R&D Portfolio Displays and Strategy Tables
A grid, a curve, a segment table and a probabilistic forecast, each catching something the other three are blind to — plus the upstream device that decides what ends up on them. Portfolio level throughout; the per-project mechanics live elsewhere.
Three management roles, coupled by project decisions
R9 opens by splitting R&D management into three roles and claiming that all three must be performed at high quality for an organisation to reach its potential. The split matters because portfolio displays are instruments of one specific role, and are routinely misapplied to another.
THE THREE ROLES OF R&D MANAGEMENT
| Role | Definition given | Shorthand |
|---|---|---|
| Strategic management | Allocating resources | Doing the right R&D |
| Operational management | Execution of projects | Doing the R&D right |
| People management | Leadership, organisation, motivation, teamwork | Described in the source as possibly the most important of the three |
Close paraphrase of the source's own definitions.
Strategic and operational management are coupled by project evaluation and project decisions. The highest-level project decisions are go/no-go — selection, continuation, modification and termination — and project strategy decisions, meaning the definition of the project and its goals, are part of the same coupling.
That is the argument for portfolio displays in one sentence. If the two halves of R&D management meet at project decisions, then the instruments that inform those decisions are what connect resource allocation to what actually happens in a laboratory.
The four displays and the question each answers
Portfolio grid
Are we balanced? Probability of success against potential commercial value, one bubble per project, four named quadrants.
Productivity curve
Are we efficient? Expected return per project against cumulative cost to completion, highest return first.
Segment return analysis
Where does the return actually sit? Return, success rate and budget share by business and technology category.
Probabilistic revenue forecast
Will the pipeline deliver? Potential revenue against probability-weighted expected revenue, over the product life cycle.
The strategy table, upstream of every display
A portfolio decision is a choice among alternatives, and its quality is bounded by the alternatives available to choose between. A grid of twenty projects that were each defined by one team, unchallenged, is a precise picture of an unexamined set of options.
The strategy table is the source's answer to that. Every major product design decision becomes a column, a wide range of options is listed beneath each, and alternative product concepts are traced as paths across the columns. The construction mechanics belong with the other project-level tools and are set out in R&D project evaluation tools. What matters at portfolio level is what it produces: several coherent, genuinely different candidate definitions of the same project, before anything is funded.
The printed example is a strategy table for a medical diagnostic system, with columns for test menu, time to answer, ease of use, pricing, quality control and educating customers. Those columns are one illustrative case from 1994, not a template — the point they demonstrate is that six independent design decisions produce a very large space of possible products, and that a team which has never laid them out has not chosen among them.
In decision-quality terms this is the alternatives dimension being enforced at the point of funding — see Decision quality in R&D for the full set of six.
Display 1 — the portfolio grid
The grid plots probability of success on the vertical axis against potential commercial value on the horizontal, with each project a bubble. Its four quadrants carry names that have long outlived the article.
THE FOUR QUADRANTS
| Quadrant | Position | What it is telling you |
|---|---|---|
| Pearls | High probability, high value | The projects everyone agrees about. Their scarcity is the interesting number |
| Bread and butter | High probability, low value | Reliable, modest, and capable of filling a portfolio without advancing it |
| Oysters | Low probability, high value | Where the upside is. Also where an underfunded portfolio quietly stops competing |
| White elephants | Low probability, low value | Work that survives on momentum. The quadrant to explain rather than defend |
Quadrant names and axes are from the source. The right-hand column is our reading of what each position implies.
The stated purpose is balancing R&D effort across trade-offs such as risk against return, and the source's balance rule is directional rather than numerical: a portfolio should have few projects of low potential value and many of high potential value, and if the reverse is true management must investigate why. The portfolio in the printed figure is diagnosed as imbalanced — an illustrative client display, not a benchmark distribution.
Display 2 — the productivity curve
The productivity curve is the display most likely to change what an organisation funds. Each project's expected return — expected value divided by expected cost — is plotted against cumulative R&D cost to completion, with projects ordered from the highest return downwards. The result is a descending curve whose horizontal axis is the whole budget.
