# A/B testing for festivals - groups and progressive sends

<Schema>{`{
  "@context": "https://schema.org",
  "@type": "HowTo",
  "name": "How to Run A/B Testing with Groups in Nevent",
  "description": "Guide to splitting attendee segments into groups for A/B testing of ticket sale campaigns and progressive sends",
  "totalTime": "PT48H",
  "tool": {
    "@type": "SoftwareApplication",
    "name": "Nevent",
    "applicationCategory": "Marketing Automation"
  },
  "step": [
    {
      "@type": "HowToStep",
      "position": 1,
      "name": "Define the Test Objective",
      "text": "Decide what you want to test: subject lines, discounts, offers, progressive sends for deliverability, or price testing. For example: discovering which subject line generates the best open rate."
    },
    {
      "@type": "HowToStep",
      "position": 2,
      "name": "Create Your Base Segment",
      "text": "Create the attendee segment you want to split. For example: Fans from Barcelona (8,000 fans). Verify that the size is sufficient for statistically significant results (minimum 100+ fans per group)."
    },
    {
      "@type": "HowToStep",
      "position": 3,
      "name": "Split into Groups",
      "text": "Select how many groups you need: 2 groups for classic A/B testing (50% each), 3-5 groups for multi-variant testing, 10+ groups for progressive sends. Nevent distributes fans automatically in a balanced way."
    },
    {
      "@type": "HowToStep",
      "position": 4,
      "name": "Assign Variants to Each Group",
      "text": "Assign different versions to each group. Group 1: Subject with emoji and 20% discount. Group 2: Subject without emoji with 20% discount. Important: change only ONE variable at a time for clear results."
    },
    {
      "@type": "HowToStep",
      "position": 5,
      "name": "Run the Campaigns",
      "text": "Send the campaigns simultaneously to all groups (for A/B testing) or progressively (Day 1: Group 1, Day 2: Group 2, etc.). For progressive sends, start with 10% of your base to catch problems early."
    },
    {
      "@type": "HowToStep",
      "position": 6,
      "name": "Wait and Measure Results",
      "text": "Wait 24-48 hours for email A/B tests, 7-14 days for conversion tests. Measure open rate, click rate, conversion and ROI per group."
    },
    {
      "@type": "HowToStep",
      "position": 7,
      "name": "Make a Decision and Scale",
      "text": "Identify the winning group (best performance on your target metric) and apply that variant to future campaigns. If it was a pilot test with 10%, now send the winning version to the remaining 90%."
    }
  ]
}`}</Schema>

<Schema>{`{
  "@context": "https://schema.org",
  "@type": "BreadcrumbList",
  "itemListElement": [
    {
      "@type": "ListItem",
      "position": 1,
      "name": "Home",
      "item": "https://docs.nevent.com"
    },
    {
      "@type": "ListItem",
      "position": 2,
      "name": "Segmentation Engine",
      "item": "https://docs.nevent.com/en/segmentation/segmentation-engine"
    },
    {
      "@type": "ListItem",
      "position": 3,
      "name": "Groups & A/B Testing",
      "item": "https://docs.nevent.com/en/segmentation/segmentation-engine/groups"
    }
  ]
}`}</Schema>

:::tip[Quick definition]
**What are groups used for in segmentation?**

An A/B testing function that splits a segment into equal random parts to test variants (subject lines, prices, offers). Requires a minimum of 1,000 contacts per group for statistical significance.

**Example:** Sónar split 20k fans into 2 groups to test subjects. "⏰ Last 48 hours" achieved 34% open vs "Early Bird available" 22% open. The winner generated +12pp × 10k fans = 1,200 extra opens.
:::

# Groups & A/B Testing

A/B testing for festival campaigns: the Groups function lets you split attendees into equal parts for testing subject lines, early bird offers and progressive sends. Optimise ticket sale conversion with scientific campaign testing.

