[CUSTOMER] — [HEADLINE METRIC — must be sourced] with Alfred

One sentence: who [CUSTOMER] is, the decision Alfred surfaced, and the sourced number it produced. Lead with the number; keep the adjectives out.

[METRIC 1][what it measures]Source: [customer-approved source + period]

[METRIC 2][what it measures]Source: [customer-approved source + period]

[METRIC 3][what it measures]Source: [customer-approved source + period]

The arc — situation, decision, outcome

1 · Situation

What was true before

[CUSTOMER]'s stack, spend and team — and the specific pain: where were the numbers scattered, what was being missed, what did a decision cost in hours or days?

Quantify the before-state: [SPEND/MO], [SOURCES], [HOURS/WK ON REPORTING].

2 · Decision

What Alfred surfaced

The specific brief, alert or recommendation — what changed, why, and what Alfred recommended, with the confidence score it carried. Quote the brief verbatim where possible.

Name the human who decided: [NAME, ROLE] actioned it on [DATE]. Alfred recommends; people decide.

3 · Outcome

What the number says

The result, measured over [PERIOD] against [BASELINE]: [METRIC — must be sourced].

State the measurement method in one line — where the number comes from and who verified it.

"[QUOTE — the customer's own words on the decision, not on the product]"

[NAME] · [ROLE], [CUSTOMER]

Rules before this ships

Every metric carries a visible source and the customer's written approval — a metric without both does not ship, and neither does the case study. No rounding up, no "up to" framing, no composite numbers. If real numbers are not available yet, use the Northwind demo scenario on page 2 and label it as demo — never as a customer.

Case study · worked example

Demo scenario — Northwind Labs is fictional

B2B SaaS · 120 people

Northwind Labs sizes $55.1K/mo of recoverable waste in its first month

Northwind Labs is the fictional demo customer used across every Alfred surface — one company, one internally consistent dataset. This page shows the template filled in correctly; none of it is a real customer outcome.

$55.1Krecoverable waste sized, per monthNorthwind Labs demo dataset

$4.8Kzero-conversion campaign caughtNorthwind Labs demo dataset

4.8xblended ROAS, +12.4% vs 30dNorthwind Labs demo dataset

Situation

Northwind Labs is a 120-person B2B SaaS company spending ~$310K/month across eight campaigns on Google Ads, Meta, LinkedIn and TikTok, with GA4, HubSpot, Stripe and Mailchimp behind them. Performance lived in ten tabs; the Monday number never matched the Friday number, and a mid-quarter pipeline of $1.84M against a $3.2M target left no room for wasted spend.

Decision

Alfred connected read-only in week one and synced a year of history. The first Daily Brief led with the sharpest anomaly — quoted verbatim from the demo:

"Kill the Meta broad campaign. It burned $4.8K this week with zero conversions — lead quality fell 14% when the new audience launched."

Alfred · Daily Brief · sources: Meta, GA4, HubSpot · Northwind demo

Priya Menon, VP Marketing, approved the kill (confidence 91) and the companion recommendation: reallocate $18K from non-brand search into Performance Max, converting at 5.1x against 3.4x (confidence 82). Both approvals — and every read behind them — are in the audit trail.

Outcome

Within the first month, Alfred sized $55.1K/mo of recoverable waste across the portfolio, blended ROAS reached 4.8x (+12.4% vs the prior 30 days), and a fatiguing TikTok hero creative (CTR −12% in a week) was rotated before it decayed ~$26K/mo of spend. Measurement method: Alfred's reconciled cross-platform dataset — the same numbers on every surface.

Why this example is labelled demo

Alfred publishes no customer metrics until they are sourced and customer-approved. Northwind Labs exists so demos and collateral can show real mechanics with visibly fictional, internally consistent numbers — drawn only from the canonical dataset, never invented per document.