# Group Categories & Best Practices

### Standard Group Categories

#### Base

**What to Include:**

* Intercept/constant term
* Long-term time trends
* Structural baseline effects

**Purpose:** Represents the KPI level when all marketing is zero - your organic baseline performance

**Typical Contribution:** Usually the largest single group, showing stable baseline demand

**Best Practices:**

* Always include the constant term in Base
* Keep Base separate from other groups
* Helps identify incremental marketing impact

#### Media/Marketing

**What to Include:**

* All advertising spend channels
* TV, Radio, Display, Video, OOH
* Any paid media campaigns

**Purpose:** Shows total marketing contribution to KPI

**Subgroups Option:**

* Can split into "Traditional Media" and "Digital Media"
* Or separate by channel: TV, Digital, Radio, etc.

**Best Practices:**

* Group all marketing together OR separate into meaningful subgroups
* Don't mix media with promotions or price
* Enables ROI calculation by channel

#### Digital (Optional)

**What to Include:**

* Online marketing channels
* Search (Paid Search, SEO)
* Social Media
* Display Advertising
* Video (YouTube, etc.)
* Email Marketing

**Purpose:** Separate digital from traditional for strategic comparison

**When to Use:**

* Different teams manage digital vs. traditional
* Budget shifting between online/offline
* Digital transformation initiatives

#### Price

**What to Include:**

* Price variables
* Price indices
* Discount rates
* Price elasticity variables

**Purpose:** Measure price sensitivity and revenue impact of pricing decisions

**Best Practices:**

* Keep price separate from promotions
* Enables price elasticity calculation
* Critical for pricing strategy

#### Promotions

**What to Include:**

* Promotional campaigns
* Sales events (Black Friday, etc.)
* Coupon/discount promotions
* Trade promotions
* BOGO offers

**Purpose:** Quantify promotional lift and effectiveness

**Best Practices:**

* Separate from baseline pricing
* Distinguish trade vs. consumer promotions if relevant
* Helps evaluate promotional ROI

#### Seasonality

**What to Include:**

* Holiday indicators
* Monthly/quarterly dummy variables
* Seasonal indices
* Weather effects

**Purpose:** Isolate seasonal patterns from marketing effects

**Best Practices:**

* Clearly separate from marketing
* Helps understand natural demand cycles
* Critical for forecasting

#### External/Market

**What to Include:**

* Competitor activity
* Economic indicators (GDP, unemployment)
* Market trends
* Category growth

**Purpose:** Control for factors outside your influence

**Best Practices:**

* Separate controllable from uncontrollable factors
* Provides market context
* Helps justify marketing performance

#### Distribution

**What to Include:**

* Store count
* Distribution expansion
* Product availability
* Geographic coverage

**Purpose:** Measure impact of distribution on sales

**Best Practices:**

* Important for CPG and retail
* Separate from marketing
* Helps evaluate expansion impact

### Industry-Specific Best Practices

#### CPG/Retail

**Recommended Groups:**

```
- Base
- National TV
- Local TV
- Digital
- Traditional Media (Radio, Print, OOH)
- Trade Promotions
- Consumer Promotions
- Price
- Distribution
- Seasonality
```

**Key Considerations:**

* Separate trade vs. consumer promotions
* Distribution is critical
* Seasonal patterns are strong

#### E-Commerce/DTC

**Recommended Groups:**

```
- Base
- Search (Brand + Non-Brand)
- Social Media
- Display (Prospecting + Retargeting)
- Video
- Email/Owned
- Affiliates
- Promotions
- Seasonality
```

**Key Considerations:**

* Digital-heavy, granular channel grouping
* Separate prospecting vs. retargeting
* Email/owned channels important

#### B2B/SaaS

**Recommended Groups:**

```
- Base
- Paid Search
- LinkedIn/Social
- Content Marketing
- Events/Webinars
- Display/Retargeting
- Organic/SEO
- Price/Packaging
- External (Market/Competition)
```

**Key Considerations:**

* Longer sales cycles
* Content and events matter
* ABM vs. demand gen separation

### Balancing Granularity

#### Too Few Groups

**Signs You Need More Groups:**

* Can't make actionable decisions
* Too many channels lumped together
* Stakeholders ask for more detail
* Can't identify best-performing channels

**Solution:** Split major groups into sub-groups

#### Too Many Groups

**Signs You Need Fewer Groups:**

* Cluttered, unreadable charts
* Groups with minimal contribution
* Difficult to see patterns
* Too many colors to distinguish

