“Test everything” is common advice, written for sites with large traffic. For a small business site with a few thousand visitors a month, a classic A/B test of a button colour would run for many months and still not give a clear answer.
How much traffic does a test need?
It depends on your current conversion rate and the size of improvement you want to detect. A common approximation (for 95% confidence and 80% power) is:
Visitors per variation ≈ 16 × p × (1 − p) ÷ (p × lift)²
where p is your conversion rate and lift is the relative improvement.
| Current conversion rate | Improvement to detect | Visitors per variation (approx.) |
|---|---|---|
| 2% | +50% (to 3%) | about 3,100 |
| 2% | +20% (to 2.4%) | about 19,600 |
| 5% | +20% (to 6%) | about 7,600 |
These are rough planning numbers; use a sample size calculator for your case. The point: small improvements need a lot of traffic. If you have 2,000 visitors a month, you can only reliably detect big differences.
The planner does this calculation.
What to do with low traffic
1. Test big changes, not tweaks
A new headline and offer, a different page structure, adding price information, a WhatsApp button in the first screen: changes that could plausibly move conversion by 30–50%+, not button colours.
2. Fix the obvious first
Many improvements don’t need testing: a broken form, a slow page, missing prices, no proof, contact options hidden on mobile. Fix them. See website conversion optimisation.
3. Before/after comparisons, carefully
Compare a period before and after a change. It’s weaker than an A/B test because other things change too (seasons, ads, offers), so:
- Use equal-length periods without promotions or holidays
- Change one major thing at a time
- Check the trend in leads or sales, not just conversion rate
4. Qualitative research
- Session recordings and heatmaps for a week: where do people hesitate, rage-click, or leave?
- Customer questions: what do people ask on WhatsApp before buying? That’s missing from the page.
- Five-person tests: ask five people who fit your audience to complete a task on your site while thinking aloud. You’ll find the biggest problems quickly.
5. Test messages in ads first
Ad platforms have far more impressions than your site has visitors. Test headlines and offers as ad variations; the winning message is a strong candidate for your landing page. See ad creative testing.
If you do run an A/B test
- Decide the sample size and duration before starting, and don’t stop early because one variation looks ahead
- Run full weeks to cover weekday patterns
- One primary metric (enquiries or purchases), not ten
- Make sure tracking records each conversion once
Frequently asked questions
Can we test with 500 visitors a month?
Not with classic A/B testing, except for very large differences. Use the qualitative methods and before/after comparisons.
What’s a good conversion rate to aim for?
It varies widely by business, traffic source and offer. Compare with your own history rather than general benchmarks.
Should we use a testing tool?
Only once traffic supports it. Until then, the time is better spent on research and obvious fixes.
If you’d like help finding and fixing what stops visitors converting, that’s part of our traffic funnel service.
Free download · Excel calculator
Test planner and sample size estimator
Estimate visitors needed per variation from your conversion rate and the improvement you hope to detect, then plan and log tests and their results.
Download itNo sign-up. Opens in Excel, Google Sheets or Numbers.
Rather have it done?
This is the kind of work our traffic funnel service covers. Tell us where you are today and we’ll reply with a plan and a fixed price.
Let’s talk