⚡
What changed — last 7D vs prior 7D
High confidence · 284K events
Revenue
£71K
+£42K (+12%)
Pipeline
£284K
+£65K vs prior
CAC ?
Cost to Acquire a Customer
= Total Spend ÷ New Customers
Target: £39 · Current: £42
= Total Spend ÷ New Customers
Target: £39 · Current: £42
£42
+£3 above target
ROAS ?
Paid Media ROAS
= Paid Revenue ÷ Paid Spend
Excludes free Organic channel
£276K paid rev ÷ £59K paid spend = 4.7×
Target: 5.0× · Display drags this down
= Paid Revenue ÷ Paid Spend
Excludes free Organic channel
£276K paid rev ÷ £59K paid spend = 4.7×
Target: 5.0× · Display drags this down
4.7×
+0.6× vs prior
Wasted spend
£18K
Action required
▲
Paid Search
+£28K
▲
Email
+£12K
▲
Organic
+£8K
+34% spike
▼
Display
−£8K
· 0.8× ROAS
Paid Search ?
Click to toggle on chart.
Shapley-weighted attribution.
Shapley-weighted attribution.
£112K
↑ 22% · 5.9× ROAS
Paid Social
£78K
↑ 8% · 4.1× ROAS
Organic Search
£58K
↑ 14% · ∞ ROAS +34%
Email
£28K
↑ 31% · 11× ROAS
Display
£8K
↓ 4% · 0.8× ROAS
⚠
£18,240 wasted spend detected.
Display at 0.8× ROAS. Reallocate £12K to Paid Search — projected +£38K uplift.
Budget context Quick decision view
Click any row to apply action
Channel
Spend
Share
Revenue
ROAS
Signal
Impact
Action
Total spend
£120,400
Total revenue
£284,000
Paid media ROAS
4.7×
excl. free organic
Optimal uplift
+£80K/mo
Revenue by channel Shapley MTA
7D
14D
30D
60D
Top campaigns
By attributed revenue · 30 days
Brand Search — Exact
Active
£48K · 7.2× · CAC £28 · CTR 4.2%
Prospecting — Lookalike LA
Active
£38K · 4.8× · CAC £45 · CTR 2.8%
Email Win-Back Sequence
High ROI
£16K · 12× · CAC £18 · Open 28%
Display Broad Audience
Wasting
£18K · 0.8× · CAC £182 · CTR 0.2%
Revenue mix Donut
By attributed channel
Full attribution report Shapley MTA
4 models · 284K events
All channels · all attribution models in parallel · confidence scores · 30-day window
Channel
Last Click
First Click
Linear
★ Shapley
Conf.
Action
Attributed revenue
£284K
Paid media ROAS
4.7×
excl. free organic
Total paid spend
£69K
Model agreement
91%
Identified upside
+£80K/mo
AI Growth Analyst · 284,192 attribution events · Shapley MTA · 6 analysis modes · No external API required
What drove revenue this week?
Channel breakdown · drivers · anomalies
Where should we cut budget?
Wasted spend · low ROAS channels · £ impact
Where to increase spend?
High ROAS · headroom · scale opportunities
What is underperforming?
Below average · declining · needs action
Build a growth plan
Full budget reallocation · forecast · risk score
Did TV drive revenue?
Offline impact · uplift · brand search lift
Who is gaining share?
Competitor visibility · SOV · keyword overlap
Full attribution report
All channels · all models · confidence
💡 Insight
💷 £ Impact
✅ Recommendation
Or ask anything in plain English
Ready. Click a prompt above for structured analysis, or type your own question below.
📺
Offline media is shaping this forecast.
TV halo effect adds est. +£28K to 90-day revenue. OOH contributes +£17K. Forecasts shown with and without offline influence.
