Six months into a new initiative, someone in the business usually knows that one of the assumptions behind it is wrong. Customers are buying for a different reason than expected. The process takes twice the skill anyone planned for. The decision-maker turns out to be a different person. The people closest to the work see it first, because they are the ones meeting customers, installing equipment and handling problems every day.
Ask where that knowledge is supposed to go, and most businesses can name a meeting where progress is reported. Far fewer can name a meeting whose job is to ask whether the premise behind the initiative still holds, as opposed to whether the initiative is on track. Without one, the finding stays where it was found. It is mentioned, perhaps written down as a risk, and the business keeps funding an initiative whose premise it has already disproved, until the disproof shows up in the results.
This article explains why plans need to be able to learn from delivery as well as direct it, why that learning so often goes nowhere, and a few simple changes that give findings a destination: separating status from learning, naming who can change direction, writing down what each initiative is testing and measuring how long it takes the business to act on what it learns.
Plans flow down, but learning should flow up
The usual model of putting a plan into action runs downward. The owner or leadership sets goals, the goals become initiatives and the initiatives become tasks. That model is useful, but it is only half the picture.
Research by Peter Morris and Ashley Jamieson, published in Project Management Journal in 2005 and drawing on case studies of four companies, described projects as both deliberate and emergent vehicles for strategy. They are deliberate because they are formal ways of carrying out an intended plan. They are emergent because, once under way, they create new conditions and new knowledge that should shape the plan itself. Information and action flow upward from projects to the strategy as well as downward. The authors cautioned that their own top-down diagram failed to capture this back-and-forth.
In one of their case companies, a pharmaceutical business, portfolios were formally reviewed and rebalanced every six months. Results emerging from trials shaped portfolio strategy, and project leaders took an active role in proposing the next phase. That company had built something to receive what its projects learned. Many businesses have not.
The researchers also ran a survey. It was small and exploratory, and the authors were clear that its results could not be treated as statistically valid. The case studies are the stronger evidence, and one of their most practical observations was that where upward learning worked, it ran through business decision-making rather than only through project reporting.
Three ways the learning gets lost
It is treated as a reporting problem. Sometimes information is lost as it passes up through layers of management. That is a real problem, but it is not this one. Here, the finding often arrives intact. It is stated clearly, minuted and understood. It simply has nowhere to go, because no meeting is set up to act on it.
It is recorded as a risk. A risk register records things that might go wrong with an initiative. A finding that the premise is wrong is not a risk to delivery. It is something that has already gone wrong with the plan. Recording it as a project risk quietly turns a strategic discovery into a delivery problem and gives it to the person least able to resolve it, usually the project lead.
It waits for the annual plan. The plan will be revisited eventually, but often months later, by people working from different information. By then the initiative has usually adjusted its scope to fit the wrong premise, and the adjustment looks like an ordinary delivery decision rather than a change of direction.
There is also a common confusion about authority. The person sponsoring an initiative can fund it, limit it and stop it. But the authority to change what the business is trying to achieve may sit with someone else, such as the owner or a partners’ meeting. When that is unclear, sponsors quietly rescope initiatives to fit findings they feel unable to raise.
Every initiative is a test
It helps to treat delivery as a form of enquiry as well as implementation. Every initiative, whether or not anyone intends it, tests the assumptions behind it: whether customers will pay what was expected, whether the technology works as hoped, whether the business can absorb the change and whether the skills exist inside. It tests them with real money and real consequences, which often makes it more informative than any research the business could commission.
That evidence is produced whether or not anyone collects it. The difference between businesses is not whether their projects learn. They all do. It is whether anything is set up to receive the learning, and whether anyone has the authority to act on it.
A finding needs a destination before a channel
When businesses notice this problem, they often respond by improving escalation: telling people to raise concerns sooner or more forcefully. But escalating a finding to someone who cannot change the plan only moves it to a more senior place where it will also be noted and not acted on.
The prior question is: who may change what this initiative is for, and do they ever hear directly from the people doing the work? In a small business, the answer is often the owner, which seems simple. In practice, the owner may hear only status updates, such as on time, on budget, pipeline numbers, while the reasons behind the numbers stay with the project lead.
Separate the status report from the learning report
A simple convention makes a large difference: keep two questions apart.
- Status: is the initiative on track against its plan?
- Learning: does the premise behind the initiative still hold, and what have we learned that would change the case for doing it?
When the two are combined, the second is almost always read as the first. A finding that customers are buying for a different reason becomes “pipeline below target, more marketing planned”. Asked separately, the same finding becomes a question about direction.
The plans that detect rather than predict article describes how to write down the assumptions that make the learning question easy to answer.
Write down what each initiative is testing
For each significant initiative, write down the two or three assumptions it is effectively testing, such as:
- customers will buy for the reason we expect;
- the process can be delivered with our current skills;
- the cost per unit will be what we estimated;
- the buyer is the person we have been targeting.
Then, for each assumption, state in advance what observation would count as evidence that it is wrong. This takes a few minutes and turns delivery into a deliberate test rather than an accidental one. It also prevents later arguments about whether the evidence really counts.
Measure how long it takes to act
The most revealing measure is cycle time from observation to decision: how long it takes from someone noticing that a premise is wrong to the business deciding anything about it. Trace one real example from beginning to end. The answer is usually longer than anyone expects.
If the business reviews direction only once a year, a typical finding waits about six months before anyone with authority considers it. A shorter cycle, such as a monthly or quarterly learning review, means findings are acted on while the cost of changing course is still low. The loss from a slow cycle is not just the money spent on the initiative. It is the difference between changing course early, on an observation, and changing course late, on a result.
