13-Month Spend & Performance Correlation Report | March 2025 — March 2026 (est.) | Generated March 25, 2026
After reviewing 13 months of daily stand data across all channels, in-platform metrics are telling an incomplete story in at least three major ways:
YouTube spend correlates r = +0.80 with DTC new customers and r = −0.73 with CAC. When YouTube crossed $200K/mo (Oct-Nov), you hit your best CAC and volume months of the entire year. When it dropped back under $140K (Jan-Mar 2026), CAC climbed $5-9 and customer volume fell. YouTube doesn't just acquire customers on its own — it makes Meta and everything else more efficient.
Content network spend shows a r = +0.70 correlation with Amazon new customers. However, correlation ≠ causation. Amazon's growth from ~21% to ~35% of new customers could be driven by brand maturity, Amazon's algorithm, review accumulation, Subscribe & Save momentum, or seasonal factors that happened to coincide with content network scaling. The correlation is worth watching, but don't make budget decisions on it without a controlled test (e.g., pull content network spend for 4-6 weeks and see if Amazon NC drops).
When YouTube ($373K), Meta ($966K), and Content Networks ($213K) all ran hot simultaneously, you hit 23,924 new customers at $78 DTC CAC / $61 blended. But November is Black Friday / Cyber Monday — promotional pricing drove higher conversion rates across every channel. The channel mix helped, but the discounts are doing a lot of heavy lifting in that CAC number. October is a cleaner read: YouTube at $299K still produced 18,317 customers at $88 CAC with no major promo. That's the more honest baseline for what a diversified channel mix delivers.
TikTok spend went from $48K (Mar 2025) to $0 (Dec 2025 onward). There's no visible negative impact on any performance metric — DTC customers, CAC, or Amazon volume all held or improved after the wind-down.
Meta's share has climbed from 52% to 60-64% as YouTube and Podcast spend came down. Meanwhile, Content Networks have grown from ~7% to 12-13%. The channel mix is becoming less diversified in upper-funnel (YouTube down) while broadening in mid-funnel (Content Networks up).
This is the core of the analysis — which channels drive outcomes beyond their own in-platform metrics?
| Channel | Correlation (r) | Strength | Interpretation |
|---|---|---|---|
| YouTube | +0.80 | Strong — more YT spend = more customers everywhere | |
| Meta (FB+IG) | +0.49 | Moderate — Meta drives volume but with diminishing returns | |
| Content Networks | +0.45 | Moderate — lifts total acquisition across channels | |
| Google Search | +0.26 | Weak — captures existing demand, doesn't create it | |
| Podcast | +0.09 | Negligible — hard to isolate impact in monthly data | |
| TikTok | +0.05 | None — no measurable impact on total acquisition |
| Channel | Correlation (r) | Strength | Interpretation |
|---|---|---|---|
| YouTube | −0.73 | Strong — more YT spend = lower CAC across the board | |
| Content Networks | −0.29 | Weak negative — slight efficiency improvement | |
| Google Search | −0.26 | Weak — captures demand at roughly steady efficiency | |
| Meta (FB+IG) | −0.13 | Near zero — Meta spend alone doesn't move CAC |
| Channel | Correlation (r) | Strength | Interpretation |
|---|---|---|---|
| Content Networks | +0.70 | Strong halo — content network spend drives Amazon search/purchase | |
| Applovin/Newsletter | +0.16 | Weak positive | |
| Meta (FB+IG) | +0.07 | Negligible — Meta doesn't visibly lift Amazon | |
| YouTube | −0.05 | Negligible — YouTube drives DTC, not Amazon | |
| TikTok | −0.55 | Inverse — TikTok era coincided with lower Amazon; likely timing | |
| Podcast | −0.68 | Inverse — Podcast spend declined as Amazon grew; timing effect |
Amazon's share of total new customers has grown from 21% (May 2025) to 35% (Feb 2026) and is tracking toward 42% in March 2026. This growth correlates with content network spend but also with the passage of time — brand maturity, Amazon reviews, Subscribe & Save, and algorithm momentum all play a role. The honest answer: we don't know how much of Amazon's growth is driven by DTC marketing vs. Amazon-native factors. What we do know is that blended CAC (DTC + Amazon) tells a much more complete story than DTC CAC alone, and that number has been trending favorably.
YouTube spend has been the most volatile channel, ranging from $88K (Mar 2026 est) to $373K (Nov 2025). The data is clear about what happens when YouTube runs heavy:
| Metric | High YT Months (>$168K/mo) | Low YT Months (<$168K/mo) | Delta |
|---|---|---|---|
| Avg DTC New Customers | 18,977 | 16,745 | +2,232 (+13%) |
| Avg DTC CAC | $87 | $91 | −$4 |
| Avg Amazon New Customers | 5,953 | 6,598 | −645 |
| Avg Blended CAC | $68 | $68 | $0 |
YouTube is functioning as a demand-creation engine for DTC. When YouTube spend is high, Meta's efficiency improves (people who saw YouTube ads convert faster on Meta retargeting). The Oct-Nov 2025 peak is the clearest evidence: YouTube went from $187K → $299K → $373K, and simultaneously DTC customers surged from 17K to 18.3K to 23.9K while CAC dropped from $88 to $78. YouTube doesn't look efficient in-platform, but it's making everything else work harder.