Rolling forecasts Bayesian + Gradient Boost
High · 87%
87% model confidence · Trained 6h ago · 24 months history · Offline halo effect included
Revenue
30 days
£96K
±8%
60 days
£204K
±13%
90 days
£318K
±18%
ROAS
30 days
4.9×
60 days
5.2×
90 days
5.5×
CAC
30 days
£40
60 days
£37
90 days
£34
Below target ✓
Forecast with vs without offline impact Offline influence layer
Base forecast = digital channels only · With offline = includes TV halo + OOH demand lift · Confidence reflects estimation uncertainty
90-day channel forecast
Projected trajectory with Plan A approved
Offline scenario modelling What-if analysis
Model the impact of changing offline spend on your digital performance forecast. Offline is not directly optimised — it acts as a demand influence layer.
Goal simulator
Drag budget to model revenue
£20K
£60K
£120K
Projected monthly revenue
£284K
at 4.7× ROAS
AI forecast insight
Based on current plan + offline halo
Revenue forecast assumes current budget mix holds. If £12K Display budget is reallocated to Paid Search, the 90-day forecast improves to £348K (+9.4%). 📺 TV halo active
adds an estimated +£28K
above the digital baseline. If TV campaign is removed, Paid Search revenue is projected to fall by £42K over 90 days. Seasonal adjustment applied for upcoming bank holidays.
Growth Plan Builder
Aggregate · Adjust · Simulate · Approve
1
Pre-filled recommendations
Auto-imported from attribution
↑
Applied
Paid Search +£12K
+£38K projected · Low risk
↑
Applied
Email +£2.4K
+£28K projected · Very low risk
↓
Applied
Display −£12K
Stop waste · 0.8× ROAS
2
Edit channel budgets
Live simulation updates instantly
Paid Search
+£12K
5.9× ROAS
Paid Social
→ flat
4.1× ROAS
Email
+£2.4K
11× ROAS
Display
−£12K
0.8× ROAS
Total
£76,400
3
Live simulation
Updates as you edit
Revenue
£364K
+£80K
ROAS
5.9×
CAC
£34
−£8
AI insight:+28.2% revenue uplift at low risk. Paid Search shift has highest marginal return. Display cut frees £12K with zero forecast impact.
4
Actions
v2: One-click push to Google Ads + Meta
Multi-scenario comparison CMO view
Click Load to edit in Growth Plan Builder
| Scenario | Budget | Revenue | ROAS | CAC | Risk | vs Current | |
|---|---|---|---|---|---|---|---|
| Current | £60K | £284K | 4.7× | £42 | Baseline | — | |
| Plan A | £76K | £364K | 5.9× | £34 | Low | +£80K (+28%) | |
| Plan B AI rec. | £72K | £348K | 5.6× | £36 | Low | +£64K (+22.5%) | |
| Plan C | £85K | £398K | 5.8× | £32 | Medium | +£114K (+40%) |
AI budget recommendations
High · 284K events
Each recommendation checks for offline influence before action · Confidence adjusted where offline campaigns are active · Click any card for full data
+£80K
Revenue uplift
5.9×
Projected ROAS
£34
Projected CAC
+£28K
incl. offline halo
One-Click Budget Push v2 Feature
Review approved plan · Push directly to ad platforms · Full rollback available
Plan A — Approved.
£76,400/mo across 4 channels. Expected +£80K revenue uplift. Approval logged in audit ledger.
Platform status
Google Ads
£30K → £42K · +£12K
Meta Ads
£24K → £24K · No change
Display Network
£18K → £6K · −£12K
Email Platform
£2K → £4.4K · +£2.4K
Before vs after
Last push
Not yet pushed
Audit entry
sha256:9b2f1a...
📺
Offline Media Integration — Estimated Impact Only.
All figures are based on observed digital uplift during campaign periods vs baseline. This is not deterministic attribution.
