R&D portfolio management: When to fund, stop, and scale projects

Most R&D portfolios have a funding gate. Few have a stop gate, and fewer still have a scale gate.

Projects clear the first stage-gate review, absorb budget for two or three years, and skip any second vote on whether the investment still earns its place. Stage-gate governance is built on milestone reviews that assess project feasibility at each gate, not just at intake.

McKinsey documented one food company running more than 560 active projects before it created a dedicated role solely to kill stalled ones. The portfolio dropped to just over 200 within three years.

R&D portfolio management runs on three decisions: fund, stop, and scale. Most companies build a process for the first one. The other two decisions are where R&D budget gets stranded.

R&D portfolio management centers on three decisions

Technology scouting surfaces candidates for funding research and development projects. Project management runs a single initiative once it is funded.

R&D portfolio management is broader than either activity. It centers on three decisions:

1. 1. 1. Fund. Commit budget to a new initiative entering the pipeline.

      2. Stop. Close an initiative and reallocate its budget.

      3. Scale. Increase resourcing for an initiative that is clearing its milestones.

Each decision keeps the portfolio aligned with business strategy and strategic objectives as those shift.

Most CTOs handle funding well. Few have a process for stopping or scaling.

R&D and technology leaders approve funding readily. Stopping or scaling an initiative asks for a different kind of judgment, and most CTOs and VPs of R&D have no formal process for it. That gap is where this framework starts.

This article lays out a working framework for all three decisions, the software capabilities that make those decisions repeatable at scale, and two client examples that show it in practice. A clear portfolio strategy determines how aggressively each check gets applied. Basic research earns its own place in the mix as a bounded allocation, exempt from near-term return scoring.

Stranded R&D investment carries a measurable cost

A stalled initiative clears every budget review on autopilot. It keeps drawing funding, immune to the criteria that would end or scale it.

PMI's 2017 Pulse of the Profession research put a number on this at the portfolio level: organizations wasted an average of $97 million for every $1 billion invested in projects that failed to deliver their promised benefits, based on 2016 survey data. That figure spans project portfolio management broadly, not R&D budgets specifically. R&D budgets that skip a stop decision run on the same mechanism.

Every dollar a stranded initiative holds is a dollar unavailable for the scale decision on a stronger initiative. A portfolio carrying three stranded initiatives already holds enough budget to scale a strong one.

The funding sits there until a stop review releases it. Risk mitigation depends on a diversified pipeline; three stranded initiatives concentrated in one thesis put more of the portfolio in jeopardy than the same three spread across different theses would.

The same effect runs in reverse for a scale decision left unmade. An initiative clearing its milestones but capped at its original budget grows slower than the market it is chasing.

That gap in growth is opportunity cost, invisible on any line-item loss report. A scale decision delayed by even one quarter measurably lowers the initiative's net present value and the strategic value it was funded to deliver.

Who owns fund, stop, and scale decisions

A stop or scale decision needs an owner separate from the team requesting the funding: a portfolio manager or a steering committee. IESE Business School research from Professor Luis Huete found that leaders and organizations with a track record of success have a harder time killing their own projects. The same logic runs in reverse.

A steering committee that scales an initiative on a sponsor's enthusiasm skips the same milestone check a stop decision requires. Risk tolerance varies by initiative, but the governance split stays fixed: project managers own delivery risk day to day, and the steering committee owns portfolio-level risk management against corporate strategy.

Ownership splits cleanly between two levels, matched to what each group can see.

R&D leads decide:

• • • Which proposals enter the pipeline

    • Which proposals get resourced first

    • How a project's scope adjusts between gates

The steering committee decides:

• • • Which stalled initiatives get killed at the next gate

    • Which initiatives get scaled with additional budget

    • How capital moves from a killed initiative to a new or scaled one

Exhibit 1: R&D leads shape the pipeline. The steering committee owns every fund, stop, and scale vote.

This split matters once two or more kill signals appear on the same initiative. An R&D lead can flag it. Only the steering committee can close it.

