Twenty-nine architectural questions that separate a modern pharmaceutical evidence operation from a legacy one. Score where you are today, set where you intend to be, and see the gap on a single radar.
Most evidence-readiness conversations stop at the table-stakes questions. Do you generate RWE? - yes. Do you have a data license? - yes. Do you have a governance policy? - yes. Those answers no longer distinguish anyone. The questions worth asking in 2026 are about the operating model behind the evidence: whether your analytical capacity is elastic or bounded by hiring, whether what your teams learn compounds or walks out the door, whether you pay again to aggregate data you already own, and whether your audit trail is created as the work runs or reconstructed afterward. This assessment scores those questions - and shows you, honestly, what closes each gap.
The twenty-two dimensions below assume you have an evidence operation to assess. If these three prerequisites aren't in place, the architectural questions won't help yet - they'll just produce a low-score radar that tells you what you already know.
Proceed. The diagnostic will surface the operating-model questions your team probably hasn't been asked yet - and the ones where the gap is largest are usually where the value is.
Pause. The architectural questions won't help until the operation exists to assess. Establish the program, the evidence activity, and the ownership first - then the radar will mean something.
Each goal addresses a layer of the evidence operation where real variance still exists - where two organizations of the same size can look completely different, and the difference costs money. The dimensions are chosen because a low score has a concrete, addressable cause.
Two of these goals - lifecycle integration and institutional memory - are where organizations are quietly diverging today. The other five - the labor model, data ownership, defensibility, governed execution, and governance of the evidence your analyses produce - are where almost no one has a strong answer yet, because the tooling that makes a strong answer possible only recently arrived. Both kinds of gap are worth seeing.
Current state filled, target state outlined - the distance between them is the work.
Every dimension is scored 0–5 with half-point steps, twice: where you are today, and where you intend to be at a defined milestone - typically 12 to 24 months out. The gap between the two is the actionable artifact.
| Score | Level | What it means for this dimension |
|---|---|---|
| 0 | Non-existent | We don't do this and haven't considered it. |
| 1 | Reactive | We do this only when a problem forces us to. |
| 2 | Emerging | Some early practice exists; it isn't yet formalized or consistent. |
| 3 | Active | We do this consistently; it's documented and owned. |
| 4 | Proactive | We do this systematically; it's measured and improving. |
| 5 | Strategic | We do this as a competitive advantage; it's a stated priority. |
Two scores, not one. A 2.0 → 4.5 gap on When the audit trail is created is a different conversation from a 4.0 → 4.5 gap on the same dimension. The first is a structural change to how you work; the second is a refinement. The radar shows both, and the gap table below sorts them so the structural gaps surface first.
A note on honesty. This assessment is scored in your browser. We record the scores anonymously - no name, no email, no cookies - to learn where evidence operations are strongest and weakest (what we record). Defaults reflect a typical mid-size biopharma evidence operation; every one is adjustable. If your current scores come out higher than the defaults, that's genuinely good news, and the gap to StromaOS is smaller. The dimensions where the gap is largest are simply the ones where the operating model - not the effort of your team - is the constraint.
The generic scale above applies to every dimension. Governance has its own ladder, because the difference between levels is not how much effort you apply - it is whether control is written down, or actually executes.
Governance depends on individual judgment, manual review, spreadsheets, email and informal escalation. It works because specific people are careful.
Policies, SOPs, roles and review expectations exist, but execution is still manual and inconsistent. This is where most regulated organizations actually sit.
Core controls are standardized, process states are visible, and selected workflows use automated gates. This is the current industry aspiration.
Governance runs inside the workflow: the Sentry Agent evaluates policy as code before dispatch, quality gates block progression, and traceability is produced by the enforcement itself rather than logged beside it.
Controls adjust by risk, product, market, lifecycle stage, deployment posture and intended use - so a new obligation becomes a configuration rather than a rebuild.
Scoring goals five through seven honestly is usually the most uncomfortable part of this assessment, and the most useful. A strong score on lineage granularity alongside a weak score on policy execution at runtime is a common and revealing pattern: it means you can reconstruct what happened, but you could not have prevented it.
Sorted by gap size, descending. The dimensions with the most distance between where you are and where you intend to be are where investment lands - and, for most operations, where the operating model itself is the bottleneck.
| Goal | Dimension | Current | Target | Gap |
|---|
Read across your largest gaps and the same four operating-model questions keep surfacing. They aren't about your team's ability - they're about the architecture underneath it. The cards that light up are the ones your scores point to most strongly.
A maturity gap isn't abstract - it shows up on a budget. The panel below translates your radar into an illustrative annual value at stake, built from the same operating-model levers the dimensions measure. It's a starting frame, not a quote; the ROI and TCO calculators let you edit every number.
Figures are illustrative and for planning only - not a quote. They scale with the gap you scored and with a ~20-study-per-year program; heavier or lighter programs move them. Every input is editable in the ROI calculator. Each underlying default is an internal operating estimate rather than a published benchmark or customer study - the assumptions and methodology panel sets out the basis for every one.
Each bucket maps to a goal on your radar. The wider the gap on that goal, the larger the value StromaOS can convert - because a wide gap means the operating model, not the effort, is what's costing you.
Both modes use the same 29-dimension framework. The team mode is where it earns its keep - because the interesting signal is usually where two functions score the same dimension differently.
The assessment produces a real artifact - a radar and a sorted gap table you can take into your own internal conversation without us. When you finish, the ROI calculator pre-populates from your scores, so the "what would this be worth" question is already framed. Both are immediate, both are free, and neither asks for your email first.