Offline Spend
£620K
3 active campaigns
Est. Incremental Rev
£184K
↑ Estimated uplift
Est. Incremental Sessions
+38K
↑ vs baseline period
Brand Search Lift
+24%
↑ During campaign
Avg Confidence
74%
Medium-high
Offline campaign library 3 campaigns
Estimated impact based on observed uplift during campaign period vs baseline · Not deterministic attribution
| Campaign | Channel | Period | Spend | Lag | Inc. Sessions | Inc. Revenue | Brand Search ↑ | Confidence | Action |
|---|---|---|---|---|---|---|---|---|---|
| Summer TV Burst | TV | 1–14 Mar | £320K | 7d | +18,400 | +£92K | +31% |
82%
|
|
| National OOH — Spring | OOH | 15 Mar–4 Apr | £180K | 3d | +12,200 | +£58K | +14% |
71%
|
|
| Radio — Brand Awareness | Radio | 5–18 Apr | £120K | 0d | +7,600 | +£34K | +8% |
61%
|
Digital uplift during offline campaigns
Sessions indexed to 100 = baseline · Campaign period shown as shaded zone
Offline contribution to total revenue
Estimated influence · Not deducted from digital channel attribution
◈ Attribution
Offline Influence overlay added. TV contributes est. 32% demand halo.
View attribution →
◫ Forecasting
Offline spend included as demand variable. 90-day forecast +£28K with offline.
View forecast →
◑ Budget Recs
TV: strong uplift → Maintain. Radio: weak uplift → A/B test or reduce.
View recs →
◆ Growth Plan
Summer TV Burst added to Plan A. Offline spend: £620K. Recalculate impact.
View plan →
Your SOV
18%
↓ −2% vs prior
Revenue gap vs leader
−£42K
/mo estimated
Total upside
+£42K
If matched leader
Index confidence
72
Medium · SEO signals
Keyword overlap
68%
with top rival
Relative performance index vs market avg
Your performance indexed vs top 3 competitors · 100 = market average · Based on SEO visibility + SOV signals
Trend vs competitors 6-period
Your SOV vs Competitor A (leader) and market avg · SEMrush weekly crawl · Keyword overlap tracked
You
Comp A (leader)
Market avg
Share of spend vs share of return Efficiency analysis
Red = share of spend · Green = share of return · Channels where red > green are over-indexed on spend
Opportunity score by channel With confidence + drivers
Composite: gap-to-leader (40%) · trend direction (30%) · spend efficiency (30%) · expand for full breakdown
Gap analysis + scenario modelling If matched competitor performance
Recommended actions Feeds growth plan · forecast · budget recs
Each recommendation shows confidence score, key drivers, and which system it feeds into. Expand "Why this?" for full reasoning.
Competitor insights applied to
◑ Budget Recs
Competitor SOV gap signals +£8K Paid Search headroom. Confidence 79%.
View recs →
◫ Forecast
Matching Comp A Paid Search performance adds +£28K to 90-day forecast.
View forecast →
◆ Growth Plan
Competitor scenario applied to Plan A. SOV target: 24% in 90 days.
View plan →
◎ AI Analyst
Ask: "Who is gaining share?" for full AI competitor analysis and strategy.
Ask analyst →
Attribution audit ledger Tamper-proof
Every event hashed & timestamped · Blockchain-ready in v2
100%
Integrity verified
284,192
Events logged (30d)
0
Anomalies
Live audit log
Conversion event attributed — Paid Search
sha256:3a9f2c...
2m ago
✓
Growth Plan A approved
sha256:9b2f1a...
4m ago
✓
Budget rule applied — Display cut
sha256:8b1e4d...
5m ago
✓
Attribution model retrained
sha256:f019d3...
6h ago
✓
Forecast model updated
sha256:2e84c9...
6h ago
✓
💬
Sentiment & NLP Overlay
Real-time NLP across tweets, reviews & Reddit posts. Sentiment shifts correlated with conversion impact.