---

## What Are Groups Used For?

### 1. A/B Testing

Split your audience to test different versions and learn what works best.

**Example:**
```
Segment: Fans from Madrid aged 25-40 (10,000 fans)
Split into: 2 groups

Group 1 (5,000 fans): Email with 10% discount
Group 2 (5,000 fans): Email with 15% discount

→ Measure which generates better ROI
```

### 2. Progressive Sends (Deliverability)

Improve your sender reputation by sending in small batches.

**Example:**
```
Segment: 20,000 fans
Split into: 10 groups (2,000 fans each)

Day 1: Send to Group 1
Day 2: Send to Group 2
Day 3: Send to Group 3
...

→ Avoid being marked as spam
```
**Tip:** Email providers (Gmail, Outlook) look favourably on progressive sends vs instant mass sends of 50K emails.

### 3. Offer Testing

Test different incentives with small groups before launching at scale.

**Example:**
```
Segment: 5,000 VIP fans
Split into: 5 groups (1,000 fans each)

Group 1: 2-for-1 on tickets
Group 2: 25% discount
Group 3: Free upgrade to VIP
Group 4: Free parking
Group 5: No offer (control group)

→ Identify the offer with the best conversion
```

---

## How Groups Work

### Key Characteristics

✅ **Consistent**: A fan will always be in the same group
✅ **Random**: Fans are distributed randomly
✅ **Balanced**: Each group has approximately the same size
✅ **Deterministic**: Based on a hash of the user_id (does not change over time)

**Example:**
```
Fan ID: user_12345
→ Will always be in Group 3 (of 10)
→ Today, tomorrow, next month... always Group 3
```

This is crucial for **consistency** in long-term A/B tests.

### Group Range

You can split your segment into **1 to 100 groups**.

### How many groups should I create for A/B testing festival campaigns?

| Groups | Size per Group | Typical Use |
|--------|---------------|-------------|
| 2 groups | 50% each | Classic A/B testing |
| 3-5 groups | 20-33% each | Multi-variant testing |
| 10 groups | 10% each | Weekly progressive sends |
| 20 groups | 5% each | Daily progressive sends |
| 100 groups | 1% each | Granular testing / very gradual rollouts |

---

## Practical Use Cases

### Case 1: A/B Test of Subject Lines

**Objective**: Discover which subject line generates the best open rate.

**Setup:**
```
Segment: Fans from Barcelona (8,000 fans)
Split into: 2 groups

Group 1 (4,000 fans): Subject "🔥 Last 24h: 20% discount"
Group 2 (4,000 fans): Subject "Do not miss it: Exclusive offer"
```

**Measurement:**
- Group 1: Open rate 42%, Click rate 12%
- Group 2: Open rate 31%, Click rate 9%

**Decision**: Group 1 subject is the winner → Use in future campaigns.

---

### Case 2: Progressive Send Over 10 Days

**Objective**: Send a campaign to 30,000 fans over 10 days for better deliverability.

**Setup:**
```
Segment: All active fans (30,000 fans)
Split into: 10 groups (3,000 fans each)

Day 1 (Monday): Send to Group 1
Day 2 (Tuesday): Send to Group 2
Day 3 (Wednesday): Send to Group 3
...
Day 10: Send to Group 10
```

**Benefits:**
- ✅ Better sender reputation
- ✅ Lower probability of spam folder
- ✅ You can adjust the message based on early results
- ✅ Reduces server load

---

### Case 3: Price Test Across 4 Variants

**Objective**: Find the optimal price for early bird.

**Setup:**
```
Segment: Fans with early bird purchase history (2,000 fans)
Split into: 4 groups (500 fans each)

Group 1: Price €50 (no discount)
Group 2: Price €45 (10% discount)
Group 3: Price €40 (20% discount)
Group 4: Price €35 (30% discount)
```

