Apr 17–23, 2026
Fewer sessions, sharper conversion: site CVR +21% on -10% traffic, blended CAC $94 vs $104, and Apr 23 hit $81 — the best day of the month.
The Shift
The Executive Read
This is the cleanest efficiency week we've had in six. We spent $42K less than the prior week (Daily Stand), acquired only 82 fewer customers, and site CVR jumped from 2.17% to 2.62% (GA4). That's a 21% lift in conversion on 10% fewer sessions — the ratio I'm looking for every week and rarely get. Blended CAC came in at $93.71, inside our $90 target band for the first time this month (Daily Stand). Apr 23 specifically printed $81 CAC on 497 new customers — our best day of April.
The single most important reason efficiency improved is that we executed last week's memo. Last week I called for killing paid traffic into /pages/compare-listicle-og (1.03% CVR) and /pages/rise-2 (1.31%) and re-routing into /products/30-servings-tin (3.46% CVR). That happened. This week /products/30-servings-tin was the #3 landing page at 30.6K sessions and 3.57% CVR (GA4) — up roughly a full point of conversion on effectively similar volume. Meanwhile /pages/rise-2 sessions stayed flat and /pages/compare-listicle-og pulled back meaningfully. The money that stopped chasing low-intent listicle traffic got converted on a PDP that closes.
The second lever was channel-level pruning. YouTube spend was cut 68.8% week-over-week — from $22.4K to $7.0K (Daily Stand). Google cut 20.5% (Daily Stand). Facebook pulled back 9.4% to $194.6K (Daily Stand). We traded $42K of gross spend for 82 customers of lost volume, which is a very good trade — YouTube was the marginal dollar that wasn't earning its keep, and cutting it compressed the denominator on blended CAC materially.
The third contributor is Meta's internal mix. Within Meta's $189.9K of attributable spend (Meta API), the ABO sandbox carrying the FOUNDER vs Mushroom Coffee listicle ads is the efficient workhorse at $133 CPA on $86.5K of spend, and the spring-sale ASC campaigns are printing $96–$101 CPA on $25.6K combined (Meta API). ASC top-ads scaling (1dc + 7dc) is heavy at $68.3K combined but landing at $141–$146 CPA — that's the rebalance target for next week. The story on Meta is that the sales event discount + concentrated top-performer ads carry the bulk of efficient volume.
Caveat on the data: Northbeam's API timed out repeatedly this week, so all Meta channel claims are on Meta's platform-reported metrics. For blended numbers I'm leaning on Daily Stand. That's directionally consistent with the pattern we usually see, but cross-check attribution in Northbeam when it's back.
What Changed in the Ad Mix
Spend came off nearly every paid channel this week, but it came off unevenly — and the unevenness is the story. YouTube got cut hardest (-69%), then Google (-21%), then Facebook (-9%). Smaller performance channels (Podcast, GeistM+FirstMedia, Impact affiliates) held flat or grew slightly. Net effect: we concentrated spend into the channels we trust and pulled oxygen from the ones we were testing into (Daily Stand).
| Channel | This Wk | Prior Wk | Δ | Signal |
|---|---|---|---|---|
| YouTube | $6,984 | $22,360 | -68.8% | Slashed |
| Facebook CA | $1,774 | $3,409 | -48.0% | Paused |
| $16,074 | $20,208 | -20.5% | Trimmed | |
| $194,608 | $214,782 | -9.4% | Held | |
| GeistM+FirstMedia | $15,223 | $14,328 | +6.2% | Held |
| Impact (affiliate) | $12,017 | $11,433 | +5.1% | Held |
| Shopify Shop | $1,499 | $1,227 | +22.2% | Nudged up |
| Superfiliate (ambassadors) | $1,116 | $1,475 | -24.3% | Trimmed |
| Total DTC Spend | $296,512 | $338,675 | -12.4% | Deliberate pullback |
Inside Meta specifically, the ABO sandbox carrying Founder vs Mushroom Coffee is doing $91.5K of spend at $133 CPA (Meta API). The OG ASC scaling stack (1dc + 7dc) is a combined $68.3K at $141–$146 CPA — fine but the worst CPA in the account. The spring-sale ASC stack is the hidden winner at $25.6K combined at $96–$101 CPA (Meta API). That's where the next dollar of Meta spend should go.
