Documentation
How to Read Guyav
A visual field manual for every metric, chart, and signal on your dashboard.
On this page▾
Getting Started
Seasonality
Strategic Output
Reference
How to Read the Dashboard
The dashboard is organized around one question: how desirable is this brand right now, and what should we do about it?
Start with the Regime - the colored badge at the top of each brand page. It tells you the brand's current state in one word. Everything else on the page explains why it's in that state and what to do about it.
Next, scan the three executive metrics: ADI (cross-brand rank), Momentum (trajectory), and Alpha (brand vs category). These answer three distinct questions - see the sections below.
If you want to go deeper, the Signal Diagnostics panel tells you about signal quality (volatility, exhaustion, trend conviction). The Strategic Brief synthesizes everything into a plain-language action plan.
The Four Regimes
A regime is the brand's current desirability state, detected by the Hidden Markov Model. Each regime implies a different strategic posture.
Low volume, positive velocity. Signal forming - watch and seed.
What to do: Watch and test. Small, targeted bets while the signal is forming.
Example: A new maison that insiders love but the wider public has not yet discovered.
Volume building with positive velocity. Demand compounding - scale and protect.
What to do: Scale media and protect supply. Capture the compounding window.
Example: A heritage house whose attention has compounded for several quarters and is decoupling from its category.
Negative velocity. Desirability eroding - intervene before erosion compounds.
What to do: Reduce exposure and audit causes. Do not defend with discounting.
Example: A previously hot brand whose attention is fading as the category rotates to newer entrants.
Flat velocity. No active demand signal - hold or exit.
What to do: Hold or exit. Re-allocate budget to active signals.
Example: A long-established brand whose attention has plateaued and is neither rising nor falling materially.
How often do regimes change?+
Regime changes are rare by design. The model includes a stability rule that requires a new state to persist for multiple consecutive weeks before it's confirmed. This prevents false regime changes from noisy data. Most brands stay in a regime for months.
Is the Dormant regime bad?+
No. Dormant simply means the brand isn't generating an active demand signal right now - velocity is flat. For a long-established brand or a heritage maison between drops, this is expected. The question is whether the dormancy is structural (the brand has lost relevance) or transitional (between cycles). Watch for the first positive velocity reading to move into Emerging.
Regime & Confidence
The regime badge at the top of each brand page includes a confidence percentage.
Confidence is the HMM posterior probability - how certain the model is that the brand is in this specific regime at this moment. It ranges from 0% to 100%.
A confidence of 87% means the model assigns an 87% probability to Heating and distributes the remaining 13% across the other three regimes.
| Range | Level | Meaning |
|---|---|---|
| > 80% | High | Strong conviction. The regime assignment is reliable. |
| 60–80% | Moderate | Reasonable conviction, but the brand may be transitioning. |
| < 60% | Low | The model is uncertain. The brand may be between regimes. Treat the assignment with caution. |
ADI Score (Desirability Index)
A cross-brand comparable 0–100 index that ranks brands by demand-signal strength within a category. Like a Zacks Rank for early demand signals.
Desirability Index
Above-average desirability within the category.
Sample subject · Feb 23, 2026
| Range | Level | Meaning |
|---|---|---|
| > 75 | Elite | Top-tier desirability. Supports premium positioning and price increases. |
| 60–75 | Strong | Above-average desirability. Healthy competitive position. |
| 40–60 | Neutral | Average among peers. Neither leading nor lagging. |
| 25–40 | Weak | Below average. Risk of losing relevance. |
| < 25 | Lagging | Bottom tier. Structural issues - brand needs fundamental reassessment. |
Component Weights
Velocity
40%
Speed of change in desirability
Alpha
25%
Brand vs category growth
Heat
20%
Social sentiment signal
Acceleration
15%
Change in velocity
ADI vs Momentum - what's the difference?+
ADI is a cross-brand comparable rank. It z-scores signals across all brands in the category at the current point in time. Think of it like class rank - it tells you where you stand vs peers right now.
Momentum is a per-brand trajectory. It tells you whether desirability is building or fading, regardless of where you rank among peers.
In demand terms, it shows whether the brand's pull is building or fading for this specific brand, regardless of absolute scale.
They can diverge: A brand with high Momentum but low ADI is building quickly but still behind peers. A brand with low Momentum but high ADI is still desirable, but momentum is fading.
What does 'Lagging' mean in ADI?+
Lagging (ADI below 25) means the brand sits in the bottom tier of desirability relative to all tracked brands in the category. It doesn't mean zero demand - it means peers are significantly more desirable right now. This requires structural intervention, not tactical adjustments.
Momentum Score
A 0–100 score measuring the brand's current trajectory - is desirability building or fading?
Momentum Score
Brand is gaining signal traction. Momentum is positive and accelerating.