Quarterly reviews catch most stalled initiatives before the cost compounds much further. Monthly reviews work better for portfolios with short technology cycles, where a competitor move or a validated risk can date an initiative within weeks.

Faster cadences also catch resource constraints sooner, compounding the cost savings from an early stop. Both cadences need to cover already-funded initiatives as well as new proposals.

Fund, stop, scale: A decision framework for R&D

The framework turns separate authority into eight concrete checks. Each decision runs on its own criteria, with a specific trigger rather than a general principle to interpret case by case.

These checks turn portfolio management processes into a repeatable, informed decision-making system instead of a fresh judgment call at every review. The checks below follow the order a steering committee works through them: from a new proposal down to an initiative that fails every scale signal too.

What determines a fund decision?

1. Confirm strategic alignment against a named investment thesis. Score the proposal against a specific thesis, such as reducing dependency on a single supplier or entering a named adjacent market segment. Strategic alignment needs that level of specificity to score as a real criterion.
2. Confirm resource allocation before approving budget. Check whether funding this proposal displaces a higher-priority initiative already funded this cycle. If it does, treat both proposals as one tradeoff decision, scored against each other in the same vote.
3. Check the technology inventory for overlap. Search the inventory before approving budget for new capability. A proposal duplicating a capability the organization already holds gets merged into the existing initiative instead of funded twice. A proposal that fails a basic technical feasibility check never reaches this step at all, and technological feasibility gets the same scrutiny when the underlying science, not just the product design, is what's unproven.

What determines a stop decision?

Four signals recur across stalled R&D initiatives:

↓ Budget burn outpacing milestone progress

↓ Market assumptions invalidated since the last gate

↓ A lower priority score than newly funded competing work

↓ Two consecutive review cycles with technical risk unchanged

1. Trigger the stop vote on any single signal. Any one of the four justifies a stop vote alone. Two or more signals together should close the vote at the next scheduled review. A lower priority score is effectively a risk-return matrix judgment: a low-impact initiative loses its claim on capital the moment a higher-return one needs it.
2. Redirect the freed budget in the same cycle. Move it to a fund or scale decision within the same review. A stop decision that parks its budget for a future cycle just creates a different kind of stranded investment. Resource efficiency depends on this step specifically, since over-allocating talent to a low-priority initiative is the same waste as over-allocating budget to one.

What determines a scale decision?

The same four checkpoints apply in reverse:

↑ Milestones cleared ahead of schedule

↑ Market assumptions confirmed by new data

↑ A higher priority score than initiatives currently funded

↑ Two consecutive review cycles with measurable improvement in technical risk

1. Trigger the scale vote on the same threshold as a stop. Any one signal justifies a scale vote alone. Two or more together should trigger it automatically, the same rule that applies to a stop. Dynamic allocation matters most here: an initiative clearing every signal needs more resourcing before the opportunity narrows, not at the next annual planning cycle.
2. Size and approve the resource increase in the same vote. Set the exact resource allocation increase and approve it in the same meeting that grants the scale decision. An unsized scale decision usually stalls at the next budget cycle. Maximized ROI comes from sizing the increase to the opportunity, not to whatever was left in the budget.
3. Reset the technology inventory baseline. After scaling, reset the baseline for that initiative so the next review measures growth against the new scope. Regular reassessment of the technology inventory this way keeps the baseline's strategic value current instead of dated.
   
   

Exhibit 2: One entry point branches to fund, stop, or scale. The steering committee's four signals decide which path an initiative takes.

Competing proposals get scored against each other in a separate step that happens before any of these checks apply. That project prioritization step draws on different prioritization tools depending on the input data available: weighted scoring, Three Horizons, the Eisenhower Matrix, and Technology Readiness Levels each fit different decision types, covered in technology portfolio prioritization.

