Loading tutorials…
Loading tutorials…
Smart Traffic is Unbounce's machine-learning router — it analyzes visitor attributes and sends each one to the variant most likely to convert THEM. Typically lifts CR 10-30% vs even A/B testing. The catch: it needs enough traffic to learn.
Who this is forMarketers on Unbounce Experiment plan or higher running landing pages with 1,000+ visitors/month per variant. If your page has under 500 visitors/month, Smart Traffic won't have data to learn from.
What you'll need
Step 1
Smart Traffic is Experiment plan and above. Build plan does NOT include it.
Check Settings → Billing → current plan.
If Build: you cannot use Smart Traffic. Either upgrade to Experiment ($149/mo annual) or stick with manual A/B testing.
If Experiment or higher: Smart Traffic is available on all pages with 2+ variants.
Smart Traffic uses ~50 visitor attributes (browser, device, OS, location, time of day, referrer, etc.) to predict which variant will convert each visitor.
It does NOT use personally identifiable info (no email, name). Privacy-safe.
Step 2
Smart Traffic needs distinct variants. 'Same page, different button color' = no signal. Variants should appeal to different segments.
Don't use Smart Traffic with trivial variants (button color, hero image). It has nothing to learn from.
Design variants for different audiences. Examples:
- Variant A: 'Get 10x leads' (volume-focused) — appeals to growth-stage SMBs.
- Variant B: 'Stop guessing — predict your pipeline' (clarity-focused) — appeals to data-driven mid-market.
- Variant C: 'Trusted by 12,000 marketers' (social-proof-focused) — appeals to risk-averse enterprise.
Smart Traffic learns which variant converts which type of visitor. A growth-stage SMB might consistently route to Variant A; enterprise to Variant C.
2-4 variants is the sweet spot. More = each gets too little traffic to learn from.
Step 3
Pages → click page → A/B Test panel → toggle Smart Traffic ON.
Open the page in Unbounce. Click A/B Test panel (top right).
If you have 2+ variants: see the Smart Traffic toggle. Click ON.
Smart Traffic enters 'Learning Period.' During this time, it routes traffic somewhat randomly to gather data on each variant + visitor type.
Learning period: ~50-200 conversions, depending on traffic volume. Typically 7-21 days.
After learning: Smart Traffic switches to 'Optimizing.' It now routes traffic intelligently — high-probability visitors to the variant most likely to convert them.
Monitor: A/B Test panel shows current state (Learning / Optimizing) + per-variant performance.
Step 4
Smart Traffic optimizes against ONE primary conversion. Pick the one that matters: form submit, purchase, signup.
Page Settings → Conversion Goal → primary goal.
Common goals: form submission (lead gen pages), purchase (ecommerce pages), button click (free trial signups), page view (engagement pages — rare).
Smart Traffic uses this single goal to determine 'best variant for this visitor.' Pick the goal that matters financially.
If you have a multi-step funnel (Page A → Page B → Purchase), set Smart Traffic on Page A optimizing for click-through to Page B. Page B optimizes for purchase.
Step 5
First 7-21 days: Smart Traffic is learning. Don't change variants or stop the test. Monitor for stability.
A/B Test panel → status: 'Learning' for first 50-200 conversions.
During learning: variant performance numbers may swing day-to-day. Don't panic. Smart Traffic is still gathering data.
Check: total conversions/day stable or growing. Per-variant traffic share roughly even (50/50 or 33/33/33 typical during learning).
After 50+ conversions per variant: status changes to 'Optimizing.' Smart Traffic now actively routes for highest-probability conversion.
Per-variant traffic share may shift: e.g., 70% to Variant A, 30% to Variant B if Variant A is winning for the typical visitor.
Per-variant CR: Smart Traffic targets ~equal CR per variant by sending the right visitor to each. So Variant A and B both end up at ~5% CR, vs random distribution where Variant A might be 7% and Variant B 3%.
Step 6
After Smart Traffic has been Optimizing for 30 days, compare against pre-Smart-Traffic baseline.
Before/After comparison: pull conversion data for 30 days BEFORE you enabled Smart Traffic vs 30 days AFTER.
Compare overall page CR (across all variants). Typical lift: 10-30%.
If lift is positive and > 10%: Smart Traffic is working. Keep it on.
If lift is < 5% or negative: Smart Traffic isn't helping much for this page. Common causes: variants too similar (no segmentation signal), not enough traffic (under 1,000 visits/mo per variant), wrong conversion goal.
Iterate: add more distinct variants, increase traffic to the page, or revise the conversion goal.