Rank 2 of 5 brands
Based on last 12 weeks - as of Feb 23, 2026
| Range | Level | Meaning |
|---|---|---|
| 80–100 | Surging | Exceptional momentum. Brand is in breakout territory. |
| 60–79 | Heating Up | Solid positive trajectory. Growth is accelerating. |
| 40–59 | Stable | Neither gaining nor losing. Steady state. |
| 20–39 | Cooling | Momentum is declining. Monitor for further deterioration. |
| 0–19 | Stalling | Very low momentum. Brand may be losing relevance. |
What period does Momentum cover?+
Momentum is based on the recent window, typically the last 12 weeks (configurable per category). The score combines trend, velocity, and acceleration within that window.
It is also cross-sectional: your score is normalized relative to all brands in the category at this point in time. A score of 61 means you're in the upper half of momentum among peers.
Can Momentum be high while ADI is low?+
Yes. High Momentum with low ADI means the brand is catching up - rising fast but starting from behind. This is a promising signal. Conversely, low Momentum with high ADI means the brand is coasting - still ranked highly but losing steam.
Alpha Score
Answers the most expensive question in marketing: "Did we grow because of us, or because the whole category grew?"
Alpha Score
Growing faster than the category. Brand-specific drivers are working.
Based on last 12 weeks of velocity data
Alpha compares the brand's growth velocity against the category average. An Alpha of 1.0 means you're growing at exactly the category rate. Above 1.0 means you're outperforming; below 1.0 means the category tide is carrying you, or you're falling behind.
| Range | Level | Meaning |
|---|---|---|
| > 1.5 | Leader | Significantly outperforming. Brand-specific drivers are dominant. |
| 1.1–1.5 | Outperformer | Growing faster than category. Differentiation is working. |
| 0.9–1.1 | Follower | In sync with the category. Growth is market-driven, not brand-driven. |
| 0.5–0.9 | Underperformer | Losing ground relative to category. Needs attention. |
| < 0.5 | Laggard | Growing at less than half the category rate. Structural issue. |
What is a Laggard?+
A Laggard (Alpha below 0.5) means the brand is growing at less than half the rate of the overall category. If category searches grew 20% but the brand grew only 8%, that's an Alpha of ~0.4 - the category is surging and the brand isn't capturing any of it. This usually points to a structural positioning problem, not a media spend issue.
What does the confidence interval mean?+
The 95% CI (e.g., 1.12–1.64) gives you the range within which the true Alpha likely falls. A narrow CI means high confidence in the score. A wide CI (e.g., 0.6–2.1) means the signal is noisy - treat the point estimate with caution.
Demand Environment
A macro context signal that tells you whether the overall market is helping or hurting your brand.
Category is in a tailwind. Pricing power is high.
No strong macro signal. Focus on brand-specific drivers.
Category headwind. Protect margins, delay launches.
The Demand Environment combines category-level interest trends with public macro proxies (luxury index performance, consumer confidence, FX effects) to determine whether the market itself is helping or hurting. A brand with strong momentum in a hostile environment is particularly impressive - and vice versa.
Signal Diagnostics
Three quality indicators that tell you how trustworthy the underlying signal is.
Signal Clarity
Low noise. The signal is clean and reliable.
Trend Direction
Trend Exhaustion
The current trend is showing signs of running out of steam.
Volatility (Signal Clarity)
Measures how noisy the underlying data is. High volatility means the signal bounces around a lot, making regime detection less reliable. Low volatility (Calm) means the signal is clean and you can trust the readings.
| Range | Level | Meaning |
|---|---|---|
| Low | Calm | Clean signal. Metrics are trustworthy. |
| Medium | Normal | Some noise. Metrics are directionally correct. |
| High | Turbulent | Very noisy. Treat exact numbers with caution. |
Trend Direction
The probability distribution over three directional states. The conviction percentage tells you how confident the model is about the dominant direction. 72% Bullish means a clear upward trend; 45% Bullish with 40% Neutral means ambiguous.
Trend Exhaustion
Estimates how much runway the current trend has left. A Sustainable trend has room to continue. Fatiguing means the rate of change is slowing. Exhausted means the trend is likely near reversal.
| Range | Level | Meaning |
|---|---|---|
| Low | Sustainable | Trend has room to continue. No signs of fatigue. |
| Medium | Fatiguing | Trend is slowing. Monitor for reversal. |
| High | Exhausted | Trend is near its limit. Expect a direction change. |
Signal Divergence
What happens when the brand signal and the category signal disagree.
Brand and category are moving in the same direction at similar rates. No divergence.
Category is rising and lifting the brand with it. Growth may not be brand-driven.
Brand is outperforming a flat or declining category. Strong brand-specific drivers.
Brand is surging but category is flat/declining. Potential hype bubble - watch for reversal.
What's the difference between Hidden Strength and Hype Risk?+
Both show a brand outperforming the category, but they differ in sustainability. Hidden Strength implies steady, organic outperformance - the brand has genuine differentiation. Hype Risk implies rapid, potentially unsustainable outperformance - often driven by a viral moment that may not translate into lasting desirability.
Changepoints & Inflections
Points in time where the underlying signal fundamentally shifted.