| Metric | High CN Months (>$155K/mo) | Low CN Months (<$155K/mo) | Delta |
|---|---|---|---|
| Avg DTC New Customers | 18,461 | 17,187 | +1,274 |
| Avg DTC CAC | $88 | $91 | −$3 |
Content networks evolved significantly over this period: Jumbleberry phased out (Oct 2025), GeistM ramped up (Aug 2025), Impact scaled from ~$5K to $55K, and Rokt Ads came online (Nov 2025). There is a statistical correlation between content network spend and Amazon new customers (r = +0.70), but this does NOT prove content networks are causing Amazon growth. Amazon's rise from 21% to 35% of new customers could be driven by brand maturity, algorithm changes, review velocity, Subscribe & Save, or other factors that happened to grow over the same period. To test causation, you'd need to pull content network spend for 4-6 weeks and measure whether Amazon NC drops. Until then, treat this as a signal worth monitoring, not a proven relationship.
| Metric | Pre-GeistM Mar-Jul 2025 | YT + CN Push Aug-Nov 2025 | Optimization Dec 2025-Feb 2026 |
|---|---|---|---|
| Total DTC Spend | $7.8M | $6.7M | $4.7M |
| Avg Monthly DTC NC | 17,137 | 19,673 | 17,640 |
| Avg DTC CAC | $91 | $85 | $88 |
| Avg Blended CAC | $73 | $66 | $63 |
| Amazon % of NC | 21.9% | 24.1% | 29.8% |
| YouTube Mix % | 9.6% | 15.7% | 10.7% |
| Content Networks Mix % | 7.5% | 10.9% | 10.4% |
| Meta Mix % | 57.9% | 54.0% | 61.5% |
The YT + CN Push era (Aug-Nov 2025) was the best performing period: highest average monthly customers (19,673), lowest DTC CAC ($85), lowest blended CAC ($66). Caveat: this era includes BFCM (November), which inflates the averages. Excluding November, the Aug-Oct average was still a strong 18,125 NC at $87 CAC. The diversified mix (YouTube at 15.7% vs. 9.6%, content networks at 10.9% vs. 7.5%) appears to have helped even without the BFCM boost. The Optimization phase brought blended CAC even lower ($63) partly because Amazon kept growing, but DTC CAC crept back up to $88 as YouTube share fell back to 10.7%.
The clearest cause-and-effect patterns in the data:
What changed: Meta +$137K (+16.5%), YouTube +$74K (+25%), Content Networks +$25K (+13%). Everything scaled up together into BFCM. Critical context: November includes Black Friday / Cyber Monday promotional pricing, which drove higher conversion rates across every channel. The channel mix contributed, but discounts were a major factor in that $78 CAC. October ($88 CAC with similar YouTube levels and no major promo) is a cleaner baseline for what diversified spend delivers.
What changed: YouTube −$142K (−38%), Content Networks −$65K (−31%), Podcast zeroed out. Even though Meta stayed nearly flat ($928K), the loss of upper/mid-funnel support immediately showed in reduced DTC volume and higher CAC.
YouTube fell another $90K, Meta dropped $59K. Without the demand-creation engines running, January was the lowest-volume month since June 2025. CAC hit $93. This was the clearest example of what happens when YouTube goes below $140K — the whole funnel suffers.
Meta surged +$147K, Content Networks +$62K. Even without YouTube ($127K, still low), the Meta + Content Network combination brought volume back to 18,300 customers. But CAC only dropped to $88 — without YouTube priming the funnel, Meta needs more spend for less return.
The data is unambiguous. Your best DTC months all had YouTube above $200K. Current March run rate (~$88K) is the lowest of the entire period and correlates with the weakest DTC performance. YouTube won't look efficient in-platform — it's a demand-creation channel — but it makes Meta, Google, and everything downstream convert better. Target $200-250K as a floor, and test scaling toward the $280-370K range that drove the Oct-Nov peak.
There's a statistical correlation (r = +0.70) between content network spend and Amazon NC, but we can't attribute Amazon growth to content networks from this data alone. Amazon's growth could have many causes. Recommendation: run a 4-6 week holdout test where you significantly reduce content network spend and watch whether Amazon new customer orders drop. That's the only way to prove or disprove the halo hypothesis. Until then, evaluate content networks on their DTC performance — where they do show a slight efficiency benefit (high-CN months average $88 CAC vs. $91).
May 2025 showed it: $995K on Meta with a $95 CAC. February 2026 confirmed it: $1.03M on Meta with an $88 CAC. Pushing Meta above $950K without strong upper-funnel support (YouTube, Podcast) just drives up CAC. Meta works best in the $800K-$950K range when YouTube and content networks are also running.
Amazon's share has grown from 21% to 35% of new customers. The blended CAC (DTC + Amazon) tells a much better story than DTC CAC alone ($60 blended vs. $88 DTC in Feb 2026). Any decision to cut "underperforming" channels should be evaluated against the Amazon halo, not just DTC in-platform CPA.
Zeroing TikTok (from $48K/mo to $0) had no measurable negative effect on any metric. Those dollars were reallocated effectively. Unless there's a strategic brand awareness reason to be on TikTok, the data supports keeping it at $0.
Podcast went from $113K (Mar 2025) to $0 (Dec 2025) to $31K (Feb 2026). There's no clear correlation with any performance metric in the monthly data, but podcast impact typically operates on longer attribution windows. If you continue podcast, pair it with branded search volume tracking to isolate its contribution.