Unclassified

31 — Packaging Strategy

Status: CANONICAL · Commercial OS
Owner: Founder + Product Marketing
Last verified: 2026-07-21

Purpose

What goes in which box and why — the feature-gating logic behind the price book (30), the upgrade paths it creates, and the fences that keep them honest.

Target audience: Product (gate implementation), sales (upgrade motions), marketing (pricing page).

1. Packaging Logic: the Trust Ladder

The tiers are not "small/medium/large." They are stages of the trust journey — each tier's headline capability is the next step of trust a team takes with AI:

Starter   "Show me the truth."        → Assembled state, no AI judgment. The show-me moment.
Growth    "Prepare things for me."    → Twin observes/explains/prepares/suggests. Trust builds on evidence.
Command   "Act — with my approval."   → PROPOSE + governed execution + TruthLayer memory. Trust pays off.
Enterprise "Govern it at scale."      → Policy customization, SSO/SCIM, SLA, data residency.

This ladder mirrors the intervention grammar (NOTICE→PROPOSE) and the control grammar — the packaging is the product philosophy, which makes the pricing page a teaching asset, not a feature matrix.

Design rule: each tier must have exactly one headline answer to "what do I get by upgrading?" — Starter→Growth: your Twin joins the bench · Growth→Command: approved actions execute + memory compounds · Command→Enterprise: your policies, your scale, your compliance.

2. Fence Design (what gates, what never gates)

Never gated (any paid tier): number of connected departments' data into the Spine (truth is whole or it's not truth) · provenance/audit visibility · export (leaving is always possible — doctrine) · security fundamentals (tenant isolation, encryption). We do not sell trust back to customers at higher tiers.

Gated by tier (the honest fences):

Fence	Why it's the fence
Twin capability depth (none → prepare → propose+execute)	The trust ladder itself; also tracks our AI COGS ($0 / ~$0.80 / ~$1.20 per seat)
Workbench count (1 → 3 → unlimited)	Maps to department expansion — the growth path (46)
Connector count (3 → 10 → 25+)	Correlates with stack complexity = value; generous enough to never block the wedge (CS needs 4–6)
Entity volume (5K → 50K → 250K)	Company-size proxy, rarely binding in-tier by design
Sync frequency (4h → 1h → 15min)	Operational urgency premium; infra-cost aligned
TruthLayer governed memory	Command's crown jewel — memory that compounds is the switching cost, priced accordingly
SSO/SCIM, custom policy, SLA, residency	Enterprise-cost realities, standard enterprise fence

Fence integrity test (run quarterly): a fence fails if support tickets show users hacking around it (fence too tight) or if <5% of upgrades cite it (fence irrelevant). Failed fences get redesigned, not enforced harder.

3. Add-ons (deliberately minimal)

Add-on	Price	Rule
Custom connector development	$2,500 one-time + maintenance in subscription	Enterprise only; roadmap-eligible connectors done free on our schedule
Additional entity volume	$99/mo per +50K	All tiers
Premium support (Growth/Command)	$500/mo	Enterprise includes it

Nothing else. Add-on sprawl is how pricing pages die; new monetization enters as tier value first, add-on only if truly orthogonal.

4. Upgrade Motions (how packaging sells)

Trigger observed	Motion	Owner
Starter team views signals daily but acts manually	"Your team is doing the Twin's prep by hand — turn it on": Growth trial 14 days	Product-led prompt + CS
Growth account's suggestions frequently executed manually in tools	"You're approving in spirit — get the approval gate + execution": Command POV on one workflow	CS (46)
Second department asks for access	Workbench expansion → tier or seat growth	Sales + CS
Security review requests SSO/policy	Enterprise conversation	Sales

Packaging KPI targets: Starter→Growth ≥ 30% in 6 mo · Growth→Command ≥ 40% in 12 mo · add-on revenue ≤ 10% of ARR (if higher, packaging is leaking value into add-ons).

KPIs: upgrade rates by path; fence-integrity findings; pricing-page → signup conversion by tier (73); expansion ARR share of NRR (72).