A product funding model is how an organisation releases money and people to product work, and the two forms in common use are project funding and product funding. Mik Kersten sharpened the distinction in his 2018 book Project to Product. He argued that an organisation running digital work through project budgets is paying for the wrong unit and measuring the wrong thing. Somebody else chose the model, usually years before the product manager arrived, and that choice sets the outer limit on how far the portfolio roles from the last page can move.
The sections below take project funding and product funding in turn. They then set out the three jobs an annual budget tries to do at once. Rolling allocation and metered funding follow. The last section gives what a product manager can change from inside a cycle nobody is about to abolish.
Project funding
Project funding releases a sum of money against a defined scope, a start date and an end date. Somebody writes a business case, a committee approves it, a team forms, the scope arrives and the team disbands. Kersten describes the pattern this way. Funding is locked for the fiscal year, so a shift in priority partway through either costs a fresh approval or waits for the next cycle.
Three consequences follow for a strategy, and every one of them shows up as a behaviour nobody chose.
- Scope becomes the unit of accountability. The team was funded to deliver a named list, so delivering the list is what success means inside the organisation, and evidence that the list is wrong arrives as a threat to the budget.
- Evidence arrives when the team has least permission to act on it. The money was committed before discovery began, so the cheapest month to change direction is the month the team is least allowed to.
- The team is temporary and the product is permanent. When the project closes, the people holding the knowledge move on, and whoever inherits the product inherits it without them.
A permanent product run by a string of temporary teams is the consequence that compounds. Every cycle adds a system somebody is maintaining without having built it. After a few years the cost of that estate sets the budget before any new work is even discussed.
Product funding
A standing team is the other arrangement. Product funding releases money to a team against an outcome, and that team stays with the product for as long as the product exists. Capacity is the unit being bought. An organisation funding a team of eight for a year has bought a year of that team's attention on a named problem, and the order of work inside that year belongs to the team.
Scope is the variable under product funding. That changes what evidence costs. Under product funding, a discovery finding that the planned feature will not move the measure costs a conversation inside the team. The same finding under project funding costs a change request. This is the mechanism the whole argument rests on, and it is why product funding depends on the leading and lagging indicators from the last module. A team funded against an outcome with no agreed measure of that outcome has a budget and no accountability.
| Question | Project funding | Product funding |
|---|---|---|
| What the money buys | A defined scope, delivered by a date | A standing team's capacity for a period |
| What the sponsor is committing to | A list of outputs and a cost | An outcome measure and a review rhythm |
| What happens at the end | The team disbands and people are reassigned | The team stays and the next period is agreed |
| What evidence against the plan costs | A change request, a fresh approval or a wait | A change of order inside the same funding |
| What travels upward | Share of scope complete against budget spent | Movement in the outcome measure and what it cost |
| Where it genuinely fits | Scope somebody else wrote, with a deadline in law or in a contract | Work whose right scope is discovered while doing it |
The bottom row matters more than most arguments about funding admit. A regulatory programme has its scope written by a regulator and its date fixed in law, so it is genuinely a project. Funding it as a standing team with an outcome measure buys nothing anybody needs.
The annual budget cycle
Both funding models usually sit inside an annual budget, and the annual cycle is where most of the damage to a strategy happens. Bjarte Bogsnes, who led the change at Statoil and chairs the Beyond Budgeting Round Table, puts the diagnosis in one line. A budget tries to do three jobs at once.
- It sets a target, which is what the organisation wants to happen.
- It states a forecast, which is what the organisation believes will happen whether or not anybody likes the answer.
- It allocates resources, which is the attempt to put money where it will do the most good.
One set of numbers cannot serve all three honestly. Because the manager is rewarded against the target, they have every reason to forecast cautiously and to ask for more resource than the work actually needs, and both of those distortions are perfectly rational. The forecast becomes a negotiating position, and the allocation becomes a claim staked in October against a year that has not happened yet. Bogsnes's remedy is to separate the three processes and let each run on its own cadence and its own numbers.
The Beyond Budgeting movement takes its name from that separation, and the practice is a good deal older than the name. Handelsbanken has run without a traditional budget since 1970, when Jan Wallander became chief executive and treated the budget as something to abolish outright. The Beyond Budgeting Round Table was founded in 1998 by Jeremy Hope, Robin Fraser and Peter Bunce. Hope and Fraser published Beyond Budgeting in 2003, and Statoil dropped its budget round in 2005 under the management model it calls Ambition to Action.
Rolling allocation
Once the three jobs are separated, allocation can run on a rhythm of its own. The forecast is refreshed every quarter and looks a fixed distance ahead, so it always covers the same horizon and never runs down to nothing in December, which is what an annual forecast has done by the autumn. Targets are set against a peer group or a trend, so a target survives a market that moved. And a request for funding is approved at the point the work is ready, which puts the decision next to the evidence for it.
Three things make rolling allocation work in practice.
- A cost ceiling, so the organisation keeps control of total spend.
- A set of decision rights naming who can release what without escalating.
- An unallocated reserve, because a rolling process with every pound already committed is an annual budget reviewed four times a year.
The reserve is the one most often missing. Without it a quarterly review becomes a quarterly recital of decisions somebody took in October, which is the annual cycle with more meetings attached to it.
Rolling allocation therefore lets a strategy change in April. Evidence about a strategy arrives on its own schedule. An organisation that can act on that evidence only in the autumn planning window has chosen to wait six months on average before responding to anything it learns.
Metered funding
Rolling allocation still funds work somebody has costed. When nobody can put a figure on the value at all, the arrangement has to change, and Eric Ries named the two versions in The Startup Way in 2017. Entitlement funding assumes that an approved team keeps receiving money, which turns the year into an exercise in protecting the allocation. Metered funding releases a small amount against a named piece of learning and releases more when the learning arrives, in the way a venture investor stages rounds and holds the next one back until the last one has shown something.
Melissa Perri applied the same reasoning to an internal portfolio in Escaping the Build Trap in 2018. She argued that an organisation should fund an idea in stages as the evidence arrives, and should treat every new idea as an investment decision. Metered funding suits the explore role exactly. A product in the defend role takes it poorly, because the work there is already known and its value comes from never stopping.
What a product manager can change inside an annual cycle
Most product managers cannot abolish the annual budget. Four things inside it are usually within reach.
- An envelope requested at the level of a problem. A request naming an outcome and a measure leaves the scope open, and a request naming twelve features has committed the year to twelve features.
- A named reserve inside that envelope for work the year has not identified yet, labelled that way in the request itself.
- Review points written into the request at the time of the request, so a decision to stop arrives on a date somebody has already agreed.
- Movement in the indicators reported alongside spend, so the next request is argued from evidence and the one after that is easier to make.
Each of these buys room to change direction while the money lasts. Room is worth having only to a team willing to use it, and the first use of it is saying no. A funded strategy that accepts every reasonable request has spent its room by April. Somebody therefore has to write the focus down as plainly as the work itself is written down.