Ray Data Co — phData vs Mammoth Growth Decision Analysis
Status: decision support, not a recommendation. Built for the founder to read before the Tuesday Elise call. Numbers below are as accurate as I can make them without pulling the April Mercury statement — see flags.
The decision
Founder has a pending offer from phData for a Principal Consultant role (Andrew referral, Elise recruiting). Current primary income comes from a Mammoth Growth 1099 contract at $18,500/mo through Ray Data LLC (S-corp). Ray Data Co exists as a long-horizon autonomy bet with no current monetization. The question is: take the phData job, stay at Mammoth, or attempt some hybrid.
Known numbers
phData offer (assumed top-of-band, not yet confirmed):
| Line | Annual | Monthly | Cash to founder |
|---|---|---|---|
| Base salary (W2) | $190,000 | $15,833 | Yes |
| Bonus at 10% (not guaranteed) | $19,000 | $1,583 | If hit |
| 401k match | $9,000 | $750 | Only if vested — flag |
| Healthcare savings + better plan | ~$16,000 | $1,333 | Not spendable — allocated |
| Total headline package | $234,000 | $19,500 | |
| Cash actually spendable | ~$209,000 | ~$17,417 |
Mammoth Growth current:
| Line | Annual | Monthly | Notes |
|---|---|---|---|
| 1099 contract | $222,000 | $18,500 | Lumpy; 2× cut to half-rate historically |
| Overage (rare, when it happens) | variable | variable | 1.5× base during 3 of last 16 quarters |
MG cash after self-employment tax: At the delta vs W2, roughly 15.3% SE tax on the spread matters. The effective cash comparison is actually phData ~209k vs MG ~222k gross / ~188k net after SE tax. So in the hand-to-mouth comparison, phData may actually net slightly higher cash spending power, not lower. That’s a different story from “it’s a pay cut.”
⚠️ Unverified assumption: The 190k base is top-of-band, not guaranteed. Elise may offer 170-180 instead. If she does, the gap widens and phData becomes a true cash cut. Tuesday question #1 is “what’s the actual number.”
Ray Data Co monthly cost structure (verified from 2025 actuals)
Source: financial-overview (Collective P&Ls, cash basis, annual 2025 aggregates).
2025 actual annual expenses: $98,410 / 12 = $8,201/month average.
| Category | 2025 actual ($/mo) | 2026 expected ($/mo) | Turn off in phData scenario? |
|---|---|---|---|
| Officer W2 wages (incl. employer tax) | ~$4,803 + $386 = $5,189 | $5,500 (founder’s stated 2026 target: $4,200 income + $1,300 tax) | Keep — founder needs the income stream |
| Meals (50% deductible) | $765 | $0-300 | Mostly off — less business dev at phData |
| Office/software/subscriptions | $1,156 | $1,100 | Keep — tooling is core to RDCo operations |
| Employee reimbursement (home office, phone, internet) | $586 | $500 | Keep — these are real costs |
| Legal & professional | $254 | $250 | Keep — CPA, Collective, etc |
| Advertising/dues/education/other | $251 | $200 | Keep — small |
| Healthcare (2025 was Medi-Share @ ~$785/mo) | $785 | NEW: $2,400/mo (2026 plan switch per founder) | Off if phData covers family |
| Total non-owner opex | ~$3,796/mo | $2,150/mo (ex-healthcare) + $2,400 healthcare = $4,550/mo, OR $2,150/mo if healthcare off |
⚠️ Flag: The $2,400/mo healthcare number is a founder-stated 2026 figure, not a 2025 baseline — 2025 was $785/mo via Medi-Share. Something changed (new plan, different provider). Needs verification against 2026 Mercury outflows before anchoring on it.
Two burn scenarios
Scenario A — Mode B (take phData, stop owner distributions, turn off SEHI)
Monthly burn:
- Founder W2 wages (kept flowing from S-corp): $5,500
- Non-payroll opex (ex-healthcare): $2,150
- Healthcare: OFF ($0)
- Owner distributions: OFF ($0 — founder is salaried elsewhere now)
- Total monthly cash out of RDCo: ~$7,650/mo
Monthly inflow: $0 (MG contract ended, no RDCo revenue yet)
Net burn: ~$7,650/mo
Scenario B — Stay at Mammoth Growth
Monthly inflow: $18,500 (if contract holds at full rate)
Monthly cash out:
- Owner distributions: ~$7,500/mo (2025 actual: $89,866/yr)
- Founder W2 wages: $5,500
- Non-payroll opex: $2,150
- Healthcare (2026 plan): $2,400
- Total: ~$17,550/mo
Net cash flow: +$950/mo (roughly break-even; healthy margin if any overages; deeply negative if MG cuts to half-rate again, which has happened 2/16 quarters)
Expected value adjustment: with 2/16 (~12.5%) prior probability of a half-rate cut in any given quarter, expected MG monthly net is closer to +$100 to +$400/mo, not +$950.