Software connects portfolio visibility to the funding decision

A fund, stop, scale framework runs on data that changes weekly: budget burn, competitive signals, technology maturity. Portfolio visibility keeps that data in front of the steering committee continuously between reviews. Structured processes built around that visibility reduce both cycle time and resource waste, since a signal that would otherwise wait for the next scheduled meeting gets caught the week it appears.

Exhibit 3: Ten capabilities turn a static tracker into a live, continuously updated fund, stop, scale review.

Ten capabilities separate a working system from a static tracker:

• Phases and gates that route the decision automatically. A configurable workflow moves each initiative through the same sequence of checkpoints, so a vote triggers on schedule instead of depending on someone remembering to raise it.
• Role-based permissions that separate authority for effective management. R&D leads update their own initiatives. The vote itself stays reserved for the steering committee, matching the split covered above.
• Budget figures tracked as fields on each initiative, aggregated automatically across the portfolio, with a full audit trail on every decision. A history of what was funded, stopped, or scaled travels with the initiative.
• Filtered views, relationship mapping, and portfolio analysis that surface overlap. Clustering related initiatives together, or tracing shared dependencies, catches a duplicate proposal before both get funded separately.
• Shared roadmaps, kanban boards, and dashboards for portfolio reviews. Milestones, phase status, and the full pipeline sit in one current view the committee works from live, during the meeting.
• Pipeline risk alerts that fire in real time, keeping portfolio performance visible between meetings. A budget overrun, a competitor move, or an invalidated market assumption reaches the steering committee the day it happens, ahead of the next scheduled review.
• A cited insights layer on each initiative. Standing questions, such as what is changing or what it means for the initiative, get answered automatically on a set refresh schedule, with every finding linked back to the patent, paper, or article behind it. Predictive analytics improves the data quality behind each finding over time.
• Comments and ratings that keep the reasoning with the record. The discussion behind a fund, stop, or scale call stays attached to the initiative, so the next review starts from what was already decided.
• Automated handoffs at defined milestones. When an initiative crosses a milestone, ownership transfers automatically and the incoming owner receives the context, tasks, and deadlines already attached, eliminating the manual data entry a stop or scale decision otherwise waits on.
• An agentic protocol connector (MCP) that lets an external AI assistant list, create, and update initiatives directly. A steering committee member can ask a connected assistant to pull draft proposals or summarize an initiative's discussion without opening the platform.

A system missing several of these still requires the steering committee to reconstruct the picture manually before every vote.

How Toyota and DRÄXLMAIER recover stranded investment

The fund, stop, scale framework holds up against two client cases already running it.

Toyota Motor Europe scores R&D proposals on one shared platform

Toyota Motor Europe runs virtual pitch events where R&D and production engineering teams submit new project proposals. Internal experts evaluate each submission on the same platform, and approved proposals convert directly into funded initiatives.

Toyota tested 40 innovation software systems before choosing ITONICS as its top-level innovation platform. More than 500 members across Europe now use it to manage over 200 R&D projects and score strategic value on the same platform, with further rollout planned across the organization.

DRÄXLMAIER cuts manual project reporting by 70 percent

DRÄXLMAIER Group centralized its idea management and pre-development workflows onto a single platform. Manual project reporting dropped by 70 percent, saving the Innovation Management department hundreds of hours a year. The team redirects that time straight into evaluating initiatives instead of compiling status updates.

That reporting time is exactly what a stop or scale review consumes when data arrives late. Removing the drag increases how often R&D portfolio management can actually make the call, since fund, stop, and scale decisions depend on the same reporting cycle.

ITONICS gives the three decisions full portfolio visibility

ITONICS is strategic portfolio intelligence built for these three decisions. It holds technology signals, pipeline status, and financial data in one system, so the fund, stop, and scale calls run on the same evidence base at every review.

Toyota and DRÄXLMAIER already run this cycle on live data. Fund, stop, scale runs as a recurring practice, checked against current numbers at every steering committee review.

Exhibit 4: Budget sits on the initiative record itself, aggregated automatically, so a stop or scale vote starts from the same figures every time.