One big structural change went live at the start of last week: a four-language PMax expansion designed to replicate the pattern of SA PMax Spanish Language ads, which has been the account's strongest creative at 7.24x ROAS / $58 CAC. The first full week of data came in with the April 13 run, and the read is mixed — details below.
All seven changes below were executed live in the Google Ads account yesterday during the April 13 session. Facebook was untouched this run — the Beginner-Setup restructure from earlier runs is still playing out and doesn't need additional intervention this week.
We've been treating "Beginner" as one audience. In reality there are four distinct people inside that label, each in a different emotional state when they see an ad. The 8-variant Fresh Copy test proved that emotional and abstract hooks don't convert — the only Setup-framed angle (Freeze-Beginner-Setup) is carrying the account at $14 CAC.
The brief from yesterday keeps everything inside Setup framing, but fans out horizontally: each sub-segment gets its own ad copy entry point, its own landing page story structure, and its own objection resolution. This expands creative surface area without abandoning the one angle we know works.
The Beginner audience is the clearest signal in the account right now — Freeze and Hold Beginner-Setup on FB, and the four new Beginner sub-persona asset groups on Google PMax. The next unlock isn't more paid angles. It's putting real people on camera inside the same framing, because a stranger saying "I did this in 10 minutes" will always beat a brand ad saying the same thing.
Two tracks running in parallel: Get Based (a paid UGC agency) is producing two dedicated briefs, and a ~40-creator affiliate network is standing by for a wider, lower-cost volume run. Same target, two delivery economics.
Roughly 40 affiliates have confirmed they're ready to create UGC. The brief is intentionally loose: produce beginner-oriented Jade Plus content, anchored on the Setup framing and the 10-minute proof point. Creators pick their own hook (unboxing, first-time setup, why-I-switched, gift-to-friend) — we're not art-directing them into one shape. The volume is the point. Forty creators shipping one or two clips each gives us ~50–80 variants to test inside the Beginner sub-persona asset groups, which is enough breadth to let the algorithm find the 2–3 that break out.
Why both tracks at once. Get Based produces the two high-polish hero pieces we already know are aligned to the winning angles. The affiliate network produces the volume and the unexpected variations — the kind of footage we wouldn't write into a brief ourselves. The hero pieces run inside Freeze and Hold Beginner-Setup on FB and the four new PMax asset groups on Google. The affiliate clips get tested against them.
The problem the tests are solving. Platform-reported ROAS is unreliable. Facebook claims 63 purchases on a recent window — Triple Whale can only attribute 28. Google Search converts exclusively on brand terms ("blockstream jade", "jade plus") — people who would likely click the organic result if the ad weren't there. And the post-purchase survey shows 60% of orders come from buyers with zero paid touchpoints. We don't know what's actually incremental until we turn things off and measure what happens to total Shopify orders.
The approach. Three sequential tests, each isolating one variable. The primary metric for every test is Shopify orders per day — not platform conversions, not pixel ROAS, not clicks. Baseline: 16.7 orders/day, $201 AOV, 85% new customer share, $19.6K monthly ad spend split ~52% FB / 48% Google.
Why this comes first. Whatever these tests find directly funds what we do next. If Facebook isn't incremental, the $10K/month we've been spending on it becomes available for other bets. The clipping program below assumes that freed budget — without it, there's no launch capital.
Where the budget comes from. This program is downstream of the incrementality tests above. If Test 1 confirms that Facebook isn't driving incremental orders (the current directional read), the ~$10K/month we've been spending there becomes available to reallocate. Clipping is the bet we want to make with that capital — lower production cost per attempt, higher surface area, closer to organic in the way it actually feels.
What clipping is. A media distribution strategy where a paid network of independent "clippers" mines long-form content (podcast appearances, interviews, keynotes) and cuts it into short-form vertical clips for TikTok, Reels, and Shorts. Clippers get paid on a CPM basis — they're directly incentivized to hunt for the moments that go viral, because that's where their earnings come from.
Why it works. The distribution economics flip in our favor: instead of us producing and boosting one clip, twenty clippers produce twenty versions and the algorithm picks the winner. The surface area of attempts is multiplied by the size of the clipper network, and we only pay for the views that actually land. It looks organic because it is organic — different editors, different framings, different hooks, all sourced from the same underlying footage.
The plan for Blockstream. Founder-led, Adam Back as the face. Source material is his existing library of long-form podcast appearances — already plenty of hours of him speaking candidly about Bitcoin, self-custody, and why Jade Plus exists. Clippers pull the strongest moments and push them out in parallel. The goal is to create the feeling that Adam is everywhere right now — the kind of ambient buzz you get when a founder is having an organic cultural moment, except we're engineering it on a CPM line item.
What I've been doing. Talked to several clipping agencies over the last two days to understand how the networks are structured, how rates are set, what the approval and brand-safety loops look like, and how attribution gets tracked back to site visits and conversions. Writing up findings and a recommended launch plan — ready to pull the trigger as soon as the Test 1 read gives us the budget.