Keith Millington.

BiP Solutions · Public sector procurement

Making the next task visible in a complex dashboard.

Redesigning Delta eSourcing dashboards around the different information and actions procurement buyers and suppliers needed to see.

Complex workflows · Team collaboration · Implementation

My role
UX Designer
Focus
Concepts, dashboards, prototypes and delivery
When
2021–2022
Delivery
Delivered with development and QA
Delta eSourcing dashboard interface with activity summaries, charts and tables.

The project at a glance

The challenge

Make organisation activity and pending procurement work understandable immediately after sign-in.

My contribution

Turned team research into three concepts, then developed buyer and supplier dashboards with the product owner, senior UX designer, developers and QA.

Delivery and evidence

Delivered a shared dashboard framework with content and filters suited to each audience.

01 · problem

An overview that hid the work to do.

Delta eSourcing is a product owned by BiP Solutions, used for public sector tendering and government contracts. If a school needs a roof repaired, it puts a tender on Delta setting out the requirements. Suppliers bid against it, and the school chooses from the bids it receives.

Its existing dashboard made it difficult to see relevant tenders, contracts and messages immediately after signing in. Users had to look elsewhere to understand what needed attention.

The platform ran on a codebase around twenty years old. Buyers and suppliers shared a need for visibility, but the details differed. A useful dashboard had to help each audience understand both its wider activity and its immediate tasks.

02 · research

Turning team research into a design direction.

The team used focus groups, questionnaires and Hotjar findings to understand the information buyers and suppliers wanted on arrival. I translated that research into three dashboard concepts.

Both audiences wanted visibility of organisation activity and pending work, but the priority content differed. Buyers needed an overview of tenders and contracts alongside messages and actions. Suppliers needed opportunities, live tenders and invitations as well as messages.

I developed the selected direction with the product owner and senior UX designer.

Dashboard areaBuyers wantedSuppliers wanted
Organisation Activity

Buyers wantedTender Types by Status and/or Value
Contracts by Status and/or Value

Suppliers wantedOpportunities
Live Tenders and their values

Pending Activity

Buyers wantedNew Messages from Suppliers
Actions required on Tenders

Suppliers wantedNew Opportunity Invites
New Messages from Buyers

Items are in order of preference within each area: the same containers, with different work inside them.

Listening to the helpdesk, not only the users

The product owner asked for a dashboard. Research with both sides showed it needed two views: buyers wanted live tenders with bid counts and type filters; suppliers wanted to see their bids, filter them and identify successful ones. Each needed an overview without opening every record.

I also sat with the Delta helpdesk team. They said their largest category of call came from people who could not see their own live tenders in one place. Staff described how awkward it was to tell customers there was no solution. This was failure demand — contact caused by the service missing something rather than by a genuine need.

With the legacy platform, including accessibility problems such as tables running off screen, I designed a new layer over data it already held. A rebuild was not on the table.

After launch, the helpdesk team reported that this category of call dropped away, with remaining contact relating to unrelated defects. That is their reported observation, not a measured reduction from call-volume data.

03 · concepts

Explore alternatives before committing to a shared structure.

I developed three concepts and took them to the product owner and a senior UX designer. After a few meetings we agreed on the third concept, which used a shared framework with audience-specific content rather than treating buyers and suppliers identically.

The core design decision was to use a consistent framework while allowing the content to reflect each audience’s work. That avoided treating buyers and suppliers as if they had identical priorities.

Original concept exploration. The selected direction was developed with the product owner and senior UX designer; this overview shows the breadth of exploration rather than a usability-test result.

Retrospective service view

Two audiences, connected through Delta

Scroll across to explore all stages → Lane labels stay in view.

01Identify and publish

Buyer goal

Get a compliant notice out to the right market.

Buyer actions

Defines the requirement, builds the notice, sets criteria and deadlines, publishes.

Buyer ↔ platformShared procurement information passes through Delta
Delta does

Notice creation, compliance fields, publication to the market.

Platform ↔ supplierOne platform connects both sides
Supplier actions

Not yet visible to them. The opportunity does not exist until it is published.

Supplier goal

Nothing yet.

Where it broke

Research evidence

Team research: focus groups, questionnaires and Hotjar. Placement within this wider lifecycle is inferred.

InferredTeam research, not mine

02Find and register interest

Buyer goal

Is anyone credible actually interested?

Buyer actions

Waiting. Watches interest build, with no clear read on who is serious.

Buyer ↔ platformShared procurement information passes through Delta
Delta does

Opportunity search, saved alerts, expression-of-interest register.

Platform ↔ supplierOne platform connects both sides
Supplier actions

Searches or receives an alert, reads the notice, decides whether to bid, registers interest, downloads documents.

Supplier goal

Is this worth the cost of bidding?

Where it broke

Nothing needing attention was visible after sign-in. Suppliers had to go looking for live tenders, invitations and messages rather than being shown them. The helpdesk absorbed the cost, reconstructing each caller's list by hand over the phone.

Research evidence

Supplier questionnaire preferences are reported in the research section above. Helpdesk interviews: the largest category of inbound calls was users who could not see their own live tenders in one place.

Helpdesk interviews: my primary research

03Clarify

Buyer goal

Answer once, fairly, to everyone.

Buyer actions

Receives clarification questions, drafts answers, publishes them to all bidders at once.

