PMax Campaigns Not Split — Data Integrity Issue
YouTube spend is currently classified as Google Search. Your channel allocation model is inaccurate until PMax campaigns are split. Misattribution Index may be understated by up to 12 points.
Configure PMax Splitter →
TikTok Platform Inflation Detected
TikTok Ads Manager is reporting 3.2× more conversions than GA4 for the same period. This is likely view-through attribution inflation. DVNT has applied a 0.6× deflation factor to TikTok conversions in all calculations.
View in Misattribution →
Direct Traffic Up 28% — Possible Tracking Gap
Direct sessions have increased 28% this month with no corresponding increase in paid spend. This may indicate a GA4 tracking break on a key landing page, or untagged campaign traffic being misclassified as direct.
Check Log Files in Integrations →
Reddit Ads Data Is 3 Days Stale
Reddit Ads last synced on 15 Jun. The Reddit allocation figures shown across this dashboard do not reflect current spend. Re-authenticate to restore live data.
Re-authenticate in Integrations →
Total Ad Spend
£284.5k
+12.4% MoM
6 channels active
Blended ROAS
2.74x
+0.18x vs target
Target: 2.5x
Blended CPA
£38.20
+£2.10 MoM
Target: £35.00
Misattribution Index
68
Moderate risk
Lower is better (0–100)
Budget vs. Attention — Channel Allocation
Your budget share vs. OFCOM raw time-spent vs. DVNT intent-adjusted attention value
Budget Allocation — Current vs. Recommended
Toggle to compare your current spend against DVNT-recommended allocation
vs
Inner = Current
Outer = Recommended
£284.5k
Current Spend
12-Week Performance Trend
Weekly spend (£k) and blended ROAS — rolling 12 weeks
Misattribution Signals
TikTok Under-Invested by 18pp
UK adults spend 35% of social media time on TikTok (OFCOM 2025), yet your allocation is just 10%. Even with a 0.8x intent discount, adjusted attention value is 28%. Budget gap: −18pp. Estimated ROAS uplift if corrected: +0.28x.
Google Search Over-Allocated by 15pp
Google Search commands 35% of your budget but only 5% of time-spent. The 4.0x intent multiplier is justified — but modelling suggests the optimal ceiling is 28–30%. Diminishing returns above this threshold.
Pinterest ROAS Outperforming at Low Budget
Pinterest delivers 3.5x ROAS on just 5% of budget. With a 2.5x intent multiplier and strong 25–45F demographic alignment, scaling to 8% could yield disproportionate returns.
Channel Performance
All Channels — Current Period
| Channel | Budget % | Time Spent % | Intent Mult | Adj. Value % | ROAS | CPA | Status |
|---|
Misattribution Index
68
Moderate — action recommended
Largest Over-Invest
+15pp
Google Search (35% budget vs 20% adj. value)
Largest Under-Invest
−18pp
TikTok (10% budget vs 28% adj. value)
Budget Gap vs. Adjusted Attention Value
Positive = over-invested, Negative = under-invested. Ideal = 0.
Ideal Allocation Scatter
Points on the diagonal line = perfectly allocated. Outliers = misattribution.
OFCOM Benchmark — Raw Time Spent vs. Your Budget
Before intent adjustment. Note: raw time-spent alone is not sufficient — see nuance factors below.
Nuance Factor Breakdown
Intent Mult × Freq Factor × Time Spent = Adjusted Value. Google's 4.0x intent multiplier reflects that users arrive with purchase intent, making short dwell time irrelevant.
Channel Nuance Factors — DVNT Calibration
| Channel | Raw Time % | Intent Mult | Freq Factor | Funnel Stage | Adj. Value % | Your Budget % | Gap |
|---|
Halo Attribution Engine
Direct & Branded Search re-attribution + OFCOM channel opportunities
Unattributed Direct Revenue
£41,200
Incremental pool above baseline
▲ 18% of total revenue hidden
Branded Search Halo Pool
£28,600
Branded PPC + Branded SEO revenue
▲ 12.5% of total revenue over-credited to Google
Total Recoverable Value
£69,800
Available for reattribution
▼ 30.5% of total revenue to redistribute
Blended True ROAS Uplift
+1.8x
vs. platform-reported ROAS
▲ TikTok tROAS: 4.2x vs. reported 1.9x
Direct Traffic Halo
Branded Search Attribution
True ROAS Table
Channel Opportunities
How the Direct Traffic Halo Model Works
Direct traffic is not a source — it is a navigational behaviour triggered by prior marketing exposure. A user sees a TikTok ad, doesn’t click, then types the URL three days later. GA4 credits Direct. DVNT credits TikTok. We establish a baseline direct volume (traffic when spend = £0, representing pure brand equity) and treat everything above that as the Incremental Halo Pool. This pool is redistributed to active channels using a Time-Decay Spend Share × OFCOM Attention Weight formula.
Direct Revenue Waterfall — Baseline vs. Incremental Pool
Halo Credit Distribution by Channel (TDSS Model)
Direct Halo Credit — Channel Level
Baseline: £18,400/mo | Incremental pool: £41,200| Channel | Spend Share | OFCOM Time % | TDSS Weight | Halo Credit £ | Halo Credit % |
|---|---|---|---|---|---|
| Meta (FB+IG) | 28% | 14.2% | 0.198 | £8,158 | 19.8% |
| TikTok | 10% | 18.6% | 0.186 | £7,663 | 18.6% |
| CTV / YouTube | 8% | 22.1% | 0.177 | £7,289 | 17.7% |
| 5% | 4.8% | 0.024 | £989 | 2.4% | |
| 4% | 3.2% | 0.013 | £536 | 1.3% | |
| Google Search | 35% | 2.1% | 0.074 | £3,049 | 7.4% |
| Unallocated Residual | — | — | — | £13,516 | 32.8% |
Branded Search: Demand Harvesting, Not Demand Generation
When a user searches for your brand name, Google did not create that intent — your broader marketing did. The DVNT model applies an Intent Discount: Google retains 15% (friction reduction), and the remaining 85% is redistributed to the upper-funnel channels that generated the brand awareness. Brand search volume from Google Trends and PPC impression data is used as a leading indicator of upper-funnel effectiveness.
