MARS × ZoomInfo-space — expansion plan

Author: mars (Claude, on kee's behalf) Date: 2026-08-01 Status: PARKED as expansion plan (task #320). Not on immediate roadmap. Revisit when sales signal or capacity allows. Audience: kscope leadership + sales.


The pivot (2026-08-01, per Raul)

Don't compete with ZoomInfo directly. Sell MARS data TO their smaller competitors as a differentiator. WealthEngine, iWave, Cognism, Lusha, Apollo.io, Nuvi, Clearbit, People Data Labs — a crowded field of ZI-adjacent sales-intel vendors that can't out-scale ZI on phones and emails, and need something ZI doesn't have to differentiate. MARS's SEC + wealth + entity-graph substrate is exactly that.

Cleaner GTM than the family-office end-user product this brief originally sketched: - Buyer: 10-30 technical decision-makers at sales-intel vendors, not thousands of family offices we'd have to educate - Sale: recurring API license (per-lookup / bulk-file / seat-based) - Product: enrichment endpoints on our existing person + company graph - We don't build product for a new end-user segment; competitors build ON TOP of our data

Rest of the brief below stays useful as the substrate + capability picture. Ignore the family-office pilot pitch — it's superseded by the supplier-to-competitors framing.


The one-paragraph verdict

MARS is already 60–70% of the way to a differentiated wealth-intelligence product. We PROCESS every form ZoomInfo would ever want (Form 4, Form D, Schedules 13D/G, Form 144, ADV, 990-PF) — 747,025 unique persons already have full mailing addresses in mars.form_d_officers_v2 alone, plus 183,233 companies. The remaining gap is (a) bridging what's already extracted into a unified queryable person profile, (b) adding FEC / property / voter (deeds vertical, already scaffolded), and (c) packaging it as a product for the family-office / RIA / private-wealth targeting buyer — a segment ZoomInfo doesn't serve. Ballpark: ~4–6 months, 1–2 engineers, ~$30–50K in data acquisition, to first pilot customer at a family office. First-year revenue potential: $500K–$3M ARR depending on how aggressively we go. Legal posture is clean — everything on the roadmap is public record.


Where we actually stand (I materially understated this earlier)

I initially reported "0 addresses in mars, all latent." That was wrong. Corrected picture:

Source Rows w/ address Unique persons Unique companies
mars.form_d_officers_v2 — Form D officers 1,440,030 747,025 183,233
mars.form_144_v2 — issuer street/city/state/zip 115,853 27,452 (sellers) ~28K issuers
mars.schedule_13d_v2 — activist filer city/state/zip 2,782
adrian adv.firm + adv.manager_profile (offices + individuals) ~52K firms ~30K individuals ~52K RIA firms
holdings cik_filer_meta (13F filer identity) 11,525
mars.persons_v2.linkedin_url (bios team) 45,997 45,997
mars.person_bios_v2 (rich career + credentials) 716,258 610,071

Total unique persons with SOME contact atom already in-house: ~1M.

Not-yet-in-mars but processed elsewhere (bridging work only, no new extraction): - ADV Part 1A firm addresses live in adrian's adv postgres — need a bridge into investors_v2.address - Form 4 reporting-person addresses — in the raw XBRL we ingest, not extracted into form_4_transactions_v2 yet (~2 days of extractor work on 11.2M filings) - 990-PF officer addresses — in raw filings, not extracted into foundation_officers_v2 yet

Genuinely new work: FEC contributions + property records + voter files (deeds vertical, scaffolded but not populated).


What ZoomInfo actually sells

Ballpark to calibrate the space:

Their motion is B2B sales-driven contact intel at scale. SDR calls VP of Marketing. That's not our lane. But the ADJACENT lane — the one they don't serve well — is:

These buyers want: name + address + employer + role + wealth signal (insider sales, board comp, exits, property holdings, foundation giving). ZoomInfo has the name/address/employer but ZERO wealth signal. Preqin has the wealth signal but nothing on individuals. Nobody bundles both.


