Visual planning for multi-project R&D teams
Multi-project R&D environments create a particular coordination problem: every team can be busy while the portfolio as a whole becomes slow. Within a multi-project R&D environment, a practical visual planning example starts with making that difference visible. A Kanban board is not merely a wall of coloured cards; it is a shared operating picture showing what is moving, what is waiting, and which delay belongs to the system rather than to an individual.
Project Visit, the Chalmers Electronic Lean Research Team (Celean), studies digital visual planning, Kanban, visual project management, enterprise systems, and lean processes. Its research and industry collaborations, including partners such as Volvo, make the case relevant beyond a generic project tool. In a research portfolio, the board must expose handoffs between experiments, platform teams, suppliers, and validation groups; otherwise teams optimise their own queues while the programme stalls.
Consider a fictional Australian engineering group, Northstar Systems, with a portfolio spread across electric vehicle charging, autonomous mining equipment, and solar microgrid controls. The programmes share firmware, test rigs, data engineers, and regulatory evidence, yet their project managers sit in Melbourne, Brisbane, and Perth and work with partners in Sydney. This example follows a weekly planning rhythm, a shared digital board, and a monthly flow review. It shows how visual planning can turn a complicated portfolio into a sequence of manageable commitments.
A realistic Australian R&D setting
Northstar’s engineering work resembles the Australian market, where companies serving mining, clean energy, defence, and medical technology combine laboratory development with remote field trials. Melbourne hosts the systems group, Brisbane contributes power electronics, and Perth teams validate ruggedised hardware near mining operations in the Pilbara. Sydney brings customers, funders, and systems integrators into delivery conversations. The result is a portfolio rich in technical uncertainty and poor cross-site visibility.
At the start of each sprint, the team gathers around a whiteboard and wall display in Melbourne, then joins colleagues dialling in from Brisbane and Perth. This familiar Australian hybrid pattern includes hot-desking, a short stand-up near the coffee machine, and a mix of in-person testing and remote software reviews. The board is updated as work happens, not reconstructed from memory on Friday.
Without a common visual language, the portfolio meeting becomes a sequence of slide decks. Each slide shows completed tasks, but not work waiting for a shared test rig, the firmware branch blocking two programmes, or the prototype that must travel from Melbourne to Perth. People attend prepared to defend progress while the real issue remains outside anyone’s immediate line of sight. A shared board changes the conversation from activity to flow.
Turning strategy into visible work
The first design decision is to define the workflow before arranging cards. Northstar uses four swimlanes—discover, build, validate, and release—rather than one for every department. A card represents a testable outcome and includes an owner, acceptance criteria, next action, due date, risk, and dependencies. Project and programme labels are tags, so cards still reveal a common bottleneck. This grammar helps new colleagues read the board quickly.
The physical board remains valuable as an information radiator for shared attention, but the digital version provides search, access, history, and links to evidence. The move from paper to pixels is well described in this review of visual project management, and it shows why a digital board should preserve the discipline of a physical Kanban method rather than simply reproduce sticky notes. Cards can carry photographs, test results, and decisions without becoming overloaded.
Each card is pulled into the next available slot, subject to a work-in-progress limit. If all three validation slots are occupied, a new validation card waits visibly instead of entering and stalling. The pull rule creates a constraint the team can discuss and redesign. A lean process therefore depends on limits rather than heroic multitasking. When a card changes state, the underlying enterprise system updates its owner and expected completion date.
Making dependencies visible
Dependency visibility matters because R&D work is often ready for one discipline but held by another. The autonomous mining team may have finished a camera algorithm while waiting for a battery enclosure, and the charging team may be waiting for the same enclosure to complete an ingress test. A portfolio board shows both waiting states only if cards carry a dependency type, needed party, and requested date. The team can then distinguish a short hold from a structural constraint.
Dependencies sit beside the affected card rather than buried in a spreadsheet, and a weekly cross-project forum reviews items that have aged beyond commitment. Owners on both sides of the handoff attend, so the discussion produces a named action and date. A colour convention marks blocked work in amber and overdue dependencies in red, with accessible labels and icons for users who cannot rely on colour alone.
The board also records why a dependency exists, reducing repeated negotiation. For the charging programme, an enclosure delay might be linked to an Australian aluminium supplier, a prototype tolerance, and a safety case. The card can attach the purchase order, drawing, and risk assessment. Because the work can involve personal information, access and retention rules should reflect the Privacy Act 1988 and Australian Privacy Principles where they apply, including information held outside the employee-records exemption.
