Proposal · Draft v0.1
Pipeline Intelligence & Decision System · Pharma

The data exists 6666.
The ssssystessssm tssso act on it
doesn't.

Every asset in your pipeline has a story. Clinical performance, regulatory trajectory, competitive position, time to market. The data is there. What's missing is a single place where it comes together and tells you clearly which asset deserves your full attention right now. That's what we're proposing to build.

Engagement Decision Intelligence — Drug Development Pipeline
First system live 10 weeks from signoff
Prepared by Harken Health and Scene
Prepared for [Contact name], [Company name]
The problem

Manual, disconnected,
and rebuilt from scratch every time.

Pipeline evaluation is built around one person. Multiple sources, no common framework, output rebuilt for every leadership presentation. It works. But the decisions are only as good as what one person can cover — and nearly impossible to audit or repeat.

Now
Today
Disconnected sources. No framework.
Asset data lives in five places. Every leadership review is a from-scratch rebuild. There is no audit trail. Decisions ride on what one person can hold in their head at the time.
No common framework No audit trail Rebuilt every cycle Single-person dependency
→
With this system
Structured inputs. Scored on every dimension.
Every asset evaluated against the same framework. Ranked against your strategic objective in under a minute. Outputs ready for leadership without reformatting. Full record of how you got there.
Common framework Scenario-driven ranking Presentation-ready Full audit trail
The stakes. The decisions made here are consequential. Back the wrong asset and you have spent significant development resource on the wrong thing. Back the right one a cycle late and the cost is the same. This system doesn't eliminate that risk. It reduces it.
What we're building

Four components.
One coherent system.

Each component is necessary for the next to work. Together they make the difference between a spreadsheet and a system you can stake a decision on.

1
Foundation
Data ingestion and normalisation
Asset data comes in via manual entry or CSV upload — clinical trial results, regulatory assessments, market forecasts, competitive intelligence. The architecture is designed from day one to support automated source integrations. Everything is normalised into a common structure so every asset is comparable on the same terms, regardless of source or format.
Manual entry CSV upload Common schema API-ready architecture
2
Engine
Seven-dimension scoring
Each asset is evaluated across seven dimensions: clinical performance, market and indication, regulatory pathway, time to market, competitive landscape, unmet need, and relative performance against standard of care. The system scores each asset, surfaces where it is strong, where it is exposed, and how it compares to the rest of the portfolio.
Clinical performance Market & indication Regulatory pathway Time to market Competitive landscape Unmet need SoC comparison
3
Decision
Scenario-driven prioritisation
Choose a strategic objective — highest likelihood of accelerated approval, fastest time to market, largest commercial opportunity. The system adjusts its weighting and produces a ranked recommendation with a clear explanation of why each asset sits where it does. Change the objective, get a new ranking in under 60 seconds.
Accelerated approval Fastest to market Largest opportunity Custom weighting
4
Defensibility
Outputs and audit trail
Every evaluation is timestamped and stored. Outputs include a prioritisation matrix, radar charts per asset, and a scenario comparison view. Everything exports to PDF, ready for leadership presentations without reformatting. When a decision is questioned later, you have a full record of what data was used and what weighting was applied.
Prioritisation matrix Per-asset radar Scenario comparison PDF export Timestamped audit log
On the first version. The MVP supports up to 10 assets, manual entry and CSV upload, fixed or user-configured weighting, 2–3 scenario modes, and single-user access. Scope is tight by design — enough to validate the system with your real data before expanding. The fuller system is built on the same architecture and follows once the MVP is in use.
Investment

Two phases.
One commitment at a time.

Phase 1 exists to eliminate the uncertainty that makes Phase 2 hard to approve. Commit to Phase 1. At the end of it, you have a fixed price and a clear picture of what you are getting before you commit to it.

Total commitment to start
£15,000 – £20,000
Phase 1 only · fixed price for Phase 2 at end of Phase 1
Phase 1 — Validation & Solution Design 2–3 weeks. Led by Harken Health and Scene. Working session with your team, data model designed, architecture decided, user journey confirmed, feature spec finalised. Deliverable: a solution design document both sides sign off before build begins. £15,000 – £20,000
Phase 2 — MVP Build 8–10 weeks. Led by Scene. Data ingestion, scoring engine, scenario modes, visualisation, PDF export, audit snapshots, single-user access. Fixed price confirmed at the end of Phase 1. £90,000 – £110,000 indicative
Ongoing — Support and licence Monthly cost once the MVP is live. Includes hosting, support, monitoring, and roadmap work scoped against actual usage. £3,500 – £5,000 / mo
Decision today Commit to Phase 1 only. Phase 2 and ongoing rates are presented at the end of Phase 1, with a fixed price and complete scope. £15,000 – £20,000
On Harken Health fees. The figures above reflect Scene's build costs. Harken Health's consultancy fees are not included and will be presented separately. The total investment will reflect both parties.
Timeline

Three phases.
First working system in ten weeks.

Nothing moves to the next phase until the previous one is signed off. Each phase has a defined output — not a progress update, an output.

Phase 1 · Weeks 1–3 · Validation & Solution Design
A signed-off solution design document.
Working session with your team. Data sources mapped. Architecture decided. User journey confirmed. Feature specification finalised. Solution design document delivered and signed off before any build starts.
Data sources mapped Architecture decided Feature spec final Solution design signed
Phase 2 · Weeks 4–13 · MVP Build
A working system loaded with your data.
Iterative build with checkpoints every two weeks. Nothing is built past a checkpoint without your sign-off. End of phase: a working system loaded with your data, tested against your real workflow.
Ingestion live Scoring engine v1 Scenario modes PDF export Audit snapshots
Phase 3 · From Week 14 · Review & Expand
Roadmap shaped by real usage.
Monthly support and licence. Roadmap to the full system built around what you actually need after using the MVP, not what we guessed up front.
30-day usage review Roadmap defined Source integrations scoped
On your time. Phase 1 requires the most from you — a two-hour session and a small number of data and access decisions. Phase 2 requires checkpoint availability roughly every two weeks. The rest is on us.
From your side

Six things we need
to start.

Most of the work is ours. A few decisions and introductions need to come from your side. None block Phase 1 from kicking off — they shape it.

1
Which systems hold the asset data, and who controls access?
An inventory of where each input lives (Veeva, Spotfire, internal portals, spreadsheets) plus a named contact for each. Determines what we can ingest in Phase 2 and what stays manual.
Action required
2
Who defines the prioritisation weights — you, leadership, or a committee?
Decides who sets the scoring model defaults in Phase 1 and who can adjust them per scenario in Phase 2.
Decision needed
3
Static data snapshots per review cycle, or live integrations?
Snapshots are simpler and ship in the MVP. Live integrations are designed-for but scoped as a Phase 3 expansion. Your call on which the MVP supports first.
Decision needed
4
Who else needs access, and at what permission level?
MVP supports single-user access by default. If multiple roles need read or edit access at MVP, we add it to the Phase 1 spec.
Decision needed
5
Data governance and compliance requirements
Internal governance, residency, retention, and any regulator-facing constraints. Documented in Phase 1 so the Phase 2 build doesn't run into them.
Action required
6
Preferred tech stack or hosting environment
Cloud vendor, deployment model (VPC, isolated tenancy, on-prem), and any platform mandates. Drives the architecture decision in Phase 1.
Information needed
On data sensitivity. Everything stays within a closed environment. No proprietary data is shared, learned from across other clients, or used outside this engagement. This is a design principle, not a feature.