How the platform extends from here.

Performix ships in stages, and each stage adds a layer of measurement rather than a new platform. MVP 1, live now, is the Protected Team Diagnostic: pick a team, run the protected survey, get one binding-constraint card. MVP 2 builds the Performance Science Library — the research source library, item bank, and evidence weights behind every diagnostic. MVP 3 adds the Capability Modeling Engine, so the Capability dimension scores against the specific basket of work rather than generic competence. MVP 4 adds Adaptive Measurement — item-performance tracking, recurring pulses, cross-company learning. V1 is the platform: integrations, benchmarks, predictive modeling, an advanced player, governance.

The order is deliberate. The diagnostic is dimension-agnostic, so it works before the deeper layers exist; each later stage sharpens the same instrument rather than replacing it. Nothing on this roadmap requires throwing away what shipped before it.

Underneath the product is Performix's research substrate — a set of independently-versioned measurement methods. Item-response theory for adaptive item selection. MaxDiff with balanced incomplete block designs for forced-choice trade-offs. Multinomial logit estimation. Monte Carlo with EVPI/EVSI for value-of-information gating. Wilson confidence intervals for the result. Cronbach's alpha for reliability. HMAC tokenization for protected feedback. These are real psychometric and statistical methods, not a language model dressed as one.

The substrate is the part that is hard to build and hard to copy — and the part most performance tools simply do not have. Performix owns it end to end: the methods and the measurement, and the playback on top — the composition, the player, the action loop.

Sales performance variance, AI transformation readiness, post-acquisition integration. Three buyers, three front-door messages, one diagnostic underneath. The CAMS conjunction-of-conditions engine is identical across all three; what changes is the buyer's question, the populations sampled, and the comparative cuts in the executive briefing.

The doors are commercial entries, not separate products — which is why the cost of opening a new one is a landing page and a set of comparative cuts, not a new build. The engagement counterpart of the same three doors is the consulting practice on peopleanalyst.com.

The UI is replaceable — any competent team can build a player. What is hard to replicate is the body of typed, validated methods underneath it, each independently versioned and independently consumable, and the privacy contract that sits as a primitive beneath all of them: min-N gates, HMAC tokenization, suppression rules, engineered so respondents cannot be re-identified and cannot believe they will be.

A competitor can copy the screens. They cannot easily copy the accumulated methods, or the protected-feedback primitive that makes honest signal possible in the first place. The defensibility compounds as the substrate grows.

Growth happens by adding methods, not by rebuilding. Capability modeling, adaptive measurement, new integrations, governance — each is either a new spoke in the substrate or a new playback surface over the same stored insights. Never a new platform.

Because every capability is modular, a new method lands as a new capability and a new surface lands as a new reader over records that are already computed and stored. The architecture is built to absorb additions, not to be torn down to accommodate them.

Each alone would not have been enough. AI made typed contracts legible at scale, so the substrate can be ingested and maintained economically. Cheap precomputed-analytics infrastructure made precompute-and-playback affordable. Organizations are fatigued with engagement-survey theater and ready for an instrument that names one thing. And the science of work-design measurement has matured enough to operationalize honestly.

The window is the intersection. The methods existed for decades; what just fell is the cost of assembling them into a usable, protected, precomputed instrument.

Enterprise analytics has no viable mobile path. Dashboards collapse on small screens; BI tools assume a desktop reader at a workstation. So the people who make the decisions — leaders, on the move — consume insights last, in a format built for someone else.

Performix delivers the diagnostic as a portable Insight Card in an iTunes-like queue on the phone: a now-playing surface, an up-next list, the evidence one tap underneath. The card is image-shaped, so it travels into email and slides too. Leadership consumes the diagnosis where decisions actually get made.