Building it
Compute expected return for every project
Potential commercial value multiplied by probability of success, divided by R&D cost to completion. The source calls this the productivity index: how efficiently a project uses R&D resources to create value.
Rank from highest to lowest
The ordering is the display. Nothing is grouped by business unit, sponsor or age.
Plot against cumulative cost
Each project occupies horizontal width equal to its remaining cost, so the axis accumulates to the total portfolio cost to completion.
Read the tail
Find where the curve crosses an expected return of one. Everything to the right of that point is expected to consume more than it returns.
The operational rule attached to the curve is deliberately not automatic. If a project's expected return is less than one, management must ask why — many such projects can and should be improved, and only some should be terminated. The curve identifies candidates for a conversation, not a cut list.
Display 3 — segment return analysis
The third display aggregates projects into categories — the whole portfolio, business areas, technology types — and reports potential return, expected return and success rate for each. The source describes it as a dollar-on-dollar return comparison across businesses and technologies.
THE PRINTED SEGMENT ANALYSIS — ILLUSTRATIVE CLIENT DATA, 1994
| Segment | Return if successful | Rate of success | Expected return | Share of R&D budget |
|---|---|---|---|---|
| Whole R&D portfolio | 8 to 1 | 50% | 4 to 1 | 100% |
| Core business | 4 to 1 | 75% | 3 to 1 | 45% |
| New business | 4 to 1 | 25% | 1 to 1 | 30% |
| New technology for existing markets | 6 to 1 | 35% | 2 to 1 | not legible in the source extract |
| Existing technology for new markets | 24 to 1 | 50% | 12 to 1 | 5% |
One organisation's illustrative figures, reproduced to show the display's structure. They are not typical values and no organisation should compare itself against them. Return if successful multiplied by rate of success reproduces expected return in every row, so those three columns are internally consistent.
Two blocks are flagged in the printed display, and they show what the analysis is for. New business takes 30 per cent of the budget for a 1-to-1 expected return, and is marked for re-examination. Existing technology for new markets returns 12 to 1 on 5 per cent of the budget, and is marked for expansion. Neither conclusion is available from a project-by-project view; both fall out of the segmentation immediately.
One caution on the success-rate column. Rates of success in a display like this are internal estimates for one organisation's own categories, not published industry rates — and how new-product success is defined and measured is itself contested, as Measuring new product success rates sets out.
Display 4 — the probabilistic revenue forecast
The fourth display combines time to completion and probability of success with peak sales revenue and product life-cycle information, producing two curves over the same horizon. Potential revenue assumes every project succeeds technically and is commercialised. Expected revenue is probability-weighted.
The printed forecast runs from 1991 to 2009, an illustrative horizon from a 1994 article that happens to show the display's real purpose: an eighteen-year view lets business managers see whether the pipeline will meet business goals, and lets deviations from long-term goals be spotted early enough to correct.
If the pipeline will not deliver what is needed
Reading the four displays together
Each display answers a different question, and each is capable of looking healthy while another looks alarming. Used together they cross-examine one another.
WHAT EACH DISPLAY CATCHES THAT THE OTHERS MISS
| Display | Catches | Blind to |
|---|---|---|
| Portfolio grid | Imbalance — too little high-value or high-risk work | Cost. A pearl and a white elephant occupy equal space regardless of what they consume |
| Productivity curve | Inefficiency — the tail returning less than it costs | Strategic fit. A high-return project can still be in the wrong business |
| Segment return analysis | Misallocation between businesses and technology types | Individual projects. A strong segment can contain weak work |
| Revenue forecast | A pipeline that will not meet the plan, early enough to act | Why. It shows the shortfall without diagnosing its cause |
The productivity curve carries one further instruction worth stating plainly, because it contradicts the most common response to a budget reduction. When budgets get tight, do not cut all projects uniformly; completely eliminate some of the weakest projects if they cannot be dramatically improved. The source's explanation for why uniform cutting persists is not cowardice but evidence: management rarely has credible assessments on which to base termination decisions. Building those assessments is the governance problem treated in Governing R&D decisions organisationally.
