Maya’s AOS Moment – 1
Maya is the CMO of a four-year-old direct-to-consumer skincare brand. Annual revenue $14M, growing 22% year on year. Customer base 380,000, of which 95,000 are active in the trailing 365 days. Marketing mix: ~40% Google and Meta paid, ~25% CRM (email and WhatsApp), ~15% influencer, ~20% organic and SEO. AOV $68. Gross margin 62%. Median purchase cycle 60 days. The CEO has been asking, quietly but persistently, why CAC keeps creeping up despite the revenue growth.
Before AOS
On paper, Maya is winning. Revenue is up. ROAS is green across the major campaigns. CRM revenue is healthy — last quarter it grew 18% year over year. Marketplace presence on Amazon and Nykaa is rising. The board is happy.
But something nags her. Every quarter, blended CAC is up. The performance marketing team explains it as auction inflation. The CRM team explains their growth as a sign of the lifecycle programmes working. The marketplace team celebrates the Amazon shelf placement. Each function is showing green; none of them is explaining the CAC creep.
She reads the Tax-onomy essay over a weekend. The line that bothers her is simple: a brand can grow revenue and still be paying for Beta on credit. The seven-bucket framework names something she has been feeling for two years: not all revenue is equal. She commissions a 30-day AOS audit with her senior analyst and her ESP partner.
The audit findings
Four weeks later, the findings land on her desk. They are worse than she expected.
| Finding | What Maya expected | What the audit showed |
| Paid Repeat Leakage | ~15% | 42% |
| Effective Transaction Tax on CRM bucket | 5–10% | 23% after 18% average Offer Tax |
| Best customers sitting in B– | Small issue | 27% of the Best cohort |
| R1 historical revenue | Unknown | $8M |
| Time to second transaction | 42 days historically | 67 days now |
| Marketplace identity capture | Not measured | 4% |
The detail behind each row matters. Paid Repeat Leakage at 42% means the retargeting campaigns the team had been celebrating as efficient repeat-revenue drivers were largely bringing back customers who were already in the database — paying Google and Meta to do what email should have done. Effective Transaction Tax on the CRM bucket of 23% — against the 7% the team had been reporting — hides 16 points of Offer Tax from an average 18% discount sitting in line items the channel-cost ledger never saw. Twenty-seven percent of her Best cohort sits in B–. Roughly 24,000 customers — proven repeat buyers — are weakening on attention before any transaction signal would have shown it. The team’s segmentation had been treating them as healthy Best because their transaction count is high. The TAT shows them drifting.
R1 represents $8M in historical revenue. Sixty-eight percent of that R1 cohort has not been reached through CRM in the last 120 days. The brand is not failing to recover them — it has stopped trying. Time to Second Transaction has stretched from 42 days to 67 days over the last 18 months. The lifecycle programmes the CRM team has been celebrating are not, in fact, compressing the second-purchase cycle. They are running alongside a slowdown. Identity Capture Rate on marketplace transactions is 4%. Of every hundred new buyers the brand acquires through Amazon, Nykaa, and Blinkit, four are ever brought into the owned database. The other ninety-six will, if they buy again, have to be paid for again — to the platform.
Maya schedules an emergency meeting with her CFO. She walks him through the numbers. The CFO’s first reaction is to challenge the methodology. The second is to ask whether the lifecycle programmes that have been showing growth are real or measurement artefacts. The third is the question Maya was waiting for: “You’re telling me 42% of our paid spend is structurally avoidable?”
Maya’s answer: “Possibly. I want to test it.” That sentence matters. She does not overclaim. She asks for a pilot.
The audit also changes how Maya sees her own team. Performance marketing is not the villain — it has been solving the problems the rest of the system handed to it. CRM is not innocent either — it has been celebrating revenue that often required heavy discounts. Marketplace is not merely distribution; it is revenue that may never compound unless identity is captured. The problem is not a person. It is the absence of an operating system.
The 90-day pilot
Maya picks three plays. She knows she should pick two; she picks three because she wants to test the full Atrium plus Meridian recovery sequence alongside the easier deployments.
Play 6 — Shift Repeat Adtech to Owned. Suppression rule applied to retargeting campaigns: any customer with a transaction in the last 90 days is excluded from prospecting and standard retargeting. Spend redirected to a new CRM reactivation flow targeting the same cohort. Matched control group retained on standard treatment. Two weeks to deploy.
Play 4 — Protect Best from Becoming Rest. The 24,000-customer B– cohort identified, segmented into a test group (12,000) and a control (12,000). Test group taken off promotional content for 30 days, replaced with utility content (skincare routine recommendations, ingredient education, founder voice notes). Test group held in a designated “Relate” stream, control held in standard treatment.
Play 5 — Recover Rest Before Adtech. Top 10,000 R1 customers by historical LTV. 30-day Atrium pilot: daily NeoMails with magnets (skin quiz, ingredient quiz, mini-stories from the founder) and no transaction ask in weeks one and two. Reactivation offers introduced gradually in weeks three and four. Matched control of 10,000 R1 customers retained on standard paid-retargeting recovery flow.
She agrees the Beta baseline with her CFO: same period prior year, adjusted for category growth. Any incremental contribution profit above that baseline counts as Alpha.