**Solution:** Consolidate small groups, use drill-down instead

#### Optimal Balance

**Aim for 5-10 groups:**

* Detailed enough for decisions
* Simple enough to communicate
* Easy to visualize
* Clear patterns emerge

**Use hierarchy:**

* High-level for executives (5-7 groups)
* Mid-level for marketing leads (7-10 groups)
* Detailed drill-down for analysts (variable-level)

### Best Practices by Use Case

#### Budget Allocation

**Group Strategy:**

* Separate controllable spend categories
* One group per budget line item
* Match finance reporting structure

**Example:**

```
- Brand Marketing (TV, Display, Sponsorships)
- Performance Marketing (Search, Social, Affiliates)
- Promotions
- Price Investments
```

**Enables:** Clear ROI by budget category, reallocation decisions

#### Executive Reporting

**Group Strategy:**

* High-level business categories
* Align with strategic priorities
* Simple, clear story

**Example:**

```
- Marketing
- Pricing
- Promotions
- External Factors
```

**Enables:** Strategic conversation, board-level reporting

#### Channel Optimization

**Group Strategy:**

* Separate each major channel
* Enable channel-to-channel comparison
* Detail where budget decisions happen

**Example:**

```
- TV
- Digital Search
- Digital Social
- Digital Display
- Radio
- Print
- Out-of-Home
- Promotions
```

**Enables:** Granular ROI comparison, tactical budget shifts

#### Customer Journey Analysis

**Group Strategy:**

* Organize by funnel stage
* Track customer progression
* Identify journey gaps

**Example:**

```
- Awareness (TV, Display, Video)
- Consideration (Content, Social, Email)
- Conversion (Search, Retargeting)
- Retention (Loyalty, Email)
```

**Enables:** Full-funnel optimization, journey mapping

### Grouping Workflow

#### 1. Initial Categorization

**Start with broad categories:**

* Base/Baseline
* Marketing
* Price/Promotions
* Seasonality
* External

**Then refine based on needs:**

* Split Marketing into channels
* Separate Price from Promotions
* Add industry-specific groups

#### 2. Stakeholder Alignment

**Involve key stakeholders:**

* Marketing leadership (channel grouping)
* Finance (budget alignment)
* Analytics team (technical feasibility)
* Executives (strategic alignment)

**Questions to Ask:**

* How do you think about marketing performance?
* What decisions will this analysis inform?
* How are budgets structured?
* What metrics do you track?

#### 3. Test and Validate

**Run decomposition with initial grouping:**

* Do contributions make sense?
* Are patterns clear?
* Can you tell the story?

**Adjust if needed:**

* Combine small groups
* Split large heterogeneous groups
* Rename for clarity

#### 4. Document and Standardize

**Create documentation:**

* Group definitions
* Variables included in each group
* Rationale for grouping decisions
* When to use this structure

**Standardize across models:**

* Use same groups for similar analyses
* Maintain consistency over time
* Enable year-over-year comparison

### Common Grouping Challenges

#### Challenge 1: Multi-Function Variables

**Problem:** Variable serves multiple purposes

* Example: Social media drives both awareness and conversion

**Solutions:**

* **Primary purpose:** Assign to most relevant group (Awareness if primarily brand)
* **Split if possible:** Use platform data to separate brand vs. performance spend
* **Hybrid group:** Create "Hybrid Channels" group
* **Document:** Note the multi-function nature

#### Challenge 2: Seasonal Marketing

**Problem:** Marketing that only runs seasonally

* Example: Christmas TV campaign

**Solutions:**

* **Group with channel:** Christmas TV → TV group
* **Separate seasonal:** Create "Holiday Marketing" group if significant
* **Best practice:** Group by channel, use time analysis to see seasonal patterns

#### Challenge 3: Test Campaigns

**Problem:** Short-term tests vs. always-on campaigns

**Solutions:**

* **Group with channel:** Test campaigns → same group as regular channel
* **Separate if analyzing:** Create "Test" group if evaluating test effectiveness
* **Typical:** Include with main channel for standard reporting

#### Challenge 4: Organic vs. Paid

**Problem:** Mix of paid and owned channels

**Solutions:**

* **Separate groups:** Paid Media vs. Owned Media
* **Within digital:** Paid Search vs. Organic Search as separate groups
* **Depends on analysis:** Separate if optimizing paid/owned mix