What Changed in Creative
The creative picture is concentrated. Two flavors of the Founder vs Mushroom Coffee listicle ad (1dc and 7dc click window variants) are eating $74.9K of spend — 39% of Meta — at a blended $128 CPA (Meta API). Those are doing the heavy lifting on volume. The higher-CPA ASC top-ads stack continues to carry scale, but the per-customer efficiency is worse. Best CPA in the account sits with the spring-sale ASC ads at $96–$101.
| Ad set / creative | Spend | CPA | CTR | CPM | Video | Status |
|---|---|---|---|---|---|---|
| Mar13 Listicle · FOUNDER vs Mushroom Coffee (1dc) | $37,598 | $122 | 1.85% | $31.35 | Ad | Workhorse |
| Mar13 Listicle · FOUNDER vs Mushroom Coffee (7dc) | $37,306 | $135 | 1.76% | $30.86 | Ad | Workhorse |
| OG Sales ASC · top-ads (7dc) | $35,643 | $141 | 1.62% | $42.53 | — | Held |
| OG Sales ASC · top-ads (1dc) | $32,664 | $146 | 1.66% | $46.89 | — | Held |
| ASC · OG Coffee Spring Sale (7dc) | $13,092 | $96 | 1.33% | $31.14 | — | Hidden winner |
| ASC · OG Coffee Spring Sale (1dc) | $12,547 | $101 | 1.34% | $31.42 | — | Hidden winner |
| Marpipe catalog ads (1dc + 7dc) | $5,223 | $141 | 1.14–1.28% | $17–$23 | — | Deprioritize |
Note on video links: the Founder vs Mushroom Coffee video_id 1801811114090882 is confirmed in the creative catalog (Meta API). The ASC top-ads adsets are collections of creatives without a single canonical video per adset, so I've marked those with an em-dash rather than fabricate a link.
What Changed in Landing Pages
This is the cleanest signal in the report. The top three LPs by sessions this week are /pages/rise-2 (41.6K, 1.42% CVR), /pages/compare-listicle-og (31.0K, 1.21% CVR), and /products/30-servings-tin (30.6K, 3.57% CVR) (GA4). The PDP is now within touching distance of the top two listicle LPs on volume while converting roughly 2.5–3× higher. That's the rebalance we called for last week showing up in the data. Bundle PDPs are still elite: /products/vanilla-chai-mocha-bundle converted at 7.55% on 5.2K sessions and /products/strawberry-matcha at 5.08% on 7.2K sessions (GA4) — these are the LPs that should be soaking up more paid traffic.
| Landing page | Sessions | Purch. | CVR | Read |
|---|---|---|---|---|
| /pages/rise-2 | 41,643 | 532 | 1.42% | Highest traffic, below floor |
| /pages/compare-listicle-og | 30,973 | 368 | 1.21% | Still leaking — cut deeper |
| /products/30-servings-tin | 30,608 | 974 | 3.57% | Hero PDP — scale further |
| / (homepage) | 21,423 | 487 | 2.50% | Strong baseline |
| /products/strawberry-matcha | 7,179 | 322 | 5.08% | Elite CVR, under-traffiked |
| /collections/shop | 6,586 | 174 | 3.04% | Above floor |
| /products/vanilla-chai-mocha-bundle | 5,204 | 362 | 7.55% | Best converter in catalog |
| /products/coffee-starter-kit | 3,396 | 83 | 2.87% | Above floor |
| /pages/nourish | 3,386 | 3 | 0.15% | Leak — pull paid |
| /pages/fix | 1,950 | 17 | 0.88% | Still a leak |
Note: GA4 delivered this week's landing-page roster only; the script did not pull the same table for Apr 10–16. Directional WoW read above is relative to the prior week's memo (Apr 13–19), which is a 3-day overlap. I'm calling out levels, not deltas, on LPs.
