Decision Intelligence Playbook — From Data to Decision by Friday

    Playbook · 02 of 02

    From data to decision — by Friday.

    The framework I deploy to turn GA4, BigQuery and Looker into the one call your board actually needs every week.

    Definition

    A dashboard isn't insight. Insight is the next correct decision.

    Analytics
    Tells you what happened. Sessions, conversions, channel splits — the historical record. Necessary, but not enough.
    Business intelligence
    Tells you how it's trending. Cohorts, segmentation, KPIs over time. Better — but still leaves the decision to the room.
    Decision intelligence
    Tells you what to do next. Move $10k from A to B, kill experiment C, double down on D. A reallocation, not a chart.

    Decision Intelligence · Framework

    The four pillars behind every weekly reallocation call.

    Each pillar maps to a specific failure mode I see across B2B and B2C estates — fragmented data, vanity dashboards, gut-feel budgets, and channels scored on the wrong number.

    01 / 04

    Single source of truth

    GA4 → BigQuery → Looker. One warehouse. One definition per metric. The CFO and the growth lead read the same number on Monday.

    02 / 04

    Decision metrics, not vanity

    Every dashboard tile must be able to move a budget line. If it can't, it doesn't ship. Pipeline, CAC payback, LTV — never impressions.

    03 / 04

    Weekly reallocation cadence

    Signal → decision → reallocation, on a 7-day loop. Underperformers get cut at week 2, not quarter 2. Compounding wins, fast.

    04 / 04

    Cohort & LTV intelligence

    Payback math that retires gut-feel CAC budgets. Channels are scored on 90-day repeat rate and contribution margin — not first-click ROAS.

    In practice

    When the data said move — we moved.

    3 anonymized reallocations · last 12 months
    B2B SaaS · DACH

    Signal

    GA4 cohort: paid-search trial→paid conversion dropped 38% WoW while branded-organic trials held flat.

    Decision

    Cut Google Ads non-brand by 45%. Redeployed budget into AEO content + lifecycle email.

    Outcome

    CAC ↓ 31% in 6 weeks · pipeline flat → +18%

    DTC E-commerce · GCC

    Signal

    BigQuery LTV model flagged Meta cohort LTV at 0.6× of Google cohort despite equal CAC.

    Decision

    Reallocated 40% of Meta spend to Google Shopping + retention SMS flows.

    Outcome

    ROAS 3.2× → 5.8× · 90-day repeat rate +24%

    Enterprise IT · EU

    Signal

    GSC + Perplexity citations showed AI-search referrals overtaking long-tail organic for solution queries.

    Decision

    Shifted 30% of link-building budget into GEO content + structured-data sprints.

    Outcome

    AI citations 8× in 90 days · MQLs +42%

    Source: GA4 · BigQuery · HubSpot · Search Console — client names withheld under NDA

    Frequently asked

    Decision intelligence — straight answers.

    What is decision intelligence in marketing?

    +

    Decision intelligence is the discipline of turning marketing data into specific, time-bound budget and channel decisions on a weekly cadence. It sits between analytics (what happened) and BI (how is it trending) — and outputs a recommendation a CEO or CFO can act on by Friday. The unit of work is not a dashboard, it's a reallocation.

    How is this different from analytics or business intelligence?

    +

    Analytics tells you what happened. BI tells you how it's trending. Decision intelligence tells you what to do next — which $10k to move from channel A to channel B this week, and what the expected payback is. It's the layer that makes the other two pay for themselves.

    Do I need BigQuery to run this?

    +

    Not always — but at meaningful scale (>$50k/mo media spend, multi-channel funnel, multiple regions) GA4's UI and sampling become the bottleneck. BigQuery gives you raw event data, joined with CRM and ad-platform spend, so cohort, LTV and payback math become trustworthy. For smaller estates, GA4 + a clean Looker layer is enough to start.

    How quickly can a weekly reallocation cadence start?

    +

    Two weeks for a clean baseline (audit, metric definitions, fix tracking gaps), one week to ship the first executive-grade dashboard, then weekly reallocation calls from week four onward. The first 2–3 cycles are calibration; meaningful reallocation typically starts at cycle 4–5.

    Who owns the dashboards after the engagement?

    +

    You do. Every Looker view, every BigQuery model, every metric definition is documented and lives in your stack — not mine. The handover is part of the deliverable, not an afterthought. Most clients keep operating the cadence in-house once the muscle is built.

    Can decision intelligence work for B2C and DTC, or only B2B?

    +

    Both. The metrics differ — B2B leans on pipeline, sales velocity and CAC payback; DTC leans on contribution margin, 90-day repeat rate and LTV:CAC — but the cadence is identical: signal → decision → reallocation, every seven days.

    Sibling playbook

    Pair with: AI search & GEO.

    Decision intelligence tells you where to move budget. GEO makes sure the buyers asking ChatGPT and Perplexity find you when they do.

    Next step

    Audit your dashboards in 20 minutes.

    Bring me your GA4 + Looker. I'll flag the three tiles that aren't earning their pixels, and the one decision your data is already telling you to make.