← Soundtrak AI Studio
Case study · The Signal Newsletter

A newsletter produced weekly, not getting enough engagement.

Property: The Signal, Andrew Braun's weekly thought-leadership newsletter · Campaign: eight-week amplification sprint, run end to end by the system · Period: 25 May to 1 July 2026

The problem

The content engine worked, but the compounding did not. The Signal Newsletter had shipped nine or more issues without missing a week, the voice was locked, and the illustrations were recognisably its own. But only about 1,000 of roughly 3,000 LinkedIn followers saw a typical Tuesday post, the Substack subscriber base sat at 25, and the soft "subscribe below" line at the bottom of each LinkedIn post had converted nobody.

The diagnosis reached during the fact-find: it was not a content problem. It was a distribution and conversion problem, and the LinkedIn post was unintentionally cannibalising the newsletter by delivering the full article with no clear CTA or signal that a richer version existed.

What was produced

Big idea

Every Tuesday is now a Trailer to the Playbook.

Rename the relationship between the two surfaces. The LinkedIn post is the Trailer: the argument, free, public, complete enough to nod at. The Substack article is the Playbook: the named case, the fair objection, and the three actions a person can take on a Monday. The audience was not subscribing because they could not see what they would be subscribing to. Once the depth differential has a name, the click stops being a leap of faith and becomes the obvious next step.

LinkedIn
The Trailer. The full argument, free, ending on the line that promises more, the Playbook by name.
Substack
The Playbook. The named case, the fair objection, the three moves worth making this week.
The stamp
The call-to-action, clearly indicating a richer version exists.

Key message

"The Tuesday post is the Trailer. The Playbook is on Substack."

Every concept in this system must survive a 15-second radio read, roughly 40 voiced words. Not because the ad will air, but because 15 seconds is a brutal forcing function for message clarity. If the proposition cannot survive a constrained read-aloud, it is not tight enough to ship. This is the campaign's message in script form:

The 15-second radio ad

[voice: dry, Australian, unhurried, a half-smile]

"Every Tuesday on LinkedIn you get the argument. Free, complete, worth a nod. That's the Trailer."

"The Playbook is on Substack: the named case, the fair objection, the three moves to make this week. Four minutes, every Tuesday."

"The Signal."

[Drafted retrospectively for this case study. The discipline now applies to every concept.]

Marketing plan

Four workstreams, one proposition. Each workstream ships as finished assets, a selection of which are shown below.

1
Forensic diagnosis
  • 18 historical LinkedIn posts analysed before a dollar moved
  • Prior boosted post autopsied: wrong objective, $175 CPM against a $45 benchmark
  • The important fact found: 14.9% of readers who clicked through subscribed
2
Conversion infrastructure
  • CTA stamp with an application kit
  • 12-pattern CTA library: closing lines, mid-post CTAs, day-2 comments, P.S. patterns
  • Weekly LinkedIn tiles with in-situ feed previews
3
Weekly Trailer cadence
  • Tuesday post, tile, and Substack Playbook opener, every week
  • Post shape and CTA variant rotate across four weeks
  • Eight of 17 assets fast-lane once the week 1 pattern locks
4
Surgical paid test
  • $1,000 locked: $500 LinkedIn, $500 capped Reddit experiment
  • Pre-registered A/B framework, CTA by post length
  • Conditions written down before launch, not improvised after

Key assets

Real assets from the campaign, exactly as they shipped. Seventeen planned; the week 1 launch package alone was eight ship-ready files. The three below show the system from destination to Trailer.

The Signal on Substack, live publication with masthead, subscribe surface and the archive of published issues

The destination, live. The Signal on Substack (ab64.substack.com): masthead, subscribe surface, and the archive of published issues every campaign asset points at. The week 1 article and hero shipped here as part of an eight-file launch package.

Hand-sketched illustration: own the backbone of your content, with owned research as a hub feeding articles and talks

The hand-sketched illustration. A Substack hero produced by the system in the newsletter's own napkin-sketch style: own the backbone of your content, with owned research as the hub feeding every article, talk, and post.

A LinkedIn Trailer post's closing CTA: the Playbook tease with a link, a 'what the full article covers' bullet list, and a free anti-AI-slop checker offered to Signal subscribers

The Trailer's CTA. Every LinkedIn Trailer closes with a conversion CTA drawn from the 12-pattern library: the Playbook tease with the link, the "what the full article covers" bullets, and a free anti-AI-slop checker offered to subscribers.

The results

The numbers

18 posts
Forensically analysed before a dollar of paid spend, surfacing the 14.9% click-to-subscribe conversion
1 day
Campaign open to approved Brief, selected Concept, and approved Plan
17 assets
In the plan. Week 1 alone shipped eight: post, tile, paid creative in two ratios, ad copy, article, hero, day-2 comment
~$4.90
Total AI cost to date, roughly 830k tokens, metered

Every figure is pulled from campaign artifacts on disk; nothing invented. Market results (reach, subscribers, cost per subscriber) land when the eight-week sprint reads out on 31 July 2026, and this page gets refreshed with them.

The phases

Every Soundtrak campaign walks the same six phases. The client approves at every red gate; the system does the work between them.

1 Brief structured interview 2 Concepts three full directions 3 Plan assets, budget, timeline 4 Production AI produces, brand-checked, one gate per asset 5 Publish step-by-step cookbooks, no technical handoffs 6 Ongoing assets become templates, each cycle costs less = client approval gate. Nothing advances through a red diamond without sign-off.

On this campaign: Brief approved, Concept selected, and Plan locked on day one. The expensive work happened once. The stamp, the CTA library, the shape rotation, and the in-situ preview patterns are reusable infrastructure, which is why eight of the 17 assets fast-lane after week 1: the weekly cycle gets cheaper as it goes, and at wrap the validated patterns graduate up so the next campaign inherits them on day one.

"The system markets the publication that markets the system."