Recent Changepoint
Dec 8, 2025
14 weeks ago
Magnitude: 2.4σ
Strong shift in trend level
A changepoint is a moment where the statistical properties of the signal changed - a shift in the mean level, variance, or trend direction. The system automatically detects these and classifies them.
Structural
A lasting shift in the signal level. The brand has moved to a new baseline. Usually triggered by major events (new creative director, scandal, product launch).
Transient
A temporary spike or dip that reverts to the prior baseline. Usually triggered by seasonal effects, one-off PR moments, or viral content.
How is magnitude measured?+
Magnitude is expressed in standard deviations (σ) of the signal. A 2.4σ changepoint means the signal shifted by 2.4 standard deviations from its previous level - a substantial move. Anything above 2σ is considered a strong changepoint.
Seasonal Pattern
Recurring calendar effects that inflate or deflate interest - Christmas gifting, Valentine's Day, fashion weeks.
Every brand has predictable seasonal rhythms. Champagne searches spike in December and May (weddings). Handbag searches rise before fashion weeks. The system decomposes the raw signal into a trend (the underlying demand-signal trajectory) and a seasonal component (the calendar effect).
The Latent Brand Equity chart on the dashboard shows the trend after removing seasonal noise. This is the cleanest signal for strategic decisions - it tells you whether the brand is actually growing or if December is just doing its thing.
Timing Shift
Is the seasonal peak arriving early, on time, or late this year?
The timing shift detects whether the expected seasonal peak is arriving ahead of schedule, on time, or behind. An early seasonal peak can signal intensifying consumer interest; a late peak may indicate market softness.
For champagne, if the Christmas spike starts in mid-November instead of early December, that's an early shift - potentially driven by earlier retail promotions or a gifting trend moving forward.
Early
Peak arriving ahead of schedule
On Time
Peak matching historical pattern
Late
Peak arriving behind schedule
Strategic Brief
An AI-generated, brand-specific action plan synthesized from all signals.
The Strategic Brief is generated by an LLM that receives the full regime analysis, momentum score, alpha score, signal diagnostics, and recent news for the brand. It produces a concise, action-oriented summary with three sections:
Situation
What the data says about the brand right now.
Implications
What this means for positioning, pricing, and media.
Actions
Concrete next steps for the marketing team.
Transition Forecast
The most likely next regime and the probability of each transition.
Transition Probabilities - from Heating
The HMM produces a transition matrix - the probability of moving from the current regime to any other regime in the next period. A 78% probability of staying in Heating means the regime is stable. The 14% chance of moving to Cooling is the most likely alternative path.
These are not predictions. They represent the learned historical pattern of transitions for brands in the category. Think of them as base rates, not forecasts.
What does 'most likely next regime' mean?+
It's the regime with the highest transition probability other than the current one. If a brand is in Heating with 78% stay probability and 14% Cooling probability, then Cooling is the "most likely next regime." It does not mean the transition will happen - only that if a transition occurs, that's the most probable destination.
Glossary
Quick reference for every term used in the dashboard.
| Term | Definition |
|---|---|
| ADI | Guyav Demand Index. A cross-brand comparable 0–100 score ranking brands by desirability within a category. |
| Acceleration | The second derivative of the signal - the change in velocity. Indicates whether momentum is building or fading. |
| Alpha | The ratio of brand growth velocity to category growth velocity. Above 1.0 = outperforming the category. |
| Changepoint | A point in time where the statistical properties of the signal fundamentally shifted. |
| Confidence | The HMM posterior probability that the brand is in the assigned regime. Higher = more certain. |
| Demand Environment | Macro context signal (Favorable / Neutral / Hostile) based on category trends and economic proxies. |
| Exhaustion | How much runway the current trend has. Sustainable → Fatiguing → Exhausted. |
| Heat | Internal social-attention signal derived from public social and community channels. Use it as supporting evidence, not as a standalone proof of desirability. |
| Laggard | Alpha below 0.5 - growing at less than half the category rate. Also: ADI below 25. |
| Latent Brand Equity | The underlying desirability trend after removing seasonal noise. |
| Momentum | A 0–100 score measuring the brand's trajectory. Combines trend, velocity, and acceleration. |
| Regime | The brand's current desirability state as detected by the Hidden Markov Model. See the four regimes section above. |
| Emerging | A desirability regime. Low volume, positive velocity. Signal forming - watch and seed. |
| Heating | A desirability regime. Volume building with positive velocity. Demand compounding - scale and protect. |
| Cooling | A desirability regime. Negative velocity. Desirability eroding - intervene before erosion compounds. |
| Dormant | A desirability regime. Flat velocity. No active demand signal - hold or exit. |
| Stability | The probability of staying in the current regime. Above 0.8 = stable; below 0.5 = transitioning. |
| Structural (changepoint) | A lasting shift in the signal level - the brand has moved to a new baseline. |
| Transient (changepoint) | A temporary spike or dip that reverts to the prior baseline. |
| Velocity | The first derivative - the speed and direction of change in desirability. Positive = building, negative = fading. |
| Volatility | How noisy the signal is. Calm = clean data. Turbulent = treat exact numbers with caution. |