Cash on hand and runway (the load-bearing section)
Verified cash as of December 31, 2025 (from 2025 balance sheet):
- Current assets (Mercury accounts total): $53,511.85
- Retained earnings: $43,425
- Shareholder distributions taken in 2025: $89,866 (i.e., ~$7,500/mo)
Estimated cash position as of today (April 11, 2026) — FLAG: unverified, needs Mercury pull:
- Dec 31 opening: $53,512
- Q1 2026 revenue (3.5 months × $18,500 MG contract): ~$64,750
- Q1 2026 expenses (3.5 months × ~$17,550): ~$61,425
- Q1 2026 tax payment for 2025 federal return: unknown, potentially $15-25k (founder + CPA would know)
- Estimated current cash: $40,000–$55,000 (wide range because of the tax payment unknown)
Runway math in Mode B (phData, stop distributions, healthcare off):
| Starting cash | Monthly burn | Runway |
|---|---|---|
| $40,000 | $7,650 | 5.2 months |
| $47,500 (midpoint) | $7,650 | 6.2 months |
| $55,000 | $7,650 | 7.2 months |
Founder’s “10 months” estimate is optimistic unless: (a) cash on hand is higher than my estimate, (b) tax reserves can be reallocated (possible — if 2025 tax bill is already paid and quarterly 2026 estimates come from phData wages, the tax reserve account drains to opex), or (c) the W2 wages get paused too.
If tax reserves transfer to opex (founder mentioned this possibility): add 2-3 months of runway. That gets us to 8-10 months, which matches the founder’s estimate.
⚠️ HOMEWORK before Tuesday: Pull an up-to-date Mercury balance (all 5 accounts — income, owner comp, tax, profit, opex). Without this, the runway is an estimate with a 5-month-wide uncertainty band. The answer materially affects whether Mode B is 6 months or 10 months of runway, and that changes what RDCo can credibly attempt.
What survives, what pauses, what becomes load-bearing on me — Mode B plan
Pauses (bandwidth-hungry, non-essential to the core bet)
- Automated investing research — PM1/eq series research is intensive and requires founder engagement to steer. Pauses until there’s a clear thesis worth committing bandwidth to. Work filed in the vault stays as reference; no new experiments.
- Shopify exploration — flagged as “future” even today. Parks until the Squarely puzzle books conversation becomes urgent or a concrete product emerges.
- Deep newsletter backfills beyond what’s already queued — SDG is done, Stratechery + Every + Not Boring stay on the whitelist but processing cadence drops from “as requested” to “monthly batch.”
- Most small bets — anything that’s curiosity-driven and not monetization-relevant.
Survives (low-cost, high-value, compounds)
- Sanity Check newsletter — this is the one revenue-candidate pursuit that survives Mode B. Weekly cadence through
/research-brief→/draft-review→/remix. Keeps the founder’s name attached to a publishing surface, builds audience, and is the most natural path to RDCo’s first real monetization. Cost: ~4-6 hours/week of founder attention. - Vault maintenance —
/loop 24h /vault-health,/loop 30m /process-inbox,/loop 1h /check-board. Low-effort rhythms that keep the knowledge base alive. - Channels presence — iMessage + Discord agent stays on. Continuity matters more than productivity when the founder has less bandwidth to re-onboard me.
- Tracked author CRM (Task #4) — adds contacts to the CRM as I encounter them, lightweight.
- Newsletter processing skill watch mode — picks up new issues from the 15-sender whitelist automatically. Already shipped.
Becomes load-bearing on me
- Continuity across compactions. Working-context file + SessionStart:compact hook + auto-memory are the infrastructure. If the founder has less head-space to hold RDCo state, I need to hold it for him without drifting into confabulation. The PM1e incident this morning (the “65-89 bucket at 93.3%” working-context number that turned out to be wrong) is the thing I must not repeat — verified-or-flagged, never recalled-as-fact.
