A weekly briefing that distills paid social, SEO experiments, CRO tests, and attribution debates into clean, annotated data stories — delivered before Monday standup.
Issue #001 ships in 14 days · No spam, unsubscribe anytime
0 readers on the waitlist
Across 214 Series B ad accounts tracked this quarter, Meta's ROAS held at 3.8x while TikTok's median dropped to 1.9x — a gap that widens when you exclude DTC brands with strong UGC libraries.
The more interesting story is in the attribution layer: 68% of accounts running incrementality tests found their MTA model was overcrediting last-touch by 1.4–2.1x.
↳ Sourced from anonymized ad account data, Q4 2025. N=214 accounts, $2M+ monthly spend. Full methodology in §04.
Key Finding
68%
of accounts overcrediting last-touch attribution by >1.4x
ROAS by Channel — Q4 2025
214 accounts · $2M+ monthly spend
Median ROAS · Error bars show IQR · Redacted channels visible in full issue
Issue #001CAC Payback Period · SaaS Median
↳ N=340 SaaS companies, ARR $1M–$50M. Median, not mean — distribution is right-skewed.
Four reads worth your Monday morning. Annotations are editorial — we tell you what actually matters.
↳ Worth reading for the confidence interval visualizations alone. The CFO-friendly framing is new.
↳ If you're running TikTok Shop and haven't re-baselining your holdout groups, this is urgent.
↳ Recovery patterns cluster around two strategies — neither involves AI content at scale.
↳ The paid vs. organic CVR gap is widening. Direct traffic still converts 2.3x organic.
Full issue includes 8–12 curated links with extended annotations and source context.
Growth marketing produces more data than any team can process. Dispatch is not a content aggregator — it is a research function. Every issue is built from primary data before secondary sources are consulted.
We don't republish press releases. We don't amplify vendor benchmarks without independent verification. When we're uncertain, we say so — with explicit confidence levels.
Editorial Standards
Anonymized ad account exports from 200+ growth teams, aggregated weekly. No individual account data is identifiable. Consent obtained via platform data partner agreements.
Trends are selected when they appear in ≥15% of accounts within a 4-week window AND diverge from the prior 12-week baseline by more than 1.5 standard deviations.
Every data story is reviewed by at least one practitioner with direct experience in the relevant channel. We distinguish correlation from causation explicitly, always.
Primary data, secondary research, and editorial opinion are labeled separately in every issue. We cite sample sizes, date ranges, and confidence levels for all quantitative claims.
"We hold ourselves to the standard of a good research analyst: show your work, acknowledge uncertainty, and never let a good narrative override inconvenient data."
— Dispatch Editorial Policy, v1.0
Issue #001 ships to the waitlist in 14 days. No teaser content, no drip campaign — just the full first issue, on time.
Issue #001 ships in 14 days · No spam, unsubscribe anytime
0 readers on the waitlist