Brand Search Volume vs. Upper-Funnel Spend
Google Trends index (0–100) vs. weekly paid social + CTV spend — 26 weeks
Pearson r
0.74
Lag (weeks)
1.8
Significance
p<0.01
Branded Search Revenue — Claimed vs. Redistributed
Google retains 15% friction-reduction credit; 85% redistributed to demand generators
Google Retains
£4,290
Redistributed
£24,310
Intent Discount
85%
Identifying Branded Traffic — Data Sources
| Method | Source | Confidence |
|---|---|---|
| Branded PPC impressions | Google Ads API — brand keyword campaigns | High |
| Branded organic clicks | Google Search Console — brand queries | High |
| Brand search volume trend | Google Trends API — brand term index | Medium |
| GA4 branded sessions | GA4 — source/medium containing brand term | High |
Branded Search Redistribution — Channel Level
| Channel | Redistribution % | Revenue Credited |
|---|---|---|
| Meta (FB+IG) | 34% | £8,265 |
| TikTok | 26% | £6,321 |
| CTV / YouTube | 22% | £5,348 |
| 10% | £2,431 | |
| 8% | £1,945 | |
| Google (retained) | 15% friction only | £4,290 |
True ROAS (tROAS) — Three-Layer Attribution
Platform ROAS is what each ad platform claims. GA4 ROAS is what your analytics reports. DVNT True ROAS adds Direct Halo Credit and Branded Search Redistribution back to each channel — this is the number that should drive budget decisions.
True ROAS vs. Reported ROAS — All Channels
True ROAS (tROAS) — Full Attribution Breakdown
Jan–Jun 2026 | Total spend: £284,500| Channel | Spend | Platform ROAS | GA4 ROAS | + Direct Halo | + Branded Halo | True ROAS | vs. GA4 |
|---|---|---|---|---|---|---|---|
| Google Search | £99,575 | 6.8x | 4.2x | +£3,049 | −£20,020 | 2.9x | −1.3x |
| Meta (FB+IG) | £79,660 | 2.1x | 1.4x | +£8,158 | +£8,265 | 3.6x | +2.2x |
| TikTok | £28,450 | 1.9x | 0.9x | +£7,663 | +£6,321 | 4.2x | +3.3x |
| CTV / YouTube | £22,760 | 1.2x | 0.6x | +£7,289 | +£5,348 | 3.8x | +3.2x |
| £14,225 | 2.4x | 1.8x | +£989 | +£2,431 | 3.1x | +1.3x | |
| £11,380 | 1.6x | 1.1x | +£536 | +£1,945 | 2.8x | +1.7x |
OFCOM + BARB Opportunity Channels — Where Your Audience Actually Is
The tROAS model tells you which current channels are under-credited. This section goes further: using OFCOM Online Nation 2025 and BARB Q1 2026 data, it identifies channels where your target audience spends significant time that you are not currently investing in.
Currently Active Channels — Scale Opportunity
TikTok
Scale — Under-invested vs. OFCOM share
OFCOM Time Share
18.6%
Your Budget Share
10%
True ROAS
4.2x
Gap
−8.6pp
Highest tROAS in the mix at 4.2x once halo credit is applied. OFCOM data shows 18–34s spend 38 mins/day on TikTok. Increasing to 18% budget share is projected to lift branded search volume by 12–15% within 6 weeks.
CTV / Streaming
Scale — Highest attention, lowest CPM
BARB Time Share
22.1%
Your Budget Share
8%
True ROAS
3.8x
Gap
−14.1pp
BARB Q1 2026 shows streaming/BVOD now accounts for 22% of total video viewing time in the UK. CTV delivers the highest attention scores of any digital channel (OFCOM: 71% full-attention viewing) yet receives only 8% of budget.
Google Search
Reduce — Over-invested, demand harvesting only
OFCOM Time Share
2.1%
Your Budget Share
35%
True ROAS
2.9x
Gap
+32.9pp
Google Search is a demand harvesting channel. At 35% budget share vs. 2.1% OFCOM time share, it is the most over-invested channel. The tROAS of 2.9x (vs. platform-reported 6.8x) confirms significant over-attribution. Branded search budget can be reduced 40–60% without revenue impact.
New Channels — OFCOM/BARB Recommended Additions
Podcast Advertising
New — OFCOM: 28% of UK adults weekly
OFCOM Reach
28%
Your Budget Share
0%
Avg. CPM
£12–18
Attention Quality
Very High
OFCOM reports 28% of UK adults listen to podcasts weekly. Podcast advertising delivers the highest brand recall scores of any audio format (72% aided recall vs. 45% for radio). Zero current investment represents a significant gap given audience overlap.
Streaming Audio / Spotify
New — OFCOM: 62% UK monthly reach
OFCOM Reach
62%
Your Budget Share
0%
Avg. CPM
£6–12
Attention Quality
High
62% of UK adults use streaming audio monthly (OFCOM 2025). Spotify’s programmatic audio offers CPMs of £6–12 with strong demographic targeting. Sequential audio-to-display retargeting can bridge upper-funnel awareness to lower-funnel conversion.
What-If Budget Simulator
Drag the sliders to model different budget allocations. Projected ROAS, CPA, and Misattribution Index update in real time.
Total Allocation
100%
Projected Outcomes
Projected ROAS
2.74x
Projected CPA
£38.20
Misattribution Index
68
Python Export
Python 3 — Budget Reallocation Script
# DVNT Media Intelligence — Budget Reallocation Script # Generated: Jun 18, 2026 | Client: Acme Corp import pandas as pd import numpy as np # ── Channel Data ───────────────────────────────────────────── channels = pd.DataFrame({ 'channel': ['Google Search', 'Meta', 'TikTok', 'CTV/YouTube', 'Pinterest', 'Reddit'], 'budget_pct': [35, 30, 10, 15, 5, 5], 'adj_value_pct':[20, 30, 28, 25, 12.5, 7.5], 'roas': [3.2, 2.8, 1.9, 2.1, 3.5, 2.5], 'cpa': [45.50, 38.20, 25.00, 55.00, 32.00, 40.00], }) # ── Calculate Misattribution Gap ───────────────────────────── channels['gap'] = channels['budget_pct'] - channels['adj_value_pct'] channels['status'] = channels['gap'].apply( lambda g: 'over' if g > 5 else ('under' if g < -5 else 'aligned') ) # ── Recommended Reallocation ───────────────────────────────── recommended = { 'Google Search': 28, 'Meta': 30, 'TikTok': 22, 'CTV/YouTube': 15, 'Pinterest': 8, 'Reddit': 5 } channels['recommended_pct'] = channels['channel'].map(recommended) total_budget = 284500 channels['current_spend'] = (channels['budget_pct'] / 100) * total_budget channels['recommended_spend'] = (channels['recommended_pct'] / 100) * total_budget channels['spend_delta'] = channels['recommended_spend'] - channels['current_spend'] print(channels[['channel','budget_pct','recommended_pct','spend_delta','status']])
Automation Scripts
Schedule Overview
Execution Schedule
| Script | Frequency | Last Run | Next Run | Status | Action |
|---|
Product Review Sentiment
Aggregated from Trustpilot, Google Reviews, Amazon — last 30 days
74% Positive
16% Neutral
10% Negative
Recent Reviews
Latest scraped reviews with AI sentiment classification
Competitor Ad Activity
Scraped from Meta Ad Library & Google Ads Transparency — this week
Trending Search Terms
Category-relevant keywords with velocity scores — Google Trends, Reddit, TikTok
Rights Expiry Alert: 2 assets expire within 30 days — Summer_Lifestyle_Hero.mp4 (talent licence, 14 days) and Brand_Music_Sting_v2.mp3 (music licence, 22 days). Pause or replace before expiry.