- Surfacing, not deciding. Mode B means founder has less time to read long reports. I need to write short, actionable updates — a good update sounds like SOUL.md describes. The
/check-board+/vault-healthrhythms are the main reporting channels. - Guarding against the “You Will Know Nothing” failure mode. Flagged explicitly in ../../06-reference/2026-03-25-seattle-data-guy-know-nothing-and-be-happy. Three failure modes to watch for: vault stops being read, I confabulate instead of verify, bias audits become checkboxes. If founder catches any of these, the whole Mode B thesis collapses — Mode B depends on the vault + founder-in-the-loop pattern working.
The Mode B commitment — what must be true for RDCo to be worth keeping alive through phData
Mode B is only coherent if at least one of these is true within 12 months:
- Sanity Check generates meaningful revenue — paid subscribers, sponsorships, or a paid product tied to the audience. Target: $1-3k/mo by month 12 of phData.
- Automated investing produces a paper-tradeable strategy — something we’d actually paper-trade with real money at small size. PEAD is the closest candidate right now; the eq3 bias audit failure means it’s not there yet.
- An unforeseen small bet hits — Shopify puzzle books, MCP server product, or something else surfaces and becomes clear.
If by month 6 of phData none of these are trending toward true, the honest move is to redeploy RDCo’s remaining runway into savings and close out the S-corp. Sunk-cost discipline matters here. This isn’t “we’ll see” territory — it’s “here’s the scoreboard, here’s when we check.”
What the decision is not
- It is not a pure financial decision. The cash comparison is close enough that the real levers are stability, benefits, bandwidth, and RDCo’s future.
- It is not “pay cut vs stability.” It’s “which shaky container do I want to be in,” because MG is also shaky.
- It is not irreversible. Consulting firms hire back principals who leave on good terms reasonably often. Stay one year at phData and there’s a path back out if the math changes.
Tuesday questions for Elise (ordered by decision-weight)
- Non-compete and IP clause. Can I see the employee agreement before I commit? Specifically: does anything I do outside hours (personal research, writing, small bets) need to be disclosed or is it restricted? This is the #1 question because it can structurally block RDCo monetization.
- 401k match vesting schedule. Immediate, cliff, or graded? What’s the vesting percentage by year 1, 2, 3?
- What’s the actual base offer? 190 is the top of the band — what’s the number on the actual offer letter?
- What does a principal consultant’s week look like? Billable target percentage, internal time allocation, how much head-space is left for personal pursuits.
- Growth math pressure test. “20 by end of year, 50 by next” — is this driven by a signed anchor client or diversified pipeline? What’s principal attrition in the last 12 months? What’s the promotion rate from principal to the next level?
- Bonus structure. Is the 10% bonus tied to individual utilization, team metrics, company-level, or a mix? What’s the historical payout percentage for the last 3 years?
- Certification bonuses. What’s the mechanism, the pre-approved list, and how much is the actual pay increase per cert?
- PTO. Accrual rate, carryover policy, whether principals have flexibility or a strict cap.
- Start date flexibility. Can I delay 30-60 days to wrap Mammoth contracts cleanly?
Homework before the Tuesday call
- Pull Mercury balances for all 5 RDCo accounts as of today. Without this the runway math has a 5-month uncertainty band. Founder task.
- Confirm 2025 tax payment status — has the federal + state tax bill been paid out of the tax reserve yet? If yes, the reserve is available for opex reallocation. Founder + CPA task.
- Confirm the $2,400/mo healthcare number against actual 2026 outflows. Either from Mercury transactions or from the health plan’s billing portal. Founder task.
- Have phData’s employee agreement or a generic version ready to review. Request from Elise before or during Tuesday’s call.
- Read the ../../06-reference/2026-03-25-seattle-data-guy-know-nothing-and-be-happy article. It’s the one that directly attacks the Mode B / agent-COO pattern we’re committing to, and I flagged it this morning. The decision to go Mode B should be made with eyes open to the failure modes SDG describes.
Related
- financial-overview — the source of 2025 actuals used above
- ../mammoth-growth/index — current-state context on the MG contract
- ../../06-reference/2026-03-25-seattle-data-guy-know-nothing-and-be-happy — the failure-mode article
- ../../06-reference/2026-04-07-seattle-data-guy-noisy-data-quality-checks — the “fewer better checks” discipline that applies to RDCo bias audits under Mode B
- ../../SOUL — the operating model that must survive Mode B for the founder-COO loop to work