Glow Brand Measurement

Stage: How It Works

Target: CS, Platform Specialists, customers

Reading time: ~15 minutes

The Problem Glow Solves

Brand teams are set up to fail. They are measured on short-term metrics, spending to nurture the 95% of customers not yet in-market, but evaluated as if they should only convert the 5% who already are.

When budgets tighten, brand spend gets cut first — because nobody can prove it works. Traditional tools are blind to brand impact: GA4 credits the last click, quarterly MMMs are too slow, and platform dashboards self-report. Nobody can answer the fundamental question: "Is our brand investment actually driving business outcomes?"

What is Glow?

Glow is Fospha's causal measurement and brand analytics platform. It uses Bayesian Causal Inference to estimate the probability that upper-funnel brand activities are driving leading indicators — which subsequently translate into conversions and average order value.

Unlike correlation-based tools, Glow tests competing explanations for why metrics move together and assigns confidence intervals to quantify how certain the model is that one variable genuinely influences another.

Leading Indicators

Glow monitors two primary leading indicators that signal future demand when awareness spend increases:

• Engaged Sessions — a measure of meaningful website engagement driven by brand activity
• Branded Search Impressions — how often people search for your brand name, a direct signal of brand awareness

These metrics provide faster feedback loops before conversion data emerges, giving brand teams early signals that their investment is working.

How the Model Works

Glow uses a Directed Acyclic Graph (DAG) framework to map relationships between inputs and outcomes. Think of it like this: if you see wet grass, you need to consider rainfall, sprinklers, and dew before determining the cause. Glow applies the same logic to brand measurement — considering multiple explanations before assigning credit.

The methodology involves four steps:

1. Test which leading indicators predict revenue shifts
2. Analyse historical patterns in metric movement
3. Calculate the likelihood each input caused outcome changes
4. Assign strength scores quantifying relationship magnitude

The framework draws on approximately two years of historical marketing and sales data to uncover statistically credible causal relationships.

Relationship Strength Categories

Glow categorises the strength of each relationship:

Causality vs. Correlation

Correlation merely shows metrics moving together. Causality identifies which activity most likely drives change. Glow's Bayesian framework considers multiple explanations, accounting for variables like promotions and seasonality. This helps marketers understand not just when metrics move together, but why — grounding insights in causal relationships rather than coincidental patterns.

Three Practical Use Cases

1. Retrospective Reporting

Identifies windows where awareness spend increased and shows how leading indicators responded. This helps form hypotheses about which channels and creative types are most effective for brand building.

2. Live Reporting

Monitors real-time changes in leading indicators to detect whether current awareness activity is translating into meaningful engagement. Helps distinguish between natural conversion lags and ineffective spending.

3. Annotating Brand Activations

Allows teams to mark product launches, offline events, and sales periods directly onto charts — connecting real-world campaigns to leading indicator movements.

Glow Rollout Roadmap

Phase 1 (Current): Visibility into leading indicators and monitoring tools to track them alongside brand spend.

Phase 2 (Coming): Early Glow Model insights quantifying relationships between demand signals and outcomes, highlighting high-impact channels.

Phase 3 (Future): Complete in-platform experience for monitoring, forecasting, and reporting on brand ROI.

Why It Matters

• Causal, not correlational — Glow uses Bayesian Causal Inference to go beyond surface-level patterns
• Faster than alternatives — leading indicator visibility from day one, model insights emerging in weeks rather than quarters
• Actionable — three practical use cases (retrospective, live, annotating) turn brand measurement into brand decisions
• Defensible — confidence intervals and strength categories give brand teams data that finance can trust

Key Takeaways

• Glow = Fospha's causal measurement and brand analytics platform
• Uses Bayesian Causal Inference with a DAG framework — not just correlation
• Monitors Engaged Sessions and Branded Search Impressions as leading indicators of brand health
• Relationship strength is categorised: Very Strong (>0.7), Strong (>0.3), Present (>0), Not Detected
• Three use cases: Retrospective Reporting, Live Reporting, Annotating Brand Activations
• Rolling out in three phases, with leading indicator visibility available now