Skip to content

Portfolio optimisation

You can’t fund every project. Portfolio optimisation finds the mix that delivers the most value from the budget you have.

  • Free 30-minute call
  • No commitment
  • You own the code
Typical industries
Public sector, Supply chain
Problem type
Resource planning
You receive
A documented model, run on your data

In short: A project portfolio optimisation chooses which projects to fund when the budget can’t cover them all. It selects the set of projects that delivers the most value and fits your budget and your staff. It also considers how projects depend on each other, how risky they are and how well they fit your strategy.

The problem

Forty proposals are listed in a spreadsheet, each with a score, and the money covers perhaps a dozen. You fund from the top of the list until the cash runs out. Two of the funded projects need the same three engineers. One small project is just below the cut-off, where the money ran out. It would have built the test bench that both of them need.

A list that is almost always longer than the budget

Almost every organisation with more than a handful of initiatives has a list of worthwhile work that’s longer than its budget. And there aren’t enough people or months to do it all. One good quarter of cash flow won’t change that. So the question that matters is which projects to fund.

Why ranking by one score fails

The usual approach ranks projects by one score and funds from the top. The score might be net present value (NPV): today’s value of the money a project should bring in. Money you get later counts for less. At 5% interest, 105 francs in a year are worth 100 francs today. It might be strategic value or expected revenue. Any such score tells you how good a project is alone. It tells you nothing about whether it belongs in this year’s portfolio.

Two projects can each look attractive and be pointless together, because they want to win the same customer with the same offer. A third may look modest and be essential, because two larger ones can’t start without it. A ranking sees neither case. It also misses the specialists that several projects need at the same time.

What changes

Today, you fund the top of the list and find in month three that two projects are waiting for the same engineers. Our model proposes a set of projects that fits the budget, the team and the order of work. The small project is in as well. It builds the test bench that two other projects can’t move forward without. You see why each project is in or out, and what swapping one would cost.

  • A model written around your own portfolio optimisation rules.
  • Results measured in the numbers you already track.
  • It slots into how you already work, and you can check every answer it gives.

What portfolio optimisation gives you

Portfolio optimisation chooses the combination of projects that gets the most value from the budget you have. For each project, we enter its cost, expected value, risk and strategic weight. Expected value is the average result once you count the chance of success. A project with a 50% chance of earning 4 million has an expected value of 2 million. We also enter which people several projects share, and which project has to be finished before another can start. You see how much the portfolio should earn. And you see how much of it is left if some projects fail. You also see how much budget you use and which strategic priorities are covered. We test it on this year’s proposals before anything is funded.

How it works

We write your resources, the work they have to cover and your rules into one model. You get a plan that says which resource does what and when. It’s the plan that fits your goals best, such as the lowest cost or the shortest waiting time. Every rule is kept at the same time, and that’s the part a spreadsheet can’t do.

Selecting a set of projects

This is a selection problem. A project’s value depends on which other projects you fund with it, so the decision is about the whole set. Project portfolio selection optimisation decides which projects are in and which are out, all at once. The model checks each yes or no together with all the others: does the whole set still fit the budget and the team?

Budget, resources and dependencies

The budget is the obvious limit: the selected projects must fit the available capital. In research and development (R&D), money is almost always the first limit you hit. Rankings miss the limit on people. Even with the money, you have a fixed pool of engineers or project managers. When two projects need the same specialists, both wait for them and finish late.

Dependencies are what surprise people. Project B only pays off if project A is finished first. If you fund B without A, the money spent on B earns nothing until A is done. The model treats capital, capacity and precedence as one problem. Precedence is the order in which projects must run. Each of these limits changes which projects can run together.

Why the best ratio still loses

Ranking by value per franc looks smarter than ranking by value. It still fails, because budgets come in whole projects. Take a budget of 10 million. Project A costs 6 and returns 9, the best ratio on the list. Projects B and C cost 5 each and return 7 each.

Fund by ratio and you take A, and then nothing else fits: 9 in value, with 4 million left idle. Fund B and C instead, and you spend all 10 and get 14. This is the 0-1 knapsack problem. 0-1 means each project is either fully in or fully out. The ratio rule is known to fall short on it. With forty proposals and shared engineers, the best set is hard to find by hand.

Return, risk and strategic fit

Return is the expected benefit: revenue, cost saved, capability gained. Almost everyone optimises it first, and most stop there. Risk reduces it, because R&D projects have very different chances of success. One may work out nine times in ten, another once in ten. Strategic fit is the part rankings miss entirely. Some projects give you know-how you lack or defend an important market. Such projects are worth more than their numbers show.

The method: integer programming

The method is integer programming. Integer means some decisions can’t be split. You fund a project fully or not at all, never 60 per cent of it. So each project gets a yes or a no. Choosing which projects get a yes within a fixed budget is the knapsack problem from the example with projects A, B and C. Its name comes from packing a bag with the most value without passing its weight limit. Yours adds limits on people and on the order of work.

We write your project selection as a mixed-integer program. Program here means a mathematical model, not software. It mixes yes-or-no choices with ordinary amounts, such as sums of money. You get the best portfolio that keeps every limit, with proof that none is better. Return and risk become separate goals, so you get the Pareto frontier: the portfolios where more return costs more risk. Our page on multi-objective optimisation shows how the plans on that curve are calculated.

Project returns are uncertain, so we calculate that curve across the full range of outcomes. Stochastic optimisation plans against many possible futures and how likely each one is. For each portfolio, it shows how often the portfolio earns less than planned, for example because two projects fail. In budget constrained portfolio selection, you then fund the set that still does well in the weaker years.

Product, public sector and R&D portfolios

The same method decides which products to build, keep or retire. Retiring one product can free the capacity another needs. Public sector programmes compete for a fixed budget under public scrutiny. Warehouse, transport and automation projects compete for the same investment budget. R&D portfolio optimisation selects the projects that bring the most expected value within your development budget.

Common questions.

Something else? Ask us directly.

Yes, and rough estimates are where it helps most. You give each project a range, low to high, instead of one number. A portfolio is the set of projects you fund. The model checks which portfolios are still good choices when the benefits turn out low and when they turn out high. Often the same core of projects is in every good portfolio. Then you only need to discuss the few projects that are sometimes in and sometimes out. For those, better estimates are worth the effort.

Yes. For each project, we enter how much it spends in each year. Each year gets its own budget limit. The model can also choose start dates, so a project that doesn’t fit this year may still start next year. That catches a common trap: a portfolio that fits the first year’s budget and then needs far more money in year two.

Then we tell the model that this project must be funded, and it picks the best set of projects around it. We compare that result with the best set without this rule. The difference shows how much benefit you give up for the decision. Sometimes the answer is very little, and everyone moves on. Sometimes it’s a lot, and the board deserves to see that number before it signs.

Sometimes it can, and if that’s the case, we’ll tell you. Standard software is built for an average operation, though. Yours runs on its own rules: your contracts, your exceptions, the things you know by heart. A custom model writes those rules in, and that’s where the better plan comes from.

You don’t need either to start. We build the model from the data you’ve already got, even if it lives in three spreadsheets and one person’s head. You get a system you run yourself. We train your team until they no longer need us.

You see a first result with the prototype, before you commit to anything bigger. We start with a free 30-minute call about your operation and the model that would fit it. Then we build a working prototype and run it on your numbers.

Tell us what you’re planning.

In a free 30-minute call with a specialist, you find out whether optimisation can help.

  • Free, no commitment
  • You talk to a specialist directly
  • A clear go or no-go