Step 7
Smart Traffic isn't set-and-forget. Refresh variants quarterly to keep CR climbing.
Quarterly: review per-variant performance + traffic share.
Identify the worst-performing variant. Replace with a new variant testing a different angle.
Smart Traffic starts a new short learning period on the new variant, then resumes optimizing.
Annual: re-evaluate entire variant set. Audience may have shifted (new product fit, new customer segment). Refresh accordingly.
Document each variant's hypothesis in your testing log. Even Smart Traffic variants benefit from clear hypotheses.
Common mistakes
Enabling Smart Traffic on a page with under 500 visitors/month
What goes wrong: Smart Traffic needs data to learn. Under 500 visits/month per variant = it never exits learning mode. Doesn't make routing decisions. No lift.
How to avoid: Drive traffic to 1,000+ visits/month per variant before enabling. Or stick with manual A/B testing on lower-traffic pages.
Designing similar variants (different button color)
What goes wrong: Smart Traffic learns to route based on visitor attributes. If variants don't differ meaningfully, it can't learn anything useful. No CR lift over manual A/B.
How to avoid: Design variants for different audiences with different headlines, value props, social proof. Smart Traffic shines when variants are distinctly different.
Wrong primary conversion goal
What goes wrong: Optimize for 'button click' when actual revenue is 'form submission.' Smart Traffic routes for clicks, not for leads. Engagement up, revenue flat.
How to avoid: Set primary goal to the conversion that drives revenue: form submit, purchase, signup. Optimize for what matters financially.
Stopping Smart Traffic during learning period
What goes wrong: Marketer sees uneven results during learning, panics, disables Smart Traffic. Misses the optimizing phase. Never realizes the CR lift.
How to avoid: Patience. Learning period is 50-200 conversions. Variant performance swings are normal during learning. Trust the system.
Running A/B tests AND Smart Traffic simultaneously
What goes wrong: Smart Traffic biases the 'A/B test' by routing high-converting visitors to one variant. Test results are invalid.
How to avoid: Disable Smart Traffic during manual A/B tests. Re-enable after declaring a winner.
Never refreshing variants
What goes wrong: Smart Traffic optimizes existing variants but doesn't suggest new ones. After 6-12 months, CR plateaus because variants are stale.
How to avoid: Quarterly variant refresh: replace worst-performing variant with a new hypothesis. Keep the testing iteration going.
Recap
Done — what's next
How to build a high-converting Unbounce landing page
Read the next tutorial
Hand it off
Smart Traffic is Unbounce's killer feature — most accounts don't use it because variant design strategy isn't intuitive. A CRO specialist at $14-16/hr can design + launch a proper Smart Traffic setup with 3-4 audience-targeted variants in 4-8 hours, typically $100-300, lifting CR 10-30%.
See specialist rates
A/B testing: find one winning variant, deploy to all visitors. Smart Traffic: route each visitor to the variant most likely to convert THEM. Smart Traffic typically lifts CR 10-30% beyond pure A/B because different visitors prefer different variants.
50-200 conversions (across all variants). On a page with 100 conversions/month, that's 1-2 months. On a page with 500/mo, it's 1-2 weeks. Patience is required.
No. Smart Traffic is Experiment plan and above ($149/mo annual). Build plan supports manual A/B testing only.
Yes — it's traffic-source agnostic. Smart Traffic learns based on visitor attributes (browser, device, location, time, etc.), not on which ad platform sent the visitor. Works with any traffic source.
Yes if both are running. Smart Traffic biases visitor routing, which invalidates A/B test statistical assumptions. Use one OR the other on a given page at a given time.
Unbounce
Unbounce's drag-and-drop builder makes shipping a landing page in 3 hours feel doable. The harder question: does the page actually convert? Most DIY pages launch at 1-3% CR. Specialists routinely ship pages at 8-15%.
Unbounce
Unbounce makes A/B testing easy — too easy. Most marketers ship 'A/B tests' that lack statistical significance, test trivial changes, and conclude prematurely. Here's how to test like a CRO professional.
Unbounce
Dynamic Text Replacement matches what visitors searched for to what they see on your landing page. Done right, it lifts Quality Score (lowering Google Ads CPC by 20-40%) AND lifts CR by 10-20%. Done wrong, it shows broken placeholder text and tanks both.
Unbounce
Unbounce can pay back its $99-625/mo fee 5-20x with the right setup. Or break even with mediocre setup. The gap is where specialists earn their fee.
Google Ads
Conversions firing in your store but Google reporting zero? You're not alone. This walks through the right setup path for your stack — including the GA4 import workflow that 80% of DIY tutorials skip.