Buyer ↔ platformShared procurement information passes through Delta
Delta does

One clarification thread, published to every bidder for fairness.

Platform ↔ supplierOne platform connects both sides
Supplier actions

Submits questions. Reads answers given to everyone, including competitors' questions.

Supplier goal

Do I understand what they actually want?

Where it broke

Messages sat away from the work. A clarification answer is time-critical, and neither side saw it on arrival.

Research evidence

Messages named by both groups as priority dashboard content.

04Bid and submit

Buyer goal

Are the bids in, and are they complete?

Buyer actions

Waiting on the deadline. Cannot see part-complete submissions.

Buyer ↔ platformShared procurement information passes through Delta
Delta does

Document upload, version handling, hard deadline lock.

Platform ↔ supplierOne platform connects both sides
Supplier actions

Assembles the response across colleagues, uploads before the deadline.

Supplier goal

Get a complete bid in before the clock stops.

Where it broke

Pending actions were invisible. Both sides had to reconstruct what was outstanding from several areas of the platform.

Research evidence

Buyer questionnaire: stated preference for organisation activity and pending activity together. Finding and suggested action shown above.

05Evaluate and award

Buyer goal

Score defensibly and award on time.

Buyer actions

Opens submissions, scores against criteria, moderates, issues the award notice.

Buyer ↔ platformShared procurement information passes through Delta
Delta does

Evaluation workspace, scoring record, award notice distribution.

Platform ↔ supplierOne platform connects both sides
Supplier actions

Waiting, with no visibility of progress. Then wins or loses.

Supplier goal

Where do I stand?

Where it broke

No organisation-wide view. Buyers could not see tenders and contracts together; suppliers could not see how they were performing overall.

Research evidence

Supplier questionnaire: stated preference for an organisation performance overview. Finding and suggested action shown above.

06Manage and report

Buyer goal

Keep the record straight for audit.

Buyer actions

Manages the contract record and reports on activity across the organisation.

Buyer ↔ platformShared procurement information passes through Delta
Delta does

Contract record and organisation-level activity history.

Platform ↔ supplierOne platform connects both sides
Supplier actions

Delivers, or reviews the outcome and looks for the next opportunity.

Supplier goal

How are we doing across all our bids?

Where it broke

Reporting meant assembling it by hand from screens built for single records.

Research evidence

Stage inferred from the procurement lifecycle rather than from the research.

InferredTO VERIFY
Retrospective two-sided service map. Buyers above, suppliers below, the platform between them.

What the wider service view makes visible

The dashboard is a doorway to the stages, not a stage itself. This retrospective map places the dashboard in the wider procurement lifecycle. Both sides return to an overview to find work and then move into individual records. That helps explain why the dashboard needed to connect activity rather than act as another isolated screen.

The same structural preference, different content. The questionnaire findings above support one framework with two configurations. The shared containers stay consistent; the tenders, invitations, messages and actions inside them reflect each audience’s work.

Waiting is part of the service. The map includes buyers waiting for interest and submissions, and suppliers waiting for publication and award decisions. Those pauses make visibility of status and pending activity worth investigating.

The support desk exposed a gap analytics alone could not explain. Helpdesk staff described calls from people trying to find their own live tenders. That is failure demand — contact caused by the service missing something rather than by a genuine need. Their reported observation after launch provides qualitative evidence.

04 · decisions

Separate the overview from the action.

I designed buyer and supplier views for tenders, contracts, messages and pending activity. Views and filters helped users move between an organisation-wide picture and the items that required attention.

The shared design approach supported consistency, while the information presented reflected the different tasks on each side of procurement. The aim was to make relevant work visible on arrival, rather than require people to reconstruct it from multiple areas.

Supplier and buyer dashboard designs. The shared structure supports consistency while each view presents information relevant to its audience.

05 · delivery

Working through implementation constraints.

I connected the views and filters in an interactive prototype to work through how people would move between tender and contract information. The prototype was a way to communicate behaviour to delivery colleagues, not just a collection of finished screens.

I worked with backend development and functional QA as data-population and implementation issues emerged. The final design needed to make sense when connected to the product’s data, rather than only with the examples used in the prototype.

The delivered interface closely followed the final prototypes after those issues were addressed. My work continued through that translation from design into implementation.

Original connected prototype showing routes between dashboard views. The connections document interaction coverage, not proof of usability.

06 · outcome

A shared framework for two distinct audiences.

The dashboard brought organisation activity, messages and pending work into a more structured view, with filtering for tenders and contracts. It made the intended next action more explicit in the interface.

The interface was delivered with development and QA. The helpdesk team later reported that calls about finding live tenders dropped away. This is a qualitative observation, not a measured call-volume reduction. Further user evaluation, including Useberry, was planned but is not documented as completed; no quantitative post-launch impact measure is available.

07 · learning

What I would evaluate next.

I would test whether each audience can identify its next task and locate relevant activity without leaving the dashboard unnecessarily. Buyers and suppliers should be evaluated separately because a shared layout can still work differently for each group.

I would also check how the dashboard behaves with sparse and extensive activity, and how clearly its filters communicate the current view. These are the questions I would prioritise in a next evaluation, rather than infer success from the number of screens or prototype connections.

The six lifecycle stages shown in the retrospective service map reflect a general public procurement lifecycle rather than Delta’s own information architecture. They provide context for the dashboard work, not evidence that every stage was researched or redesigned.

Next case studyDistrelec