#### Challenge 5: Cross-Channel Campaigns

**Problem:** Integrated campaigns across multiple channels

**Solutions:**

* **Individual channels:** Assign each channel separately
* **Campaign group:** Create campaign-specific group if measuring integrated effect
* **Best practice:** Use individual channels, analyze campaign timing separately

### Advanced Grouping Techniques

#### Hierarchical Grouping

**Create multiple levels:**

**Level 1 (Executive):**

```
- Marketing
- Non-Marketing
```

**Level 2 (Marketing Leadership):**

```
- Traditional Media
- Digital Media
- Promotions
```

**Level 3 (Channel Managers):**

```
- TV
- Radio
- Search
- Social
- Display
```

**Implementation:** Use different grouping configurations for different audiences

#### Dynamic Grouping

**Adjust groups based on time period:**

* **Pre-digital era:** Fewer digital groups
* **Digital transformation:** More digital granularity
* **Current period:** Balance based on spend mix

**Use case:** Historical analysis with changing channel mix

#### Scenario-Based Grouping

**Different groupings for different questions:**

**Question: "What's the ROI of our media?"** → Group: All Media together

**Question: "Which channel performs best?"** → Group: Separate each channel

**Question: "Digital vs Traditional?"** → Group: Two major groups

**Flexibility:** Change grouping based on analytical goal

### Grouping Checklist

Use this checklist to validate your grouping:

**Completeness:**

* \[ ] All variables assigned to a group
* \[ ] Constant term in Base group
* \[ ] No variables orphaned

**Clarity:**

* \[ ] Group names are clear and descriptive
* \[ ] Names are consistent across models
* \[ ] Business stakeholders understand groups

**Logic:**

* \[ ] Variables in same group are logically related
* \[ ] Groups match business structure
* \[ ] Grouping enables key decisions

**Balance:**

* \[ ] 5-10 groups (optimal range)
* \[ ] No groups with only 1-2 variables (unless strategically important)
* \[ ] No single group dominating with 20+ variables

**Actionability:**

* \[ ] Groups align with budget allocation
* \[ ] Enable channel comparison
* \[ ] Support optimization decisions

**Communication:**

* \[ ] Easy to visualize
* \[ ] Clear color assignment
* \[ ] Enables storytelling

### Examples of Effective Grouping

#### Example 1: National Brand Launch

**Context:** Launching new product nationally

**Groups:**

```
Base: Constant, Pre-launch trend
National TV: All national TV campaigns
Local TV: Regional TV support
Digital Brand: Display, Video, Social (brand)
Digital Performance: Search, Retargeting (conversion)
Retail Support: Trade promotions, POS
Launch Events: Special events, sampling
```

**Rationale:** Separates launch investment types, enables ROI by tactic

#### Example 2: E-Commerce Growth

**Context:** Scaling e-commerce business

**Groups:**

```
Base: Organic baseline
Upper Funnel: Display Prospecting, Video, Social Brand
Mid Funnel: Content, Social Engagement
Lower Funnel: Search, Retargeting, Shopping Ads
Owned: Email, SEO, Affiliates
Promotions: Discounts, Free Shipping
```

**Rationale:** Funnel-based for customer journey optimization

#### Example 3: Retail Chain Expansion

**Context:** Opening new stores while maintaining existing

**Groups:**

```
Base: Existing store baseline
National Media: TV, Radio (all markets)
Existing Markets: Marketing in current locations
New Markets: Marketing in expansion areas
Promotions: Sales events, coupons
Distribution: New store openings
Price: Pricing strategy
```

**Rationale:** Separates expansion from maintenance, measures new market investment

### Tips for Success

**Start Simple:**

* Begin with 5-7 broad groups
* Refine based on stakeholder feedback
* Don't overthink initial setup

**Stay Flexible:**

* Grouping isn't permanent
* Adjust as business evolves
* Different groupings for different questions

**Focus on Decisions:**

* Group to enable actions
* Match budget structure
* Support strategic choices

**Test with Real Data:**

* Run decomposition early
* Check if groups tell clear story
* Iterate based on insights

**Document Everything:**

* Record grouping decisions
* Explain rationale
* Share with team

**Get Buy-In:**

* Involve stakeholders early
* Align with business language
* Ensure groups make sense to decision-makers

### Next Steps

After establishing group categories:

* Proceed to **Color Assignment Strategy**
* Configure **Adjustment Parameters**
* Run **Decomposition Analysis**
* Analyze **Group-Level Performance**