What to carry forward
- Strategic and operational R&D management meet at project decisions. Portfolio displays are the instruments of that joint, and belong to the strategic role.
- Generate alternatives before you display a portfolio. A grid of unchallenged project definitions is a precise picture of an unexamined choice.
- Rank by expected value divided by expected cost. The tail below an expected return of one is the finding, and it warrants a conversation rather than an automatic cut.
- Segment the portfolio. Misallocation between business and technology categories is invisible project by project and obvious in one table.
- Publish expected revenue, not potential revenue. The credibility cost of the optimistic curve arrives a few years later, reliably.
- Every number in the printed displays is one organisation's illustrative data from 1994. The structures transfer; the values do not.
Frequently asked questions
What is the difference between the portfolio grid and the productivity curve?
The grid shows balance — how projects are distributed across probability of success and potential value — and treats each project as a point regardless of cost. The curve shows efficiency, ranking projects by expected value divided by expected cost and laying them out across the whole budget. A portfolio can look balanced on the grid while a quarter of its cost sits below an expected return of one.
Should we terminate everything with an expected return below one?
The source does not say so. Its rule is that where expected return is less than one, management must ask why — many such projects can and should be improved, and only some should be terminated. The curve is a way of finding the conversations that need to happen, not a decision rule.
Can I use the segment return figures from the source as benchmarks?
No. They are illustrative data from one anonymised organisation in 1994, reproduced to show the display's structure. The internal arithmetic is consistent, which makes the table useful as a worked example, but the values carry no information about any other organisation.
Why does the source object to potential revenue forecasts?
Because potential revenue assumes every project will succeed technically and be commercialised, which it calls clearly unrealistic. Its observation is that well-managed organisations present that curve and lose significant credibility a few years later, when the pipeline delivers the probability-weighted result instead.
How does a strategy table relate to portfolio management?
It sits upstream. Portfolio choice is a choice among alternatives, so the quality of any display is bounded by the quality of the project definitions on it. The strategy table generates several coherent alternative concepts for the same project and exposes disagreement between functions before funding, which is why the source treats a shared product vision as a funding gate.
Do these displays require special software?
They were drawn for a 1994 article and none of them needs anything a spreadsheet cannot do. What they need is the underlying per-project estimates — cost to completion, probability of success and potential commercial value — which is the expensive part and the part the source spends most of its length on.
References and source attribution
- R9 - improving R&D decisions and execution. Practitioner methods article in a journal for R&D and technology managers, 1994; 8 printed pages; a synthesis of a decision-analysis toolkit illustrated with anonymised client displays, containing nine figures. Not empirical research.
- Eleven copyrighted journal articles on R&D project management, supplied as a reading set for a literature review and profiled for this library. Front matter, abstracts, framework sections, tables and figures were read; article bodies were not reproduced, and all content here is paraphrase.
- Figures reproduced in outline from R9: a strategy table for a medical diagnostic system; an R&D portfolio grid; an R&D productivity curve; a portfolio segment return analysis; and a probabilistic new-product revenue forecast. All values are illustrative client data, and one cell of the segment analysis was not legible in the extract used.
- Externally cited within R9 and not independently verified: a study by an industrial research association, cited to a 1984 reference, reporting that 9 of the 12 most important actions R&D organisations can take to improve productivity involve strategic management of R&D.
- Supplied teaching source for this library (research methods and research process materials). Used here for page conventions and voice only; it does not treat R&D project management.
Suggested questions for Ask KEVOS
- Turn my project list into a portfolio grid and tell me which quadrant is over-weighted.
- Calculate expected return for my projects and show me where the curve crosses one.
- How should I segment my R&D portfolio for a return comparison across businesses?
- Draft the expected-revenue version of a pipeline forecast I currently present as potential revenue.
- What alternatives should a strategy table generate for a product I describe?
- How do I explain to a board why we should cut some projects entirely instead of trimming all of them?