Pravesha Jain
All work
Case study · Research-led product design

Vendor & Contractor Management

When everyone owned a spreadsheet, no one owned the truth.

A research-led case study: how nineteen interviews reframed a dashboard request into a shared operating model.

I led discovery end-to-end, reframed the brief, and designed one shared operating model — with a focused view for every role.

My role
Research lead & design partner
Scope
Discovery → strategy → shipped product
Outcome
Shipped · in production use
Impact
Better visibility, less breach risk, faster contractor decisions.

Scope of the problem

Why this was hard

Across teams, the same data looked different — and decisions suffered.

Fragmented data

Vendor and contractor information lived across spreadsheets, emails and shared folders.

Low confidence

Every decision needed manual verification before anyone would act on it.

Manual, error-prone work

Reconciliation happened outside the system, by hand, every month-end.

No shared ownership

Nobody trusted one source enough to stop keeping their own.

12
stakeholder groups
6
user archetypes
1
shared operating model

The problem was never a missing dashboard. Trust in the shared data had broken down — so everyone rebuilt the truth by hand.

Finance Controller

Method

Five moves from insight to a defensible direction

  1. 01

    Understand

    On-site ethnography, role by role.

  2. 02

    Diagnose

    Data quality and broken hand-offs — not a missing screen.

  3. 03

    Set principles

    Trust, clarity, accountability.

  4. 04

    Explore

    Benchmark, concept, validate with real users.

  5. 05

    Prioritise

    Sequence by impact and effort.

Research synthesis

Nineteen interviews, clustered on one wall

Data confidence32 notes

Which number is correct?

P03

We don’t trust the source

P07

Different teams, different reports

P05
Fragmentation41 notes

Too many spreadsheets

P01

I keep my own report

P14

No single place to go

P11
Ownership28 notes

Not clear who owns what

P04

Finger-pointing, slow escalations

P08

Who changed what?

P02
Workflow friction35 notes

Re-entering the same data

P06

Approvals are slow

P10

No visibility into status

P13
What users need29 notes

One place for everything

P15

Role-specific views

P17

Real-time & trustworthy

P19

Fig.Reconstructed affinity wall — notes grouped into the themes that became archetypes and priorities. Dot-votes mark what the room agreed mattered most. Interview IDs anonymised.

Product thinking

How the problem reframed itself

  1. Initial assumption

    They need a dashboard.

  2. Research insight

    They have a trust problem.

  3. Product pivot

    Design a shared operating model.

  4. Final strategy

    A decision platform, per role.

Principles the evidence earned

Every principle traces to something we saw

01

Reduce ambiguity

Teams argued over which number was right.

One governed definition for every metric.

02

Respect role boundaries

People ignored information irrelevant to their job.

Dedicated role-specific workspaces.

03

Preserve flexibility

Every team had bent the process to fit reality.

Configurable views, not a rigid flow.

04

Governance in the workflow

Offboarding failed because it was an afterthought.

LDAP locks & approvals inside daily actions.

Strategy

Three strategic product decisions

01

Don’t replace spreadsheets — replace fragmented ownership.

Why

The spreadsheets were a symptom of broken trust, not the disease.

Business implication

One governed source the whole organisation could stand behind.

02

Design around roles, not data.

Why

One screen can’t serve six conflicting needs without becoming noise.

Business implication

A focused workspace per role on a shared spine.

03

Treat governance as part of the workflow.

Why

Offboarding and approvals failed whenever they were left to admins.

Business implication

Compliance and audit-readiness by default, not by effort.

One trusted source, many purposes

Role-specific views on a shared data spine.

Not one bloated dashboard — a focused view per role, drawn from one trusted, governed source.

One data spine. Many role lenses.

Each role gets what they need to act — nothing more, nothing less.

Pillar Head

Strategic overview

$98.7M+6%

contracted value

Finance Controller

Spend & accruals

$12.6M+5%

spend YTD

Delivery Manager

Performance & utilisation

85%+4%

utilisation

VMO Lead

Vendor & demand

342+7%

vendors · 11 at risk

HR Manager

Lifecycle & headcount

1,248+3%

headcount

Contractor

Personal profile

78%on track

my utilisation

Role lens layer
Personalized KPIsSmart filtersContextual alertsWorkflow actionsSaved viewsDrill anywhereExport & share

Trusted data spine

Governed, validated and continuously updated.

Contract & Terms
Vendors & Contacts
Workflows & Approvals
Financials & Payments
Performance & Risk
Documents & Files
Source systems
ERP / FinanceE-SourcingCLMProcurementSupplier PortalBI / ReportingOther Systems
Business impact

Single source of truth

Eliminates duplicate data and conflicting reports

Faster decisions

Role-specific insight cuts cycle time

Lower risk

Consistent governance and audit readiness

Stronger collaboration

Everyone works from the same trusted data

Information architecture

The structure held; only the fidelity changed

  1. SketchRapid exploration
  2. WireframeHierarchy stabilised
  3. AnnotatedInteraction validated
  4. Final UIVisual system applied

Fig.Sketch to shipped UI: hierarchy, navigation and information architecture stayed constant. The visual system came last.

The outcome

In use. Trusted. Adopted.

In use

The single source of truth for vendor and contractor data across the org.

Trusted

Teams act on the platform’s numbers instead of re-checking their own.

Adopted

Rolled out across regions — and the private spreadsheets were retired.

For the first time, we can answer critical contractor questions in minutes, not days.Pillar Head

Reflection

Research was the product decision.

I began believing the challenge was interface complexity. Research showed the real problem was organizational fragmentation — six teams, no shared model, and no reason to trust a central system.

So the work shifted from screens to a shared operating model: designing for confidence, governance and decision quality, not just usability. The UI came last — and only because the trust beneath it finally held.

One operating model.
One trusted source.
One product teams finally believed.

  1. Research
  2. Synthesis
  3. Product
  4. Governance
  5. Adoption

This project wasn’t about replacing spreadsheets. It was about giving six teams a shared system they trusted enough to replace them.