What is Glow?
Glow is a causal measurement and brand analytics platform that helps you understand how brand investment influences long-term business outcomes.
Using Bayesian Causal Inference, Glow estimates how likely your brand activities are to drive leading indicators such as Engaged Sessions and Branded Search Impressions, and how changes in these indicators translate into downstream outcomes like Conversions and Average Order Value (AOV) over time.
By quantifying the strength and direction of these relationships and revealing how their effects unfold over time - Glow helps marketing and finance teams connect brand activity with future business performance.
Where we’re headed — Now, Next, Later
We’re approaching Glow’s rollout in three stages:
-
Phase 1 – You’ll receive visibility into leading indicators (metrics that signal future demand) and tools to monitor how they grow, so you can track them alongside brand spend.
-
Phase 2 – You will get early insights from our Glow Model, quantifying the relationship between demand signals and outcomes, and highlighting channels that drive demand.
-
Phase 3 – Glow will evolve into a complete in-platform experience, giving your team the ability to monitor, forecast, and report on brand ROI directly within Fospha.
Why Glow Matters?
Glow bridges the gap between marketing performance and financial impact.
Brand investment is often one of the largest, yet least measurable, drivers of growth. This creates misalignment between CMOs who want to build future demand and CFOs who need to see ROI today. Glow brings both sides together by connecting brand activity to business outcomes, giving marketers the confidence to defend, optimize, and plan budgets with measurable evidence.
Glow helps marketing teams answer three long-standing questions in brand measurement:
-
How do I see which metrics are signals of future demand? (Available now)
If brand investment isn’t influencing conversions in the short term, it can be hard to prove its impact. Glow surfaces leading indicators - showing how awareness activity is driving future demand. This gives you faster feedback loops to report on how brand investment is performing and prove that it is creating measurable demand.
-
How do I know which awareness activity is actually building future demand? (Coming soon)
-
Traditional MMMs and brand lift studies take time to reveal the impact of brand investment, making it harder to double down on what works in the moment. Glow highlights which channels are driving demand, so teams can confidently invest their budgets.
-
-
How can I plan today’s budgets when I can’t see brand’s long-term payoff? (Coming soon)
-
Marketers are often forced to choose between hitting this month’s numbers and investing in future growth. Glow gives you a long-term forecast of how your digital brand media influences outcomes, so you can speak the same language as finance, and plan for the future.
-
What can Glow be used for right now?
In this first phase of the Glow Alpha, the dashboard acts as a reporting and discovery tool, helping you understand how brand activity influences the leading indicators that predict future growth.
Glow currently supports three practical use cases for marketing teams:
1. Retrospective Reporting
Purpose:
Measure how leading indicators shifted in response to historical changes in awareness spend
How it works:
Glow identifies suggested windows - periods where awareness spend increased and shows how leading indicators such as Engaged Sessions and Branded Search Impressions responded over time.
Outcome:
Use these insights to identify which past awareness activity drove measurable movement in the leading indicators. Through this, you’ll be able to form hypotheses about which channels, activations or creative types drove impact, helping shape future upper-funnel investment accordingly.
2. Live Reporting
Purpose:
Monitor changes in leading indicators in real-time to detect early signals of brand impact
How it works:
Glow measures real-time changes in Engaged Sessions and Branded Search Impressions and how they are responding to current awareness spend.
Outcome:
Glow helps you spot whether current awareness activity is translating into meaningful engagement.
-
If leading indicators rise (e.g., Engaged Sessions or Branded Search Impressions increase) while awareness spend also grows, this suggests your activity is building future demand — even if blended ROAS hasn’t yet improved.
-
If leading indicators plateau or decline while awareness spend increases, it may signal that your awareness activity isn’t growing brand signals. This could be because other factors are impacting their leading indicators. For example, pulling back spend significantly, could lead to a drop in engage sessions.
By reading these signals together, you can judge whether awareness investment is driving meaningful engagement or not.
3. Annotating brand activations
Purpose:
Annotate brand activations in the dashboard to see how they align with movements in leading indicators
How it works:
Add annotations — such as product offline events or brand activations — directly onto overtime charts to see how leading indicators moved during those moments.
Outcome:
Create a unified view of your brand’s impact across digital and offline activity. By annotating key brand moments — such as product launches, sales periods, or offline events — you can connect real-world campaigns to movements in leading indicators. This helps your team visualise all brand activity in context, revealing what drove changes in performance and enriching Glow’s causal insights with real-world context.
How the Glow Model Works
The easiest way to understand Glow’s model is through an analogy.
|
Imagine you walk outside and notice the grass is wet. There are a few possible reasons — it might have rained, the sprinklers could have been on, or there was morning dew. To figure out why, you’d draw on prior evidence: weather reports, sprinkler schedules, and the time of day. Based on this evidence, you’d judge which explanation is most likely. |
Glow’s model works in a similar way.
|
It uses Bayesian inference — a statistical framework that updates the probability of one event causing another as new evidence becomes available. Glow analyses how changes in your upper-funnel activity (such as impressions, brand search, or engaged sessions) relate to shifts in downstream outcomes like conversions or average order value (AOV). It then estimates how likely each relationship is to have influenced those outcomes. Using your historical data, the model quantifies how much of the observed change can be statistically explained by the metrics it sees — while distinguishing between what can be accounted for and what cannot. Each relationship is assigned a confidence score, showing how strongly the model believes one metric has influenced another.
|
Here’s how it does that:
-
Identify leading indicators: Glow tests which leading indicators truly predict revenue shifts— such as Branded Search and Engaged Sessions.