Variant Generator
Generate structured A/B/n test matrices from your existing creatives. Isolate one variable per test for statistically clean results.
Copy Variant Generator
Image Variant Planner
Platform Spec Compliance Check
Meta Feed (1:1)
Compliant
TikTok (9:16)
Needs crop — safe zone check
Pinterest (2:3)
Compliant
Google Display (300×250)
File size exceeds 150KB limit
Creative Performance Leaderboard
Cross-channel creative ranking by DVNT Creative Score — a composite of CTR, CVR, ROAS, and Fatigue Index. Same asset tracked across all platforms via asset fingerprinting.
Active Creatives
24
Across 6 channels
Top Creative ROAS
4.8x
UGC_Review_Cut_01 — TikTok
Fatigued Assets
3
CTR declining ≥15% WoW
Avg Creative Score
62/100
↑ 4pts vs last month
Creative Score by Format
CTR Trend — Top 3 Creatives (12 weeks)
Full Creative Leaderboard — All Channels
Creative → Attribution Bridge
Connects creative performance to the Halo Attribution model. Identifies which creative types drive the most direct traffic and branded search — and credits them accordingly.
UGC Video
TikTok + Meta Reels
Halo Direct Credit£12,400
Branded Search Uplift+18% WoW
Lag to Search Peak4–6 days
True ROAS (incl. halo)5.2x
View in Halo Engine
Lifestyle Static
Meta Feed + Pinterest
Halo Direct Credit£8,200
Branded Search Uplift+9% WoW
Lag to Search Peak7–10 days
True ROAS (incl. halo)4.1x
View in Halo Engine
CTV / Pre-roll
YouTube + DV360
Halo Direct Credit£14,800
Branded Search Uplift+24% WoW
Lag to Search Peak3–5 days
True ROAS (incl. halo)4.7x
View in Halo Engine
Product / Catalogue
Google Shopping + Meta DPA
Halo Direct Credit£3,100
Branded Search Uplift+3% WoW
Lag to Search Peak1–2 days
True ROAS (incl. halo)3.4x
View in Halo Engine
Community / Reddit
Reddit Promoted Posts
Halo Direct Credit£1,800
Branded Search Uplift+6% WoW
Lag to Search Peak5–8 days
True ROAS (incl. halo)3.8x
View in Halo Engine
DVNT Insight
Creative type recommendation
CTV pre-roll drives the highest branded search uplift (+24%) with the shortest lag (3–5 days). Combined with UGC video on TikTok, these two formats account for £27,200 of halo revenue — 39% of the total recoverable pool.
Recommendation: prioritise CTV and UGC video in the creative mix. Reduce reliance on static product ads for upper-funnel objectives.
See Budget RecommendationsGoogle
Meta
TikTok
Pinterest
Reddit
CTV / YouTube
Google Search & PMax
OFCOM Intent Multiplier: 4.0× · Current Allocation: 35% · DVNT Recommended: 28–30%
3.2×
ROAS
+7%
Over-Allocated
£99.6k
Spend
Strategy
Tactical Actions
Tracking & Setup
Creative Specs
Account Blueprint GEM
DVNT Strategic Position
Google Search commands the highest intent multiplier (4.0×) in the DVNT framework because users arrive with a near-formed purchase decision. Short dwell time is a feature, not a flaw — the user has already done their research elsewhere. This means reducing Google budget purely on time-spent grounds is a strategic error.
However, the current 35% allocation exceeds the DVNT-modelled ceiling of 28–30%. Spend above this threshold enters diminishing returns territory — you are bidding against yourself on branded terms and over-indexing on bottom-funnel at the expense of mid-funnel channels that feed it.
However, the current 35% allocation exceeds the DVNT-modelled ceiling of 28–30%. Spend above this threshold enters diminishing returns territory — you are bidding against yourself on branded terms and over-indexing on bottom-funnel at the expense of mid-funnel channels that feed it.
Campaign Architecture
- Brand Search: Always-on, exact/phrase match. Protect share of voice. Budget: 15–20% of Google total.
- Non-Brand Search: Category + competitor terms. Smart Bidding (tCPA or tROAS). Budget: 40–50%.
- Performance Max: Feed-driven, asset-rich. Requires strong conversion history (>50/month). Budget: 30–35%.
- Display / Demand Gen: Retargeting only. Cap frequency at 3×/week. Budget: 5–10%.
Misattribution Risk
Google's last-click dominance in legacy attribution models inflates its apparent contribution. Switch to data-driven attribution (DDA) in Google Ads and cross-reference with GA4 path analysis. PMax in particular cannibalises branded search — monitor brand vs. non-brand split weekly. If brand search share exceeds 40% of Google spend, you are paying for traffic that would have arrived organically.
DVNT Recommendation
Reduce Google from 35% → 28% of total budget. Redirect the freed 7% (≈£19.9k/month) to TikTok (under-indexed at 10% vs. 18% attention share). Maintain or increase non-brand search investment; the reduction should come from PMax and Display, not from high-intent keyword campaigns.
Priority Tactical Actions
1
Audit PMax Asset Groups for Brand Cannibalisation
Run a Search Terms report filtered to PMax. If branded queries exceed 35% of PMax impressions, add brand terms as negative keywords at campaign level. This prevents PMax from claiming credit for organic brand searches.
2
Switch All Campaigns to Data-Driven Attribution
Navigate to Conversion Actions → Attribution Model → Data-Driven. This redistributes credit across the full path and will typically reduce Google's apparent contribution by 12–18%, giving a more accurate picture of true channel value.
3
Implement Target ROAS Bidding on Non-Brand Campaigns
Set tROAS at 10–15% above your current actual ROAS to give the algorithm room to optimise. Requires minimum 30 conversions in the last 30 days per campaign. Do not apply to campaigns in learning phase.
4
Enable Enhanced Conversions for Web
Pass hashed first-party data (email, phone) via the Google tag or GTM. This recovers conversion signals lost to ITP/cookie restrictions and typically improves conversion modelling accuracy by 15–25%.
5
Reduce Display Budget to Retargeting Only
Prospecting via Google Display Network consistently underperforms vs. paid social for awareness. Restrict GDN to retargeting audiences (site visitors, cart abandoners) with a 3×/week frequency cap. Redirect prospecting budget to TikTok or Meta.
