Ask a portfolio team how to prioritize projects and you will usually be shown a spreadsheet. Down the side, every candidate project. Across the top, five or six criteria: strategic alignment, financial return, risk, customer impact, sometimes urgency. Each criterion carries a weight, each project gets a score from 1 to 5 against each one, and the weighted total sorts the list. It is transparent, it holds up in a board meeting, and it takes the shouting out of the room.
It also has no clock in it. And the question it is being asked to answer is a question about time.
How do you prioritize projects?
To prioritize projects is to decide the order in which they receive your scarcest capacity. A useful method answers three questions: what each project loses for every week it waits, how much of your constrained team it needs, and how many projects that team can carry at once. Every other criterion is an input to those three.
The weighted scoring model deserves credit. It puts every project through the same criteria and replaces the loudest sponsor with a number, a real gain over ranking by instinct.
Where it breaks is the first of the three questions. If you want to know how to prioritize projects in a way that changes the money, the figure you need is not how valuable each project is but what it costs to wait for it. Flow Economics calls that figure drag cost: the value your organization loses for every week a project is not yet delivering. A scoring model measures worth. Priority is decided by the price of waiting, and the two can point in opposite directions.
Why does a weighted scoring model get the order wrong?
Look at what each criterion measures. Strategic alignment is the same whether the project starts this month or next spring. Financial return is usually a lifetime figure, which barely moves if delivery slips a quarter. Risk and customer impact are properties of the project, not of the calendar. Add them up and you get a timeless score: a number that would come out the same if the project started today or a year from now.
A timeless score can tell you which projects belong in the portfolio. It cannot tell you which one goes first, because first only matters when waiting costs something, and waiting costs different projects very different amounts.
Some projects lose almost nothing by waiting. A platform consolidation that saves $5,000 a week simply starts saving a week later. Other projects lose value they will never recover. An integration built for a partner's program earns only while that program runs, and every week late is a week of revenue that is gone for good. The two can carry similar lifetime values. Their drag costs can differ by a factor of four, and that difference is the whole of the sequencing decision.
What does getting the order wrong cost?
An illustrative example. Thornbury has two projects waiting for its data engineering team. That team is the constraint: the one group whose capacity limits how much the whole portfolio can finish. It can work on one project at a time.
Platform consolidation. Saves $5,000 a week once live, for at least 5 years. Needs 8 weeks of data engineering. Lifetime value around $1.3 million. Weighted score: 8.1.
Partner integration. Earns $20,000 a week, but only during a 26 week partner program that has already begun. Needs 6 weeks of data engineering. Lifetime value at most $520,000. Weighted score: 6.4.
The scoring model puts the consolidation first, and nobody in the room would argue. So the consolidation goes live at week 8 and the integration at week 14. Now count what the waiting cost: 8 weeks at $5,000 for the consolidation, plus 14 weeks at $20,000 for the integration. That is $320,000.
Reverse the order. The integration goes live at week 6 and the consolidation at week 14: 6 weeks at $20,000, plus 14 weeks at $5,000. That is $190,000. Same team, same 2 projects, same 14 weeks, and $130,000 less value lost, by doing first the project that scored lower.
Then there is the option most portfolios actually take. Both projects are labeled high priority, so the team is split across them, and neither finishes before week 14. Even ignoring the cost of switching, the drag cost is 14 weeks at $5,000 plus 14 weeks at $20,000: $350,000. Running both at once is worse than either sequence, including the wrong one. The ranking never got the chance to matter, because nothing was ranked out of the team's week.
Why can't project-by-project scoring see this?
Because each row in the spreadsheet is scored as if it were the only claim on your people. Flow Economics calls that assumption the Project Illusion: the belief that if every project is assessed and run well on its own terms, the portfolio will perform well.
A project's alignment and lifetime return are facts about that project. The cost of making it wait, and the delay it imposes on everything queued behind it at the constraint, are facts about the portfolio. A spreadsheet of independent rows has no column for them.
Isn't urgency already one of the criteria?
The objection you are already forming: most scoring models include urgency, so time is already in there.
It is in there as one criterion among 6, weighted at perhaps 15% and scored from 1 to 5. That does two things to it. It compresses a fourfold difference in weekly drag cost into a difference of two or three points, and then it dilutes those points with five criteria that have no clock at all. At Thornbury, scoring the integration 5 for urgency against 2 for the consolidation, at a 15% weight, moves the gap by less than half a point. The integration still loses by more than a point.
Urgency scored as points is the right instinct measured in the wrong unit. Stated in dollars per week, it is not one criterion to be weighed against the others. It is the number the others are there to help you estimate.
How to prioritize projects in five steps
You do not need to abandon the scoring workshop. Change what comes out of it.
1. Name the constraint. Find the one team or skill most projects end up waiting on. If you are not sure which it is, How to Find the One Resource That Sets Your Delivery Speed sets out the method.
2. Use your scores to decide membership, not order. A timeless score is good at telling you whether a project belongs in the portfolio at all.
3. Put a dollar figure on waiting. For every project that passes, estimate its drag cost: what it earns or saves per week once live, and whether a late week is postponed or lost for good. Rough figures are fine. You need the ranking, not the decimals.
4. Divide by time at the constraint. Drag cost per week divided by weeks of constraint time gives each project's return on your scarcest hours, the logic behind what Flow Economics calls value per constrained resource hour. At Thornbury that is $3,333 per constraint week for the integration against $625 for the consolidation. Highest goes first. Stop Ranking Projects by How Big They Are works through this step in more detail.
5. Draw the cut line. Decide how many projects the constraint can carry at once, start only those, and let the rest wait in order. A ranking that does not change what the constraint works on next Monday is not a priority list. It is a list. Start Less to Finish More covers how to set that limit.
A scoring model answers the question of what is worth doing. Prioritization asks a different one: what does it cost to make each piece of work wait? Answer that in dollars per week and the order of your portfolio stops being a debate. It becomes arithmetic.
Leave a Reply