-
Looks at historical patterns: It analyses how these indicators have moved together over time.
-
Tests different explanations: The model calculates how likely each input is to have caused a change in outcome, while accounting for other variables that could also have an effect.
-
Quantifies confidence scores: Each relationship is assigned a strength score, showing the strength of the relationship identified between inputs and leading indicators.
By combining these probabilities, Glow moves beyond surface-level correlations — helping you initially understand to what degree awareness causes movement in leading indicators.
In the future, we’ll help you understand which awareness activities build future demand.
FAQ's
Q: How to use Glow alongside Fospha’s core product
A: Glow and Fospha are designed to work together — one tracking performance efficiency, the other uncovering leading indicators of long-term outcomes
Use Fospha’s core dashboards to optimize active spend and diagnose changes in efficiency metrics.
Use Glow to understand to what degree awareness causes movement in leading indicators — identifying whether awareness investment is creating the early signals that lead to growth over time.
For example:
-
When leading indicators rise before conversions do, Glow provides confidence that brand spend is working — signalling future revenue uplift, even before it appears in performance reports.
-
When ROAS dips as awareness spend scales, Glow helps confirm whether the drop reflects a natural lag between brand exposure and conversion, rather than declining efficiency.
By combining both views, you can act with confidence in the short term while keeping sight of the longer-term effects of your marketing investment.
Q: Why should I trust these leading indicators?
A: Glow’s leading indicators are grounded in both your historical data and real-world validation across hundreds of brands.
Using around two years of your own marketing and sales data, Glow’s Bayesian inference model tests how leading indicators — such as Engaged Sessions and Branded Search Impressions — respond when awareness investment changes.
This enables the model to identify which signals consistently move in line with brand activity and have the strongest statistical relationship with downstream outcomes like conversions and revenue.
To ensure reliability beyond a single dataset, these indicators have been validated across more than 360 brands and markets, confirming that they behave consistently as early signs of future demand.
They’ve also been shaped through qualitative feedback from our conversations with brands and performance marketers to ensure the metrics are intuitive, reasonable and representative as signals of brand impact.
By combining rigorous statistical modelling with cross-brand validation, Glow gives you confidence that the indicators you see are actually reflective of brand performance.
Q: What data does Glow use?
A: Glow analyses around two years of your historical marketing and sales data to uncover meaningful relationships between your marketing activity and business outcomes.
It uses data from Google Analytics (including engaged sessions and impressions) and paid media platforms (such as PPC – Brand) to understand how shifts across your awareness, consideration, and conversion activity influence key outcomes like conversions and average order value (AOV).
By using these long-term trends, the model can identify statistically robust relationships between your awareness activity (such as impressions and spend), engagement signals (like site visits or sessions), and downstream business outcomes (such as conversions or AOV).
This historical depth allows Glow to distinguish short-term fluctuations from consistent, long-term patterns, ensuring the insights it surfaces are grounded in meaningful evidence rather than day-to-day noise.
Q: How does Glow determine 'relationship strength?'
A: Glow measures how strongly two metrics move together over time, based on patterns in your historical data.
It does this using a Bayesian inference model, which estimates the probability and magnitude of one metric influencing another — for example, how changes in awareness activity might drive changes in engaged sessions or branded search.
Each relationship is summarised by a strength coefficient, which represents how confidently the model can link movement in one variable to movement in another, after accounting for other factors.
Higher coefficients indicate stronger, more consistent relationships observed across your data.
For example, a strength coefficient of 0.5 suggests a moderate and statistically credible link - when one metric increases, the related metric tends to move in the same direction about half as strongly, on average.
Glow categorises these coefficients as:
-
Very Strong = strength coefficients > 0.7
-
Strong = strength coefficients >0.3
-
Present = strength coefficients >0
-
Not Detected: No statistically significant relationship found
Note: “Not detected” does not mean no relationship exists — it often reflects limited or highly variable data that prevents the model from estimating the relationship with confidence.
Q: How is causality determined rather than correlation?
A: Correlation shows that two things move together — for example, awareness spend and brand search — but it doesn’t explain why. Causality, on the other hand, helps uncover which activity is most likely driving the change.
Glow uses a Bayesian causal modelling framework, built around a Directed Acyclic Graph (DAG).
This structure maps the relationships between multiple inputs and outcomes — such as awareness spend, engaged sessions, and conversions — and tests a range of possible explanations for how they interact.
By analysing historical data through this framework, Glow estimates the probability that one variable has a causal influence on another, while accounting for other factors that could also play a role (for example, promotions, seasonality, or overlapping campaigns).
The model then returns results with confidence intervals, showing how certain it is that a given input is genuinely influencing an outcome.
Think of it like this: students who get good grades often study a lot — but that doesn’t mean studying alone causes high grades. Some have better teachers or quieter homes.
Glow’s causal model works the same way: it considers multiple explanations and identifies the factors most likely to be truly driving performance.
By going beyond surface-level correlations, Glow helps marketers understand not just when metrics move together, but why — grounding insights in statistically credible causal relationships rather than correlation.
Looking for more on Glow? Check out all our resources below: