There is a cost that almost every large organization incurs at the beginning of every significant project, and almost no organization measures it: the time a project lead spends finding the right people before the project can actually begin. This pre-assembly phase has no formal name in most project management frameworks. It does not appear on a project plan. It happens in the margins, in messages and calls and informal conversations, and its cost is absorbed into the general overhead of getting work done.
In a 5,000-person organization, a project lead managing a cross-functional initiative might need to locate people with three or four specific expertise profiles across departments they have limited prior relationship with. Our observations from early-access pilots suggest this search can take anywhere from five to fifteen working days, depending on the uniqueness of the expertise needed and the breadth of the project lead's informal network. That is a significant fraction of the early weeks of any project that has a defined delivery window.
What the search actually looks like
The process project leads follow to assemble expertise is strikingly consistent across organizations of different types and sizes. It follows a roughly predictable sequence that, because it is so familiar, rarely gets examined critically.
The project lead starts with their immediate network: people they know personally who might either have the relevant expertise or know who does. They make several contacts and receive several referrals. Each referral requires follow-up: a message or call, a description of what is needed, a conversation about whether this person is the right fit or whether they know who is. That conversation may yield another referral, or it may result in the discovery that this person is not the right fit but is the best available option given the timeline. At some point, the search stops, not necessarily because the best possible people have been found, but because enough time has passed that the project lead has to work with whoever they have.
The problem is that this process samples the informal network much more than it samples the organization as a whole. A project lead's referral network is a subset of the organization. It is biased toward people in adjacent departments, people of similar seniority, and people who are already well-connected enough to appear in informal referral chains. Individual contributors who do deep specialized work and have long tenure often have disproportionately relevant expertise for technical and operational projects, but they are systematically undersampled by informal referral-based search because they tend to have smaller informal networks.
The two kinds of cost
The first and most visible cost is the time spent searching. A project lead spending two weeks on pre-assembly is effectively not doing the project work those two weeks were supposed to support. In organizations that run many concurrent projects, this multiplies. If twenty project leads each spend two weeks per quarter on expert-finding searches that could be completed in a day with better information infrastructure, the aggregate cost is substantial, though it never appears in any budget line.
The second cost is less visible but arguably more significant: the cost of finding the wrong person, or finding the right person two months late. Projects staffed with the people who happened to be findable rather than the people with the best expertise are less effective than they would have been. Assumptions that get built into early project phases because the right expertise was not available when it was needed often persist through the project lifecycle even after the expertise becomes available later. The early knowledge gap shapes the design of the work in ways that are hard to reverse.
One operations team we worked with in early access was planning a compliance process redesign. During the pre-assembly search, the project lead identified two people from compliance and one from operations to form the core team. The knowledge map we built alongside this team subsequently showed that there was an individual in the IT security group whose expertise was directly relevant to the regulatory interface the team was trying to redesign. That person was not found during the original search. They were found weeks later, after certain design decisions had already been made. Some of those decisions had to be reconsidered. The project team attributed the rework to scope creep; the actual cause was a failure in the initial expert-finding process.
Why the problem gets harder as organizations grow
Expert-finding search is correlated with organizational size in a non-linear way. In a 200-person organization, almost everyone is within one or two degrees of separation from the relevant expertise. The informal network covers enough of the organization that search quality is reasonable. At 5,000 people, the informal network covers a much smaller fraction of the total expertise distribution. At 20,000 people, a project lead's personal network is a tiny sample of the available expertise, and the probability that the informal referral chain will terminate at the best-matched person is low.
Large Japanese enterprises, in particular, tend to develop deep expertise that is concentrated within business units and often within specific teams within those units. The tenure patterns typical of large JP organizations, where people spend substantial portions of their career in the same company, mean that experts accumulate deep knowledge over long periods. But that knowledge often stays within the domain where it was developed, because the organization does not have a mechanism for advertising its existence to the broader employee population.
What a better baseline looks like
We are not arguing that the informal network search should be eliminated. It serves social purposes as well as knowledge-finding purposes, and the relationships built during the search process have their own value. What we are arguing is that the search should not need to start from scratch every time.
An organization with good knowledge infrastructure gives a project lead a starting point that is much closer to the right answer before any informal outreach begins. A knowledge graph that maps what people have actually worked on, what expertise they have demonstrated in practice, and who they have collaborated with closely gives the project lead a filtered candidate set that reflects actual expertise rather than informal network proximity. The informal network then validates and refines that set rather than generating it from nothing.
The practical question is how to build that infrastructure without requiring a multi-year HRIS replacement project. The approach that has worked best in our early-access work starts narrow: map the expertise landscape for a specific function or initiative, using the data sources that are already available, and build the search habit from there. The value of even a partial knowledge map over no knowledge map is significant. Complete coverage is a longer-term goal. The first project kickoff that runs in two days instead of two weeks pays for the initial mapping work.