Why Performance Is Improving
We pulled $42K of spend out of the worst marginal channels — YouTube (-69%) and Google (-21%) (Daily Stand) — which compressed the CAC denominator. Inside Meta, spend concentrated on the FOUNDER vs Mushroom Coffee listicle creative and the spring-sale ASC ads ($96–$133 CPA), which are the cleanest performers in the account (Meta API). That paid traffic flowed predominantly to /products/30-servings-tin (3.57% CVR) and the homepage (2.50% CVR) instead of sub-2% listicle LPs — a direct execution of last week's directive. The site converted 21% harder (GA4), revenue grew 11.6% on 10% fewer sessions (GA4), and blended CAC landed at $93.71 (Daily Stand). Apr 23 in particular benefited from the Tuesday–Wednesday LTO drop ($56K in email-attributed revenue across two sends) (Klaviyo) — but the structural efficiency gain was in place before that email lift.
Double Down
- Scale the FOUNDER vs Mushroom Coffee Mar13 Listicle ABO ad to $50K/day combined across 1dc + 7dc variants. Currently at $10.7K/day combined at $122–$135 CPA (Meta API) — it's the most efficient high-spend ad in the account and it's not showing fatigue (CTR 1.76–1.85%, CPM $31).
- Double the spring-sale ASC stack to $50K combined (from $25.6K this week). $96–$101 CPA across 1dc + 7dc is the best CPA in Meta (Meta API). We're under-funding the cheapest new customer on the account.
- Route another 10K sessions/wk of Meta traffic from /pages/rise-2 to /products/30-servings-tin. Rise-2 is at 1.42% CVR on 41.6K sessions; 30-servings-tin is at 3.57% on 30.6K (GA4). Every 10K sessions rerouted is ~215 additional purchases at flat spend.
- Start funneling paid traffic into /products/vanilla-chai-mocha-bundle (7.55% CVR) and /products/strawberry-matcha (5.08% CVR) (GA4). Both are under 8K sessions/week — they can easily absorb 3–5× more traffic without CVR collapse.
- Cut paid to /pages/nourish (0.15% CVR) and /pages/fix (0.88% CVR) to zero. 5.3K combined sessions, 20 purchases (GA4). These are pure leaks — flush them from the planner.
- Hold YouTube at $1K/day test budget only. The 69% cut worked — don't rebuild it back to $22K/wk without a specific new creative thesis to test.
Watch
- ABO sandbox is 46% of Meta spend ($86.5K of $189.9K) (Meta API). Two creatives — 1dc and 7dc variants of the same Founder vs Mushroom Coffee listicle — are doing most of that work. If this ad fatigues or CPM inflates on the Memorial Day / Father's Day buyer surge, we lose the efficiency cushion fast. We need a shadow-top-performer test queued now.
- Volume is shrinking, not just CAC. New customers down 2.5% WoW and subscribers down net 1,455 (Omni). The efficiency win is real, but if we keep pulling oxygen off channels and spend drops another 10%, we'll hit subscriber-attrition math fast. Efficiency has to stabilize, then we reinvest — not keep shrinking.
- /pages/compare-listicle-og is still getting 31K sessions at 1.21% CVR. Despite the rerouting, the listicle is still a top-3 LP and a top-3 drag on blended conversion (GA4). If the Meta account continues sending 7dc ASC traffic there by default, next week's CVR gain could snap back. This needs a forced LP swap in the ad builder, not just a directive.
- Gorgias volume spike: 3,834 tickets this week vs 94 the week prior (Gorgias) — driven by a jump in Siena-handled subscription cancel tickets. CSAT is 4.87 and sentiment is positive, so this looks like either a data-pipeline change or a Siena volume-reclassification event. Worth confirming with CX that volume is ticketing hygiene, not a churn surge that would undercut the CAC win.