Who is in the room
What a meeting can hear depends on who is in it. If the people closest to the work never attend the discussion where direction is decided, their findings arrive as second-hand reports rather than arguments made by someone with standing. In a small business, that might mean inviting the project lead, the installer or the salesperson running a new offer to the meeting where the owner reviews direction, rather than passing their view through a manager.
The Morris and Jamieson case companies hinted at something similar. Some deliberately used the title “project leader” rather than “project manager” to signal that the role included shaping direction, not just executing tasks. Titles matter less than the underlying question: does the person who knows have a seat where it counts?
Why the people who know often say least
People doing the work may hold back findings about direction for understandable reasons. They may believe strategy is not their job. They may worry that questioning the premise will look like an excuse for poor results. They may have raised concerns before and seen nothing happen. Over time, they learn that the business wants status, not challenge.
The bad news early article looks at how businesses teach people whether it is worth speaking up. The changes in this article help by making the learning question routine, so raising a finding about direction becomes expected rather than exceptional.
A worked example
This is an illustration. An electrical contracting business with 35 staff launches a commercial solar and battery offer. The plan assumes that small business tenants in commercial buildings will buy systems that pay back within about five years, and that the business’s existing electricians, with some extra training, can deliver the work.
Monthly status reports go to the owner: number of enquiries, quotes and contracts. After six months the numbers are disappointing: 40 enquiries, but only 3 signed contracts. Each report notes “conversion below target, additional marketing planned”.
The project lead, however, has noticed something the reports do not say. Of the 40 enquiries, 28 came from tenants, and none of those converted, because tenants usually needed the building owner’s approval and did not stay long enough to benefit. All 3 contracts came from the 12 enquiries made by building owners, a conversion rate of 25%. The lead has written “tenant conversion low” in the risk register.
The owner makes three changes:
- A separate learning item is added to the monthly meeting: what have we learned that would change the case for this offer? The project lead attends and presents it.
- The assumptions are written down: customers are tenants, they decide on payback, the existing team can deliver. Each has a stated test.
- The cycle time is traced: the tenant pattern was first visible in month three but reached a decision only in month eight, a gap of five months.
At the first learning review, the business decides to redirect the offer towards building owners and property managers, and to check the accreditation and design requirements that apply to the systems it installs before scaling. Marketing spend aimed at tenants stops. The original plan was not foolish. Its premise was simply wrong in a way that only delivery could reveal, and the business now has a way to hear such findings within a month rather than half a year.
How this applies to a small Australian business
In small businesses, the distance between the person who knows and the person who decides may be short, but the meeting between them is often missing. Practical steps:
- Name who can change direction for each significant initiative.
- Separate status from learning in reviews, and ask the learning question explicitly.
- Write down what each initiative is testing, and what would count as evidence against each assumption.
- Trace one real finding from observation to decision and measure the time taken.
- Invite the people doing the work into the discussion where direction is decided.
- Review the risk register for entries that are really findings about the premise.
- Shorten the review cycle for uncertain initiatives.
Signals worth watching
- Risk register entries that are really statements about the business case.
- Scope changes without any change to the original case.
- Business cases whose assumptions have never been restated.
- Project leads who describe the plan as fixed.
- Long gaps between someone noticing a problem with the premise and anyone deciding anything.
- Status reports that explain poor results with “more effort planned”.
Common mistakes
- Treating findings about direction as delivery risks.
- Combining status and learning in one report.
- Escalating to people who cannot change the plan.
- Waiting for the annual plan to revisit direction.
- Keeping the people who know out of the room where direction is decided.
- Never measuring how long it takes to act on what is learned.
Frequently asked questions
Is this only relevant to large projects? No. Any initiative based on assumptions, such as a new product, service, market or system, produces learning that should shape the plan. In small businesses, the fix is often as simple as asking a separate question at an existing meeting.
How often should the learning question be asked? For uncertain initiatives, monthly is usually reasonable. For more stable work, quarterly may be enough. The aim is to keep the time from observation to decision short.
What if the owner is also the project lead? Then the risk is different: the owner may be too invested to see the finding. Ask someone else, such as a staff member, adviser or customer, to review the assumptions periodically and say plainly whether they still hold.
Won’t this make people question every plan constantly? Not if the questions are tied to written assumptions and agreed tests. People raise findings about specific assumptions, not general doubts.
What should happen after a finding is raised? A decision: continue, adjust, pause or stop, recorded with the evidence and the next review date. A finding that produces only discussion has still gone nowhere.
Questions to ask
- Who in our business may change what an initiative is for, and when did they last do so based on what delivery revealed?
- What are our main initiatives testing, and would we recognise evidence against their assumptions?
- How long does it take from someone noticing a premise is wrong to a decision being made?
- Which risk register entries are really findings about the plan?
- Do the people closest to the work believe they can say the plan is wrong?
- Is there any meeting where the person doing the work and the person deciding direction are both present?
Bringing it together
Initiatives do more than carry out plans. They test them, with real money and real customers, and the people closest to the work usually see the results first. Give that learning somewhere to go: name who can change direction, separate status from learning, write down what each initiative is testing, bring the people who know into the room and measure how long it takes to act. Otherwise, the business discovers its plan was wrong at the same moment as everyone else, from the results, rather than from the months of evidence it produced and did not use.
Source: KEVOS notes, drawing on P. W. G. Morris and A. Jamieson, “Moving from corporate strategy to project strategy”, Project Management Journal (2005). Examples and figures in this article are illustrations.