Tracking Health Checklist
Google Tag (gtag.js) — Base tag firing on all pages
VerifiedPurchase Conversion Action — Firing with transaction value
ActiveEnhanced Conversions — Hashed first-party data not yet configured
Action RequiredData-Driven Attribution — Currently on Last Click
Update RequiredAudience Lists — Remarketing lists populated (Site Visitors, Cart Abandoners)
ActiveGoogle Ads API — Not connected for automated reporting
Not Set Up Creative Specifications
RSA Headlines
15 max
30 chars each · Pin 1–2 only
RSA Descriptions
4 max
90 chars each
PMax Images
20 assets
1200×628 + 1200×1200
PMax Videos
5 min max
16:9 + 1:1 + 9:16
Display Banner
300×250
Also: 728×90, 160×600
Ad Strength
Excellent
Target: all RSAs at Excellent
Copy Framework — DVNT Search Principles
Headlines should lead with the user's intent, not the brand. Match the search query language exactly in at least 3 of your 15 headlines. Use price transparency where competitive ("From £X"). Include a social proof signal ("Trusted by 10,000+ customers"). Descriptions should expand on the headline promise and include a clear CTA. Avoid generic CTAs like "Learn More" — use action-specific language: "Shop the range", "Get your free quote", "Book your demo".
Account Blueprint & Automation — DVNT Google Ads Gem
Campaign Type Decision Matrix
The DVNT gem applies a structured decision framework before recommending any campaign type. Campaign type selection is driven by conversion data quality, feed completeness, creative asset availability, and commercial priority — not by Google's default suggestions.
| Campaign Type | Use When | Avoid When |
|---|---|---|
| Brand Search | Always-on — protect share of voice, exact/phrase match | Never pause unless brand spend exceeds 20% of total Google budget with no incremental lift evidence |
| Non-Brand Search | Category + competitor terms. Smart Bidding (tCPA/tROAS) with ≥30 conversions/month | Insufficient conversion volume — use Max Conversions first to build data |
| Performance Max | Conversion tracking, assets, feed quality, and commercial grouping are strong enough to support automation | Brand cannibalisation risk, poor transparency, weak value data, or need for strict query/channel control |
| Demand Gen / YouTube | Clear creative/audience strategy beyond bottom-of-funnel search capture | User expects these to replace intent-led Search without appropriate measurement |
| Exclude / Review | Entity is non-commercial, unavailable, duplicate, policy-sensitive, or missing critical data | Client explicitly approves testing with a documented reason |
Bid Strategy Decision Framework
Bid strategy selection is based on conversion data quality and value reliability — not campaign type alone. The gem enforces a prerequisite check before any Smart Bidding strategy is applied.
| Data Situation | Recommended Starting Strategy |
|---|---|
| New account / limited conversion data | Manual CPC or Maximise Clicks with strict controls |
| Lead gen with validated conversion volume | Maximise Conversions → Target CPA once lead quality is confirmed |
| Ecommerce with reliable conversion value | Maximise Conversion Value → Target ROAS after value tracking is trustworthy |
| High-margin segmented products | Separate targets/budgets if volume and business priority justify control |
| Mixed-margin catalogue, no margin data | Avoid overconfident ROAS segmentation — request margin/custom-label fields first |
| Existing campaigns in learning-sensitive state | Phased bid changes + monitoring — do not wholesale switch strategies |
Automation Roadmap — Semi-Automated OS First
The DVNT gem defaults to a semi-automated operating system before recommending full Google Ads API integration. Automation is decision support plus controlled execution — not a replacement for commercial judgement. Human approval is always required before publishing new campaigns, restructuring live accounts, or altering budgets materially.
| Phase | Tooling | Human Gate |
|---|---|---|
| MVP | Intake forms, audit workbooks, Google Ads Editor CSV/upload sheets, QA checklists | Strategist reviews all rows before posting to live account |
| Operational | Bulk upload templates, decision matrices, monitoring routines, approval gates | Client approves budget/bidding changes; strategist approves structural changes |
| Scale | Google Ads API integration — only after account-building logic and QA process are proven | Dry-run mutation preview required before any live API execution |
Google Ads Editor Upload Sheet — Standard Structure
When building or restructuring campaigns, the gem generates a Google Ads Editor-ready upload sheet as the primary build artefact. All rows default to Paused or Draft status until manually reviewed and approved in Google Ads Editor.
| Sheet Tab | Contents | When Required |
|---|---|---|
| Campaigns | Campaign names, types, budgets, bid strategies, status, geo, language | Always |
| Ad Groups / Asset Groups | Ad group names, default bids, campaign assignment, status | Always |
| Keywords | Keyword text, match type, bid, status, final URL, QA notes | Search campaigns |
| Negative Keywords | Brand negatives for PMax, competitor exclusions, irrelevant query blocks | Always — especially PMax |
| RSAs / Assets | Headlines (15), descriptions (4), asset group assets, pinning instructions | Search + PMax |
| Audiences | Audience signals, remarketing lists, customer lists, observation/targeting notes | PMax + audience-led structures |
| QA Flags | Rows requiring human review — status: Pass / Warning / Blocker / Do Not Upload | Always — mandatory before posting |
Import Workflow: Generate upload sheet → Review outside Google Ads → Import into Google Ads Editor → Resolve import errors → Second manual review inside Editor → Post approved changes to live account only. The sheet is never an automatic publishing mechanism.
QA Guardrails & Approval Gates
The gem enforces mandatory QA checks before any build file is delivered. These gates exist to prevent unsafe live-account automation and protect existing performance history.
TRACKING GATES
✓ Conversion actions verified firing
✓ Enhanced conversions configured
✓ Consent mode implemented
✓ Attribution model set to DDA
✓ GA4 cross-reference active
✓ Enhanced conversions configured
✓ Consent mode implemented
✓ Attribution model set to DDA
✓ GA4 cross-reference active
STRUCTURE GATES
✓ Naming conventions applied
✓ Brand negatives added to PMax
✓ URL quality checked (live, relevant)
✓ Bid strategy prerequisites met
✓ Budget assignments explicit
✓ Brand negatives added to PMax
✓ URL quality checked (live, relevant)
✓ Bid strategy prerequisites met
✓ Budget assignments explicit
APPROVAL GATES
✓ Strategist reviews all rows
✓ Client approves budget changes
✓ Compliance check for restricted categories
✓ Rollback plan documented
✓ Post-launch monitoring set up
✓ Client approves budget changes
✓ Compliance check for restricted categories
✓ Rollback plan documented
✓ Post-launch monitoring set up
NEVER AUTOMATE WITHOUT APPROVAL
✗ Publishing new campaigns live
✗ Restructuring live accounts
✗ Changing conversion goals
✗ Altering budgets materially
✗ Pushing account-wide changes
✗ Restructuring live accounts
✗ Changing conversion goals
✗ Altering budgets materially
✗ Pushing account-wide changes
Research Sources — DVNT Gem Framework
The gem balances three source categories rather than relying on Google's guidance alone. Official documentation covers product mechanics; independent PPC practitioners provide real-world trade-offs; analytics sources cover measurement and attribution.
| Source Category | Use For | Examples |
|---|---|---|
| Official Google Ads docs | Product mechanics, API capabilities, Editor workflows, bulk upload limits, policy | Google Ads Help, API docs, Editor Help, Developer Blog |
| Independent PPC practitioners | Account structure trade-offs, bidding caveats, PMax critiques, real-world optimisation | Search Engine Land, PPC Hero, WordStream, Optmyzr, Define Digital Academy |
| Analytics & measurement | Conversion tracking, GA4, consent mode, attribution, first-party data, CRM import | Google Analytics docs, GTM docs, privacy/consent documentation |
Meta — Facebook & Instagram
OFCOM Intent Multiplier: 1.2× · Current Allocation: 30% · DVNT Recommended: 28–32%
2.8×
ROAS
Aligned
Allocation
£85.4k
Spend
Strategy
Tactical Actions
Tracking & Setup
Creative Specs
DVNT Strategic Position
Meta's 30% allocation broadly aligns with its 25% OFCOM attention share (1.2× intent multiplier). The platform's primary strength is reach, frequency, and creative storytelling — not bottom-funnel conversion. The critical issue is that 68% of Meta spend is currently in conversion-objective campaigns, which underutilises Meta's true value as an awareness and consideration engine.
The DVNT recommendation is to rebalance to a 40% awareness / 60% conversion split, and to treat Meta as the primary channel for building the audience pools that Google Search then converts.
The DVNT recommendation is to rebalance to a 40% awareness / 60% conversion split, and to treat Meta as the primary channel for building the audience pools that Google Search then converts.
Funnel Architecture
- Upper Funnel (40%): Broad audience, Advantage+ Audience, video views / reach objective. Build brand recognition and seed retargeting pools.
- Mid Funnel (30%): Engaged audiences, website visitors 30-day, video viewers 75%. Traffic or engagement objective. Consideration-stage content.
- Lower Funnel (30%): Cart abandoners, purchaser lookalikes, CRM uploads. Conversions objective, tCPA or ROAS bidding. Dynamic product ads.
Attribution Concern
Meta's default 7-day click / 1-day view attribution window inflates contribution significantly. Switch to 7-day click only for a more conservative and comparable measurement. Enable Meta Conversions API (CAPI) with event deduplication — this recovers 15–30% of signals lost to iOS 14.5+ restrictions and is the single highest-impact tracking action available on Meta.
2025 Platform Priority
Advantage+ Shopping Campaigns (ASC) are now the default recommendation for e-commerce. They consolidate prospecting and retargeting into a single AI-managed campaign. Test ASC against your manual campaign structure with a 30-day holdout. Advantage+ Audience (formerly Broad Targeting) consistently outperforms manual interest targeting in mature accounts with strong conversion history.
Priority Tactical Actions
1
Enable Conversions API (CAPI) with Event Deduplication
Implement server-side CAPI alongside the Meta Pixel. Use the event_id parameter to deduplicate browser and server events. This is the single most impactful action for Meta performance — expect 15–30% improvement in attributed conversions and lower CPAs within 2 weeks of implementation.
2
Switch Attribution Window to 7-Day Click Only
In Ads Manager → Settings → Attribution Setting, change from 7-day click / 1-day view to 7-day click only. This gives a more accurate and comparable view of Meta's true contribution. Expect reported ROAS to drop 20–35% — this is not performance declining, it is measurement becoming more honest.
3
Launch Advantage+ Shopping Campaign Test
Create an ASC with 20% of your current conversion budget. Run for 30 days alongside your existing manual campaigns. Compare CPA, ROAS, and incremental reach. Meta's internal data shows ASC outperforms manual campaigns by 12% on average for e-commerce accounts.
4
Consolidate Ad Sets — Reduce Audience Fragmentation
If you have more than 5 active ad sets per campaign, consolidate. Fragmented ad sets split the learning budget and prevent any single ad set from exiting the learning phase. Target: 2–3 ad sets per campaign, each with a minimum £50/day budget.
5
Implement Creative Refresh Cadence
Monitor frequency at the ad level. When frequency exceeds 3.5× for a 7-day window, creative fatigue is likely. Implement a 4-week creative refresh cycle. Maintain 3–5 active creatives per ad set to allow the algorithm to self-optimise toward the best performer.
Tracking Health Checklist
Meta Pixel — Base code firing on all pages
ActivePurchase Event — Firing with value and currency parameters
VerifiedConversions API (CAPI) — Server-side connection not established
Not Set UpAttribution Window — Currently 7-day click / 1-day view (inflated)
Update RequiredCustom Audiences — Website visitors, cart abandoners, purchasers active
ActiveEvent Match Quality — Score below 7.0 (target: 8+)
Improve Required Creative Specifications
Feed Video
4:5 ratio
1080×1350px · Max 15s for best delivery
Stories / Reels
9:16 ratio
1080×1920px · Keep key content top 80%
Carousel
1:1 ratio
1080×1080px · 2–10 cards
Primary Text
125 chars
Before truncation on mobile
Headline
40 chars
Shown below image/video
Safe Zone
Top 14%
Keep text/logo out of bottom 20%
TikTok Ads
OFCOM Intent Multiplier: 0.8× · Current Allocation: 10% · DVNT Recommended: 18–22%
2.1×
ROAS
−8%
Under-Allocated
£28.5k
Spend
Strategy
Tactical Actions
KPI Optimiser
Creative Specs
DVNT Strategic Position
TikTok is the most significantly under-invested channel in this account. At 10% of budget against an 18% OFCOM attention share, the gap is the largest of any channel. The 0.8× intent multiplier reflects that TikTok users are in a passive discovery mode — they are not actively searching, but they are highly receptive to brand storytelling and product discovery.
TikTok's strength is upper-to-mid funnel: building brand awareness, seeding product discovery, and driving consideration that converts on other channels. Treating it as a direct-response channel is the most common strategic error.
TikTok's strength is upper-to-mid funnel: building brand awareness, seeding product discovery, and driving consideration that converts on other channels. Treating it as a direct-response channel is the most common strategic error.
Campaign Architecture (from DVNT TikTok Gem)
- Awareness (40%): Video Views / Reach objective. TopView or Top Feed for maximum impact. Budget: £50/day minimum.
- Consideration (35%): Traffic or Engagement objective. In-Feed Video, Spark Ads from creators. Target: 25–44 demo.
- Conversion (25%): Website Conversions, tCPA bidding. Requires 50+ conversions/month for stable learning. Collection Ads for e-commerce.
Creative Framework (Hook Law)
The DVNT TikTok gem is unambiguous: the first 3 seconds determine everything. The hook must be established before the 3-second mark. After that:
- 0–3s: Hook — movement, question, or pattern interrupt. Zero script delay.
- 3–12s: Body — deliver core message rapidly. Show, don't tell.
- Last 3s: CTA — single, clear action with visual prompt.
2025 Platform Features
Smart+ Catalog Ads — AI-automated for e-commerce. Optimises creative, audience, and bidding automatically. Test for product catalog accounts.
GMV Max — For TikTok Shop merchants. Maximises Gross Merchandise Value.
Spark Ads — Boost organic TikTok content. Highest-trust format. Prioritise creator content over brand-produced assets where possible.
GMV Max — For TikTok Shop merchants. Maximises Gross Merchandise Value.
Spark Ads — Boost organic TikTok content. Highest-trust format. Prioritise creator content over brand-produced assets where possible.
Priority Tactical Actions
1
Scale Budget from 10% → 18% of Total
This is the single highest-impact budget reallocation in the account. Increase TikTok from £28.5k to £51.2k/month. Do not increase all at once — scale in 20% increments over 6 weeks to avoid disrupting the learning algorithm. Monitor CPM and frequency weekly.
2
Implement TikTok Pixel + Events API (CAPI)
Install the TikTok Pixel via GTM and enable Advanced Matching (pass hashed email/phone). Then implement the Events API for server-side signal. This is critical for conversion campaigns — without CAPI, iOS 14.5+ signal loss will prevent the algorithm from exiting learning phase.
3
Do Not Touch Campaigns in Learning Phase (Days 1–7)
TikTok's learning phase requires 50 conversions to stabilise. During days 1–7, monitor daily but make zero changes. After day 7, make one change at a time with 3–5 days between adjustments. Violating this is the most common cause of TikTok campaign underperformance.
4
Test Spark Ads with Creator Content
Source 2–3 UGC creators in your category. Boost their organic posts as Spark Ads. This format consistently outperforms brand-produced content on TikTok by 30–40% on CTR. Budget: £500–£1,000/month per creator for initial test.
5
Set Up Segment-Specific Creative Variants
Create separate ad sets for Gen Z (18–24) and Millennial (25–34) audiences with tailored creative. Gen Z: kinetic, trend-native, fast cuts. Millennial: relatable pain points, practical demonstrations. Do not run the same creative to both segments.
KPI Optimiser — from DVNT TikTok Gem
Select your KPI and current performance status to get targeted actions
Optimisation Timeline (from DVNT TikTok Gem)
- Days 1–7 (Learning Phase): Monitor daily. Make zero changes. Target 50 conversions to exit learning.
- Days 8–14 (Initial Optimisation): Make one change every 3–5 days. Start with bid adjustments before audience changes.
- Days 15–30 (Performance Tuning): Weekly adjustments. Introduce creative variants. Expand audiences if CPA is stable.
- CTR < 0.5%: Review creative hook within 3 days — this is a creative problem, not a targeting problem.
- Frequency > 5: Refresh creative or expand audience immediately.
Creative Specifications
In-Feed Video
9:16
1080×1920px · 9–15s optimal
Hook Deadline
3 seconds
Core message must land by 3s
File Size
500MB max
MP4 or MOV format
Collection Ad
9:16 + cards
Cover video + 3 product cards
Safe Zone
Bottom 20%
Keep key content above CTA overlay
Captions
Required
60% of TikTok watched on mute
Pinterest Ads
OFCOM Intent Multiplier: 2.5× · Current Allocation: 5% · DVNT Recommended: 8–10%
3.5×
ROAS
−3%
Under-Allocated
£14.2k
Spend
Strategy
Tactical Actions
Tracking & Setup
Creative Specs
DVNT Strategic Position
Pinterest delivers the highest ROAS in this account (3.5×) on just 5% of budget — a clear signal of under-investment. The 2.5× intent multiplier reflects Pinterest's unique position as a planning and discovery platform: users are actively curating future purchases, not passively scrolling. This makes Pinterest one of the most efficient bottom-of-funnel channels for product categories with visual appeal.
The DVNT recommendation is to scale to 8% immediately, with a path to 10% if ROAS holds above 3.0× at higher spend.
The DVNT recommendation is to scale to 8% immediately, with a path to 10% if ROAS holds above 3.0× at higher spend.
Campaign Architecture (from DVNT Pinterest Gem)
- Performance+ Conversions (60%): AI-managed. Requires ≥50 events/week. Budget ≥5× target CPA. Best for e-commerce with strong conversion history.
- Consideration / Traffic (25%): Keyword + interest targeting. Drives qualified traffic for categories with longer consideration cycles.
- Awareness (15%): Broad reach, CPM bidding. Builds brand recognition in planning contexts.
Audience Strategy
Pinterest's strongest audience is 25–45F with household income >£40k — high purchase intent, high AOV. Keyword targeting on Pinterest is uniquely powerful because users actively search for inspiration in your category. Build keyword lists around product discovery terms ("living room ideas", "sustainable fashion") not just purchase terms. Actalike audiences (Pinterest's lookalike equivalent) perform well when seeded from purchaser lists.
Tracking Requirement
Pinterest Performance+ Conversions requires ≥50 conversion events per week to optimise effectively. If current weekly conversions are below this threshold, use Consideration objective until volume builds. Pinterest Base Tag must be verified and CAPI should be enabled for maximum signal recovery. Event deduplication is required when running both tag and CAPI simultaneously.
Priority Tactical Actions
1
Scale Budget from 5% → 8% Immediately
Pinterest is the highest ROAS channel in the account at 3.5×. Increase from £14.2k to £22.8k/month. The 2.5× intent multiplier and strong conversion history justify this as the second-highest priority reallocation after TikTok scaling. Monitor weekly conversion volume to ensure ≥50 events/week threshold is maintained.
2
Verify Pinterest Base Tag and Enable CAPI
Confirm the Pinterest Tag is firing on all key pages (product views, add to cart, checkout, purchase). Then implement Pinterest CAPI via the Conversions API. This is essential for Performance+ Conversions to function correctly and for recovering signal lost to browser restrictions.
3
Build Keyword-Rich Pin Descriptions
Pinterest functions as a visual search engine. Pin descriptions should include 2–3 relevant keywords naturally. Boards should be keyword-optimised. This improves both organic and paid distribution. Target: 100-character descriptions with primary keyword in first 50 characters.
4
Test Performance+ Shopping with Product Feed
If you have a product catalogue, connect it to Pinterest and test Performance+ Shopping campaigns. These use AI to automatically match products to relevant search queries and user interests. Requires a verified product feed with accurate pricing and availability data.