Overall sentiment
72
▲ +6 vs last week
Posts analysed
8,342
Last 30 days
Positive
54%
▲ +8% WoW
Negative
18%
▲ −3% WoW
Conversion impact
+£14K
Est. sentiment uplift
Sentiment trend vs conversion rate NLP overlay
All sources
🐦 Twitter/X
⭐ Reviews
🤖 Reddit
Sentiment score (0–100)
Conv. rate %
Negative spike events
Source breakdown NLP
Positive · Neutral · Negative by source
Top topics Entity extraction
Most mentioned themes this period
Sentiment shift → conversion impact Attribution overlay
Significant sentiment events correlated with conversion rate change in the 72h window
Live post feed NLP classified
All
Positive
Negative
⚑ Flagged
🧠
Model Intelligence — Attribution Transparency Layer (V1)
Shapley is the default model. All outputs include model used, confidence score, and data volume. Expand any insight for full explainability.
● Live · 284,192 events
Default model
Shapley MTA
Primary source of truth
Avg. confidence
87
High · All channels
Events analysed
284K
30-day window
Model agreement
91%
Across 4 models
Stability (WoW)
±3%
Very stable
Attribution model comparison All models · parallel output
🧠 Shapley = default
Channel
Last Click
First Click
Linear
★ Shapley
Confidence
Agreement
ROAS shown per model. Shapley
is the default for all recommendations. Agreement = % closeness across all four models.
Confidence score breakdown V1 logic
Weighted composite: model agreement (40%) · data volume (30%) · stability (30%)
Forecast model Regression · V1
Trained on 24 months of spend vs revenue. Upgradeable to gradient boosting in V2.
Inputs:
Spend by channel · Revenue · Date range · Seasonality flag
V2 upgrade: Gradient boosting + Bayesian smoothing + MMM-lite for offline lag effects
V2 upgrade: Gradient boosting + Bayesian smoothing + MMM-lite for offline lag effects
Budget recommendations Marginal ROAS logic · Shapley-backed
High ROAS → increase · Low ROAS → decrease · Low confidence → flag for validation
V2 upgrade:
Marginal ROAS curves · Diminishing returns modelling · MMM-lite for offline spend attribution
API endpoints API-first · REST
All model outputs exposed via REST API. Attach metadata to existing UI components.
Example API response — Shapley attribution
⚡
3 decisions to unlock +£80K/month
Each backed by marginal ROAS modelling, saturation analysis, model agreement and data volume. Not a summary — a proof. Expand any card to see the full evidence.
+£80K
per month · 84% confidence
284K events · Shapley model
⚠ Cost of inaction — 90-day trajectory if nothing changes
Do nothing
£284K
Revenue
£42
CAC
4.7×
ROAS
→
With this plan
£364K
Revenue
£34
CAC
5.9×
ROAS
·
Sensitivity range
Worst case
+£52K
Expected
+£80K
Best case
+£104K
Top 3 decisions — expand for full proof
Applied to Growth Plan · Flagged for One-Click Push · Logged to Audit Ledger · Confidence: 84% · Sensitivity: £52K–£104K
Before vs after Full impact projection
Current state vs applying all three decisions — modelled from marginal ROAS curves and Shapley attribution
API endpoint GET /decision_summary
Returns top 3 decisions with full proof objects — marginal ROAS, saturation, model agreement, sensitivity range
∿
Marginal ROAS & Response Curves
Diminishing returns modelled per channel using log-curve regression. Marginal ROAS = derivative of response curve at current spend point.
Response curves — spend vs revenue
Budget reallocation simulator POST /simulate_budget
Drag budget from one channel to another — see real-time impact on revenue, ROAS and CAC using response curve modelling
→
+£38K
Revenue Δ/mo
+1.2×
ROAS Δ
−£6
CAC Δ
82%
Confidence
Shift £12K from Display (0.8× marginal ROAS, 92% saturated) to Paid Search (7.2× marginal ROAS, 61% utilised). Expected +£38K/mo revenue at 82% confidence based on response curve modelling.
API response — POST /simulate_budget
Marginal ROAS table GET /marginal_roas
Current spend · current ROAS · marginal ROAS · saturation level · recommended direction
