Auto-winner: automatic winner selection based on a data threshold
- Sticky variants: each visitor always sees the same variant
2. Abandoned carts (recovery)
In abandoned cart recovery, you create variants of the recovery message and the platform finds out which one recovers more. You configure:
- Variant name plus the template for the 1st and 2nd reminder
- Send delay (15 to 1440 minutes)
- Interval between reminders (1 to 168 hours)
- Allocation weight (1 to 100)
- Auto-winner threshold plus statistics reset
The platform serves the variants with weighted random allocation and tracks events: sent, clicked, recovered, so it can automatically surface the winner based on actual recoveries.
3. SMS and Viber campaigns
In campaigns you can test different elements of the message: the copy, the call to action (CTA), the send time and even the Sender ID, and keep whatever performs best for your next sends.
How the mechanism works
- Weighted allocation: you define what share of the audience sees each variant.
- Real-time metrics: impressions, conversions/recoveries and rates per variant.
- Auto-winner: once enough data has been collected (threshold), the winning variant is selected automatically.
- Consistency: the same user always sees the same variant.
What is worth testing
- Headline / message opening: often the biggest difference.
- Offer: 10% off vs free shipping vs a gift.
- CTA: "Buy now" vs "Complete your order".
- Send time: time/day of sending, recovery delay.
- Sender ID: which Sender ID inspires more trust.
Best practices
- One variable at a time: so you know what caused the difference.
- Enough data: do not draw conclusions from a handful of impressions.
- A clear metric: define up front what "winning" means (conversion, recovery rate).
- Continuous improvement: small gains compound into big wins.
Conclusion
A/B tests turn marketing from guesswork into science. With iSend's built-in A/B tests for popups, carts and campaigns, you optimize continuously, with automatic winner selection and no manual calculations.
Frequently asked questions
Optimize with data, not guesswork
See iSend's A/B tests for popups, carts and campaigns.