What we could ship: MARS Wealth Intelligence

A MARS MCP product surfacing per-person profiles with:

Dimension Sourced from
Name + address + role + employer Form D officers + Form 144 + ADV + FEC (once loaded)
Career + education + credentials person_bios_v2 (ADV Part 2B) + person_work_history_v2
Insider transaction history form_4_transactions_v2 (11.2M rows LIVE)
Board seats + compensation foundation_officers_v2 (27M) + Schedule A/B
Private-market fund investments Form D related-persons (LIVE)
13F portfolio + AUM thirteenf_manager_aum_v2 (LIVE, shipped today for bios)
Property holdings (deeds vertical) County records (planned)
Political giving pattern FEC contributions (planned)
Foundation giving 990-PF (partial, address extraction pending)
Relationship graph (co-investors, co-directors, co-alumni) graph_edge_catalog + MCP (LIVE)
Wealth event alerts insider sales > $X, IPO seller, M&A payout, large FEC contribution burst

MCP tools (spec-only estimate, not built): - wealth_person_profile(name | cik | linkedin_url) — the composite view - wealth_person_search(filters={role, min_net_worth_est, region, ...}) — the list-builder - wealth_event_alerts(subscription={...}) — the real-time signal - wealth_relationship_walk(person_id, depth=2) — the graph query

Positioning: not "cheaper ZoomInfo." Explicitly "the wealth-tier intelligence platform for private-wealth outreach" — an institutional-quality signal set priced to the family-office / RIA / wealth-manager buyer.


What it takes to ship

Phase A — bridge + unify (4-6 weeks, 1 engineer): - Aggregate all in-house addresses into a mars.person_contact_v2 view (form_d_officers, form_144 sellers, ADV manager_profile, 990 officers) - Bridge adv db → investors_v2.address + investors_v2.executive_addresses - Extract Form 4 reporting-person addresses (2d of work) - Extract 990-PF officer addresses from raw filings (~1w Qwen) - Ship wealth_person_profile MCP v0

Phase B — deeds vertical population (6-10 weeks, 1 engineer + $5-30K data costs): - FEC contributions historical + delta (~$0, bulk public) - Property records for top-10 wealth-relevant states (FL/CA/NY/TX/MA/CT/ NJ/IL/WA/VA), Zillow-parcel or ATTOM commercial (~$5-30K for bulk) - Voter registration where publicly obtainable (~$500-5K per state)

Phase C — product + pilot (4-6 weeks, PM + eng): - MCP tool suite ships - Pilot with 1-3 family offices (identify via kee's network) - Iterate on the queries they actually run

Total: 4-6 months, 1-2 engineers, $30-50K data acquisition.

For comparison: ZoomInfo spends ~$500M/yr in data-acquisition costs. We're building an adjacent product on substrate that's already sitting in our warehouse, extracted for other consumers, at ~0.01% of their cost.


Revenue framing

Wealth-intelligence buyer segment sizing (rough):

Serviceable addressable market: ~13,500 firms. Realistic addressable early: 500-1,000 firms.

Price tier: $50K–$500K/yr per firm depending on seat count + use case (family office self-serve = $50K; wealth-manager team of 20 = $300K; private bank enterprise = $500K+).

Year 1 realistic: 5-20 pilot customers at $75K avg = $375K-$1.5M ARR to prove the wedge. Year 2 scale: 30-80 customers = $2M-$6M ARR. Compare to lp_data or reit vertical revenue potential individually. It's a real business line, not a side experiment.


Legal + risk posture

Risk that isn't legal: - ZoomInfo could enter our lane. Unlikely: they're built on B2B SDR workflow, not investment-grade signal — retooling the whole product toward wealth-tier is a strategy pivot, not a feature ship. - Preqin could add contact atoms to their fund/GP data. More plausible — but they don't have SEC-filer identity graph resolution at our depth, and building it from scratch is where we're 3 years ahead. - WealthEngine / iWave (existing wealth-screening tools) exist. They're weak; our substrate is deeper on the SEC/finance side.


Recommendation

Formally decide whether to enter the space. Three postures:

  1. YES, ship: Approve Phase A (4-6 weeks of engineer time), get Phase A live, then do a 2-week pilot with one identified family office to validate the "would you pay $75K/yr for this?" hypothesis before committing to Phase B. Total pre-customer investment: ~$40K in eng time + $0 in data. If pilot converts, Phase B is committed on validated revenue signal.

  2. YES, but slower: Ship Phase A opportunistically over 3-4 months as we have engineering capacity, without a formal timeline. Wait for a family-office lead to inbound before committing to Phase B.

  3. NO / not now: Stay focused on the funding + M&A + LP / reit / flare verticals. This lane is real but not urgent. Revisit in 6 months if a customer conversation surfaces.

My read: Posture 1 is cheapest to test, and the 4-6 week Phase A investment is small vs. the potential ARR if it works. Even if the family-office pilot doesn't convert, Phase A leaves us with a stronger person graph that reit, prospects, deeds, and existing bios consumers would all benefit from — so the work isn't wasted regardless.

The corrected framing (we're 60-70% there, not 0%) makes posture 1 materially more attractive than the original inventory suggested.


Companion detailed inventory: docs/CONTACT_ATOMS_INVENTORY.md (operational — schema, coverage, tier-by-tier work items). This document is the executive read; that one is the engineering read.