Tracking Health Checklist
Pinterest Base Tag — Installed and verified
ActivePurchase Event — Firing with order value
VerifiedConversions API (CAPI) — Not implemented
Not Set UpWeekly Conversion Volume — Below 50 events/week threshold for Performance+
Below ThresholdProduct Feed — Not connected to Pinterest
Not Connected Creative Specifications
Standard Pin
2:3 ratio
1000×1500px · Vertical preferred
Video Pin
2:3 or 9:16
1080×1920px · 6–15s optimal
Carousel
1:1 ratio
1000×1000px · 2–5 cards
Title
100 chars
Include primary keyword
Description
500 chars
2–3 keywords, natural language
Brand Safe Zone
Top 75%
Avoid bottom 25% — UI overlap
Reddit Ads
OFCOM Intent Multiplier: 1.5× · Current Allocation: 2% · DVNT Recommended: 3–5%
1.9×
ROAS
−2%
Under-Allocated
£5.7k
Spend
Strategy
Tactical Actions
Targeting Framework
Tracking & Setup
DVNT Strategic Position
Reddit's 47% YoY growth (OFCOM 2024) makes it the fastest-growing platform in this account's channel mix. The 1.5× intent multiplier reflects Reddit's unique value: users are actively researching, comparing, and seeking peer validation for purchase decisions. This is especially powerful for considered purchases, B2B, tech, finance, health, and gaming categories.
Reddit ads work best when they feel native to the community — not like traditional display advertising. The tone must be authentic, informative, and community-aware.
Reddit ads work best when they feel native to the community — not like traditional display advertising. The tone must be authentic, informative, and community-aware.
Campaign Architecture (from DVNT Reddit Gem)
- Awareness / Reach: Broad subreddit targeting, CPM bidding. KPI: Reach, CPM, Brand Lift.
- Consideration: Interest + community targeting, CPC bidding. KPI: CTR, CPC, Engaged Sessions. Keyword environments show 29.6% CTR boost.
- Conversion: Retargeting + keyword targeting. Requires verified Pixel/CAPI. KPI: CPA, ROAS.
Community Intelligence
Reddit's targeting is unique: you can target specific subreddits where your audience actively discusses your category. This is the most precise contextual targeting available on any platform. Research your top 10–15 relevant subreddits before campaign setup. Look for subreddits with >100k members and active daily discussion. Avoid overly niche subreddits (<10k members) — insufficient scale for paid targeting.
Creative Tone (Critical)
Reddit users are among the most ad-literate audiences online. Ads that feel like ads are punished with downvotes and negative comments. Effective Reddit creative:
- Leads with value or information, not a hard sell
- Uses Reddit-native language and references
- Acknowledges the community context
- Includes a specific, relevant CTA
- Headline under 150 characters — problem-first or proof-led angles outperform direct value propositions
Priority Tactical Actions
1
Fix Reddit Pixel — Re-Authorise Connection
The Reddit Ads integration is currently showing an error status. Re-authenticate the connection immediately. Without a working pixel, conversion campaigns cannot function and retargeting audiences cannot be built. This is blocking all lower-funnel Reddit activity.
2
Implement Full Event Ladder (not just Page Views)
The Reddit Pixel should fire on: Page View, View Content, Add to Cart, Lead/Signup, and Purchase. Currently only base code is verified. Implement the full event ladder via GTM. This unlocks conversion-objective campaigns and enables retargeting audiences based on funnel stage.
3
Research and Map Top 10 Relevant Subreddits
Before scaling budget, identify the 10–15 subreddits where your target audience is most active. Use Reddit's own search and third-party tools (Redditlist, Subreddit Stats). Look for: >100k members, active daily posts, purchase-relevant discussion. This targeting list is the foundation of your Reddit strategy.
4
Test Three Headline Angles Simultaneously
Per the DVNT Reddit gem: run three headline variants in parallel — (1) Direct Value Proposition, (2) Problem-First/Subreddit-Native Hook, (3) Proof-Led/Price-Transparent. Each under 150 characters. Run for 14 days minimum before declaring a winner. The problem-first angle typically wins on Reddit.
Targeting Framework (from DVNT Reddit Gem)
Interest Targeting Vectors
Broad persona buckets and hobby demographics. Use for prospecting campaigns. Reddit's interest categories are less granular than Meta — supplement with community targeting for precision. Recommended interests for most categories: Technology, Finance, Shopping, Lifestyle, Entertainment.
Community Subreddit Seeding
Target specific subreddits where deep subject engagement happens. This is Reddit's unique advantage — no other platform offers this level of contextual precision. Avoid mixing too many subreddits in one ad set — keep thematically related communities together for clean performance data.
Conversational Keywords
Target contextual phrases and consumer inquiries frequently observed in discussion threads. Reddit keyword environments show a 29.6% CTR boost vs. isolated context setups. Focus on research-stage queries: "best [product category]", "[product] review", "[competitor] alternative".
Brand Safety & Exclusions
Reddit requires explicit brand safety management. Exclude: NSFW subreddits (enabled by default — must be manually excluded), controversial political communities, and any subreddits where your brand would be contextually inappropriate. Build an exclusion list before campaign launch and review monthly.
Tracking Health Checklist
Reddit Ads Connection — Authentication error, requires re-authorisation
Auth ErrorReddit Pixel — Base code only, no event ladder
PartialConversions API (CAPI) — Not implemented
Not Set UpPurchase Event — Not firing (requires full event ladder)
Not ActiveRetargeting Audiences — Cannot build without pixel events
BlockedCTV / YouTube
OFCOM Intent Multiplier: 1.0× · Current Allocation: 8% · DVNT Recommended: 8–10%
2.4×
ROAS
Aligned
Allocation
£22.8k
Spend
Strategy
Tactical Actions
Creative Specs
DVNT Strategic Position
CTV and YouTube occupy a unique position in the channel mix: they are the only channels where lean-back, full-screen, high-attention video consumption is the default mode. BARB data confirms that UK adults spend an average of 3h 12m/day watching video content across CTV and YouTube combined — second only to linear TV.
The 1.0× intent multiplier reflects a neutral position: CTV is not a high-intent channel, but it is a high-attention channel. Its primary value is brand building, product demonstration, and retargeting at scale.
The 1.0× intent multiplier reflects a neutral position: CTV is not a high-intent channel, but it is a high-attention channel. Its primary value is brand building, product demonstration, and retargeting at scale.
Campaign Architecture
- YouTube Skippable In-Stream (50%): 15–30s. Brand storytelling, product demos. CPV or tCPM bidding. Target: broad audiences + in-market segments.
- YouTube Non-Skippable (20%): 15s max. High-impact brand moments. tCPM bidding. Use sparingly — high frequency can cause brand fatigue.
- CTV Programmatic (30%): Via DV360 or third-party DSP. Full-screen, unskippable. Premium inventory. Use for brand awareness and retargeting high-value audiences.
Measurement Challenge
CTV is the hardest channel to attribute in this mix. Do not judge CTV on last-click ROAS — it will always appear to underperform. Use brand lift studies (Google Brand Lift for YouTube), view-through conversion windows (7-day max), and incrementality testing to measure true contribution. BARB panel data can validate reach against linear TV for a unified video view.