**Measurement (example):**
| Group | Price | Conversion | Total Revenue | Revenue/Fan |
|-------|-------|------------|---------------|-------------|
| 1 | €50 | 15% (75 sales) | €3,750 | €7.50 |
| 2 | €45 | 22% (110 sales) | €4,950 | **€9.90** ✅ |
| 3 | €40 | 28% (140 sales) | €5,600 | €11.20 |
| 4 | €35 | 35% (175 sales) | €6,125 | €12.25 |

**Decision**:
- If maximising revenue/fan: Group 4 (€35) is the winner
- If you need volume: Also Group 4
- If you want a profit/volume balance: Group 3 (€40)

---

### Case 4: Gradual Rollout of a New Feature

**Objective**: Launch a new mobile app feature gradually.

**Setup:**
```
Segment: Fans with the app downloaded (15,000 fans)
Split into: 20 groups (750 fans each)

Week 1: Enable feature for Groups 1-2 (10%)
Week 2: If no bugs → Groups 3-6 (30% cumulative)
Week 3: If all OK → Groups 7-15 (75% cumulative)
Week 4: Full rollout → Groups 16-20 (100%)
```

**Benefits:**
- ✅ You catch bugs with a small audience first
- ✅ You reduce the impact of critical issues
- ✅ You can iterate based on early adopter feedback

---

## Best Practices with Groups

### 1. Minimum Size per Group

For statistically significant results:

### What is the minimum audience size for A/B testing festival emails?

| Test Type | Minimum Recommended Size |
|-----------|--------------------------|
| Email A/B test | 100+ fans per variant |
| Price test | 50+ fans per variant |
| Landing page test | 200+ fans per variant |
| Progressive send | No minimum (depends on your base) |
**Caution:** With fewer than 50 fans per group, results may be statistically inconclusive (too much noise).

### 2. Test Duration

- **Email A/B test**: 24-48 hours (wait for opens to stabilise)
- **Conversion test**: 7-14 days (full purchase decision cycle)
- **Engagement test**: 30+ days (behavioural pattern)

### 3. Variables to Test (One at a Time)

❌ **Bad — You change multiple things:**
```
Group A: Subject "Offer", 10% discount, CTA "Buy", red image
Group B: Subject "Exclusive", 20% discount, CTA "Reserve", blue image
```
**Problem**: You will not know which variable caused the difference.

✅ **Good — You change ONE variable:**
```
Group A: Subject "Exclusive Offer"
Group B: Subject "Do Not Miss This"

(Everything else the same: discount, CTA, design)
```
**Result**: You know exactly which subject line works better.

### 4. Control Group

Always include a **control group** (no change) to measure the real impact.

**Example:**
```
Group 1: Email with 15% discount
Group 2: Email with 25% discount
Group 3 (CONTROL): Email without discount

→ You measure whether the discount actually increases conversion
```

---

## Frequently Asked Questions

### Do groups change over time?

No. A fan will always be in the same group (based on a hash of their user_id).

**Example**: If Fan X is in Group 5 today, they will still be in Group 5 in 6 months.

### Can I select multiple groups for a campaign?

Yes. You can send to:
- A single group: Group 1
- A range of groups: Groups 1-5
- Specific groups: Groups 1, 3, 7

### How do I split 13,456 fans into 10 groups?

Nevent does it automatically:
- Groups 1-6: ~1,346 fans each
- Groups 7-10: ~1,345 fans each

The difference is minimal (±1 fan).

### Can I see which fans are in each group?

Yes, in the segment preview you can filter by a specific group and view the list of fans.

---

## Next Steps

Now that you understand groups:

1. **[See Full Use Cases](./use-cases)** — Real examples with groups
2. **[Best Practices](./best-practices)** — Optimise your A/B tests
3. **[FAQ](./faq)** — Answers to common questions

---

**Ready for scientific testing?** 🧪

[Examples of A/B testing: subject lines, prices, offers with groups →](./use-cases)