Coordinating distributed teams
Coordination starts with recognition that the teams are not all online at once. Perth sits two to three hours behind Melbourne depending on daylight-saving dates, so a 9:00 a.m. Melbourne review would exclude useful field feedback. Northstar rotates the early session and publishes a digest for later arrivals. In industrial R&D, that is a direct scheduling decision, not a minor administrative detail. The board timestamps decisions and comments so late contributions remain visible.
Daily contact is deliberately short. Teams use a stand-up to name blocked work and coordinate the next action, then use the board for deeper issue discussion. Facilitators can find useful prompts in Themeraki People, especially for turning a busy review into a sequence of decisions rather than a round of status updates. A facilitator also keeps the meeting within its time box and records unresolved choices for the next touchpoint.
Digital access needs deliberate boundaries. Some prototype data cannot be viewed from a personal device, while remote contractors may need a summary that excludes sensitive details. Enterprise permissions, secure records, and auditable changes are therefore part of the visual method, not optional extras. Hybrid and flexible work arrangements should be reflected in team agreements and policies that align with the Fair Work Act 2009 and applicable awards. Where physical boards are used, safe access and setup are consistent with Work Health and Safety Act 2011 duties.
Measuring flow instead of activity
A visual system becomes meaningful only when the team uses it to improve flow. Activity measures such as tasks completed can make a congested board look successful, so Northstar combines a small set of operational measures. Median lead time shows how long an outcome waits from request to acceptance; cycle time separates active work from queue time; and the age of the oldest blocked item exposes persistent constraints.
At the monthly flow review, the team starts with the board, not a slide. It charts throughput and WIP over time, then selects one troublesome transition for an experiment. If firmware integration waits two days for a reviewer, the team may create a rotating review pair and a one-page decision checklist. Over the next four weeks, median integration cycle time falls from 2.4 days to 1.7 without adding headcount. These are illustrative figures, but the method is real.
The review also checks whether WIP limits are causing useful focus or artificial bottlenecks. Sometimes a limit should move from validation to design; sometimes unclear acceptance criteria make rework invisible. Quality signals, such as escaped defects or repeated test failures, sit alongside speed measures so faster motion does not hide poor outcomes. Team leaders discuss constraints as process issues, keeping the review candid and useful.
A practical starting recipe
Starting with a large enterprise rollout can turn visual planning into another software migration. A pilot is more likely to reveal whether the card grammar, limits, and meeting rhythm work in practice. Choose one product stream, establish a baseline, and run the routine for four to six weekly cycles before adding complexity. The following measures form a practical starting recipe for Northstar and comparable teams.
- Define one shared workflow and use the same card fields across every project; create a small legend explaining status, ownership, age, and dependency signals.
- Make each card describe an observable outcome with acceptance criteria, a named owner, the next action, and a date; move tasks to a project dashboard when they no longer need coordination.
- Set explicit WIP limits for the stages with the greatest constraint, begin conservatively, and review a limit when evidence shows work is blocked rather than focused.
- Reserve a short daily or twice-weekly cross-project touchpoint for dependencies, decisions, and aged cards; rotate the time and publish a digest so remote participants in Perth and Sydney are not penalised for their location.
- Link the board to the existing system of record for requirements, test results, risks, and purchase orders; allow the board to coordinate work without creating a second database that teams must reconcile.
- Review flow and quality every month, choose one improvement experiment, and record the expected result, owner, and review date; treat a failed experiment as information about the process rather than as a failure by an individual.
The value of this recipe lies in the feedback loop rather than the board technology. Each cycle tests whether the current card design supports the work and whether the coordination routine is removing or creating delay. Once the team can explain why a card moves, who acts next, and which constraint receives attention, the board can scale across a programme without losing its operational meaning.
Keeping visual planning alive
Long-term adoption depends on treating visual planning as a management habit rather than a project deliverable. Templates, shared definitions, training for new team members, and regular removal of stale cards keep the system trustworthy. Leadership also models the desired behaviour by discussing flow and dependencies without turning the review into individual performance theatre. When a policy changes, the board and enterprise process should change together, with the reason recorded for future users.
Northstar’s eventual result is not a perfect board but a portfolio that can absorb surprise. When a supplier delays an enclosure, the affected teams see the same constraint and meet it before a launch date slips. When a new regulator asks for evidence, the relevant card makes the owner, artefact, and next decision easy to find. The durable lesson is simple: visible active work and dependencies allow the next constraint to be addressed before it becomes a missed commitment.