DVNT Recommendation
Maintain current 8% allocation. The priority is improving measurement quality — switch from CSV upload to YouTube API connection for real-time data. Implement Google Brand Lift study on the next major campaign. Test YouTube Demand Gen campaigns (formerly Video Action) as a bridge between awareness and conversion objectives.
Priority Tactical Actions
1
Connect YouTube API (Replace CSV Upload)
CTV/YouTube is currently on CSV upload — the only channel in the account not on API sync. Connect via Google Ads API to enable real-time data flow into the misattribution engine. This will also unlock automated reporting and budget pacing alerts.
2
Run Google Brand Lift Study on Next Campaign
Brand Lift studies measure the actual impact of YouTube ads on brand awareness, consideration, and purchase intent via surveys. Minimum budget: £50k/campaign. This is the only way to accurately measure CTV's contribution to upper-funnel metrics. Results feed directly into the DVNT intent multiplier calibration.
3
Test YouTube Demand Gen Campaigns
Demand Gen (formerly Video Action Campaigns) combines YouTube In-Feed, In-Stream, and Shorts into a single AI-managed campaign. It bridges the gap between awareness and conversion objectives. Test with 15% of YouTube budget for 30 days. Requires strong creative assets in multiple formats.
4
Implement Frequency Caps on Non-Skippable
Non-skippable ads without frequency caps are the fastest route to brand fatigue on CTV. Set a maximum of 3 impressions per user per week for non-skippable formats. For skippable, 5 per week is acceptable. Monitor brand sentiment in the Scraped Intelligence tab for early warning signals.
Creative Specifications
Skippable In-Stream
15–30s
Hook in first 5s (before skip)
Non-Skippable
15s max
1920×1080px · MP4
YouTube Shorts
9:16 · 60s max
1080×1920px · Vertical only
CTV Programmatic
16:9 · 15–30s
1920×1080px · MP4/MOV
Companion Banner
300×60px
Shown alongside in-stream ads
Captions
Required
40% of YouTube watched on mute
Connection Health
Google Ads
Synced 10:45 today
Meta Ads
Synced 10:42 today
TikTok Ads
Synced 10:40 today
Pinterest Ads
Synced 10:38 today
CTV / YouTube
CSV — 17 Jun 09:00
Reddit Ads
Auth error — re-auth needed
GA4
Primary source · live
Log Files
Not configured
How your data flows into the misattribution engine
Tier 1 — Ad Platforms
Google
Meta
TikTok
Tier 2 — Analytics
GA4
Tier 3 — Log Files
Server Logs
DVNT
Engine
Engine
Outputs
Dashboard
Recs
1
Ad Platform Connections
Connect each advertising platform via OAuth2 API or CSV upload. Pulls spend, impressions, clicks, conversions, and ROAS directly into the misattribution engine.
6 Connected
Manual CSV Import
Supports Google Ads, Meta, TikTok, Pinterest, Reddit, DV360 exports
Drop a channel report CSV here, or click to browse
Column headers are auto-mapped. Supported: date, campaign, spend, impressions, clicks, conversions, revenue PMax Campaign Splitter 2 PMax campaigns detected
Performance Max campaigns distribute budget across 6 inventory types simultaneously. DVNT has detected PMax campaigns in your Google Ads account. Without splitting, YouTube spend is hidden inside Google Search, and your channel allocation model is inaccurate.
Google Campaign Taxonomy 22 campaigns mapped
Each Google campaign is tagged to a DVNT channel bucket. This is the source of truth for how Google spend is split across the misattribution model. PMax campaigns show their estimated split.
All (22)
Search (8)
PMax (2)
YouTube (4)
Shopping (3)
Display (3)
Untagged (2)
| Campaign Name | Type | DVNT Channel Bucket | Monthly Spend | Brand / Non-Brand | Status |
|---|
2
Analytics & Attribution Source
Connect one lead analytics tool as your primary conversion truth. This cross-validates ad platform attribution claims and feeds path analysis into the misattribution engine.
GA4 Active
Why this matters
Every ad platform over-reports its own contribution. Google claims last-click credit. Meta claims view-through credit. TikTok claims assisted credit. Without an independent analytics source, you are reading four different stories about the same purchase. Your analytics tool is the neutral arbiter — it sees the full path, not just the last touchpoint each platform wants to claim. The DVNT engine uses this data to calculate a platform inflation factor for each channel, which adjusts the misattribution score accordingly.
Primary Source
Google Analytics 4
Property: UA-DVNT-001 · Data API
Sessions & users by source/medium
Conversion paths & assisted conversions
Revenue & ecommerce events
Data-driven attribution model
Audience overlap analysis
Adobe Analytics
Report Suite API · Workspace Data
Report Suite data feeds
Workspace segments & calculated metrics
Attribution IQ models
Cross-device journey data
Requires Report Suite ID + API key
Segment / Rudderstack
Webhook Event Stream · Write Key
Real-time purchase & lead events
Custom event properties & traits
Server-side tracking (no cookie dependency)
Identity resolution across devices
Requires Segment/Rudderstack write key
3
Website Log File Ingestion
Raw server logs provide unsampled, cookie-independent traffic data. No JavaScript required — every request is captured, including bots (filtered before analysis).
Not Configured
Log files vs. analytics: GA4 and Adobe Analytics rely on JavaScript tags and cookies — they miss ~15–30% of real traffic (ad blockers, bots, server-side requests). Server log files capture every HTTP request made to your web server, giving you a complete, unsampled picture. The DVNT engine applies a bot filter (removing known crawler IPs and non-human user agents) before ingesting log data. This is especially valuable for understanding organic search behaviour and validating paid traffic volumes independently of platform-reported data.
Sync Log
All tiers · last 50 events
Recent Sync Activity
| Source | Tier | Auth Method | Last Sync | Records | Status | Action |
|---|
Platform Settings
Data Sources
Configure which OFCOM and BARB datasets are used for benchmarking. Currently using OFCOM Online Nation 2025 and BARB Q1 2026 viewing data.
Intent Multipliers
Adjust the nuance factors applied per channel. Default values are based on industry benchmarks and OFCOM intent research. Customise per client vertical.
Alert Thresholds
Set the misattribution gap % that triggers a critical alert. Default: 10% gap = warning, 20% gap = critical.
Weekly Report Email
Automatically send a PDF summary of misattribution scores and recommendations to stakeholders every Monday at 08:00 GMT.
Client Management
Manage multiple client workspaces. Each client has isolated data, separate channel connections, and independent intent multiplier settings.
API Keys & Auth
Manage OAuth2 tokens and API keys for all connected platforms. Tokens are encrypted at rest and rotated automatically every 90 days.