The hard part is between the pieces
A regional transportation study needs a rider survey, a data pipeline that joins fare and schedule records, a forecasting model, an economic analysis and a set of public meetings. A public health program needs community outreach, a case-tracking system, a dashboard for leadership and an evaluation of whether the outreach worked. Each piece is a familiar discipline. The project fails or succeeds at the places where they meet.
Program offices increasingly write requirements like these, because the problems they are solving cross disciplines. The buying question that follows is structural: buy each piece from a different vendor, or buy the whole thing from one lead that brings in the specialists. Both are legitimate. They put the integration work in different places, and that choice decides who spends the most time on the project.
This guide covers where the seams form in mixed work, who carries them in each structure, how to set up a single accountable lead, and how to evaluate an offeror's integration approach without being sold a diagram.
Where the seams form
Seams are the places where one discipline's output becomes another's input. In mixed government projects they show up in the same few places every time.
- Definitions. The survey team defines a trip one way, the data team another, the model a third. Each is reasonable. Together they produce numbers that do not reconcile.
- Data handoffs. Field data arrives in a format, structure or timing the system builders did not expect, and weeks go into cleaning that nobody priced.
- Schedule dependencies. The analysis cannot start until the survey closes, and the dashboard cannot be tested until the analysis produces outputs. A slip in one piece moves all of them.
- Quality standards. The research team documents its methods carefully; the software team tests its code; nobody owns the question of whether the final number shown to leadership is right.
- Change. A finding in the field changes what the system needs to capture. In a split structure, that change has to be negotiated across contracts.
- Accountability. When the result is wrong, each vendor can show its own piece met its own requirement. The program office is left holding the gap.
None of these are failures of competence. They are the natural result of capable teams optimizing their own piece. Somebody has to own the whole, and the structure of the award decides who that is.
Who carries the seams when the work is split
When a mixed project is split across several vendors, the integration work does not go away. It moves to the government, and usually to the COR. The COR becomes the person who reconciles definitions, sequences the schedules, translates between teams, chases the late handoff and decides whose problem a defect is.
That is a demanding role. It requires technical fluency across every discipline involved, time for a standing coordination meeting, and the authority to resolve conflicts between vendors who each answer only to their own contract. Most CORs carry several contracts and a program job besides. When the integration load lands on them, it competes with everything else, and the seams get attention only when something breaks.
A familiar example: the survey vendor delivers a clean file on time, coded to its own scheme. The data vendor loads it, finds that the geography codes do not match the administrative records, and asks the COR which to use. The analysis vendor, meanwhile, has already built its model on the administrative codes. Each vendor did its job. The COR now spends two weeks brokering a fix that none of the three contracts assigned to anyone.
There is also a contractual limit. Each vendor is obligated only to its own statement of work. When vendor A's late data delays vendor B, B may have a legitimate claim for schedule relief, and A may have met its own terms. The government absorbs the cost of the coordination failure.
Split and integrated delivery, side by side
| Dimension | Split across vendors | One accountable lead |
|---|---|---|
| Who owns the result | Each vendor owns its piece; the government owns the whole | The lead firm owns the whole, including the seams |
| Schedule | Several schedules, reconciled by the COR | One integrated schedule with cross-discipline dependencies built in |
| Definitions | Each team brings its own; conflicts surface late | One shared glossary and data dictionary, agreed early |
| Change | Negotiated across contracts, often with modifications to several | Absorbed inside one plan, with one modification when scope truly changes |
| COR workload | Coordination, translation and dispute resolution | Oversight of one plan, one report and one point of contact |
| Invoicing | Several invoices against several orders | One invoice against one order |
The integrated column is not automatically better. It concentrates risk in one firm, and it works only if that firm can actually integrate. That is why the evaluation of integration capability matters as much as the evaluation of any single discipline.
What one accountable lead actually provides
The phrase "single point of contact" appears in nearly every proposal and means very little on its own. One accountable lead is something more specific: a firm that is contractually responsible for the whole result, with a named person who leads the integration and the authority to direct every specialist on the team.
A single point of accountability. When the result is wrong, there is one firm whose job it was to make it right. That firm cannot point to another vendor, because there is no other vendor.
An integrated schedule. The survey close date, the data delivery, the model run and the dashboard test sit on one schedule, with the dependencies between them visible. When the survey slips, the lead re-plans the downstream work the same week.
Shared definitions. Before fieldwork starts, the lead produces one glossary and one data dictionary that every specialist uses. The survey instrument, the database schema and the model inputs all trace to it.
One report and one invoice. The COR reads one status report that covers every discipline and approves one invoice against one order. The administrative load drops in proportion to the number of vendors removed.
How to structure it
The structure that works is a lead firm with named specialists and partners. The lead holds the award, carries the integration, and performs a substantial share of the core work itself. Specialists and partner firms are named in the proposal for the disciplines the lead does not perform, with their people, hours and commitments shown alongside the lead's.
Three things make this structure hold. First, the integration lead is a named person with real hours, not a project manager title on the org chart. Second, the specialists are committed in writing before award, so the team the government evaluates is the team that performs. Third, the integration work is described as work in the statement of work, with its own tasks and deliverables, rather than assumed.
For the program office, the requirement should ask for the whole outcome and describe the disciplines involved, rather than writing separate scopes that each read like a standalone contract. Name the cross-discipline deliverables explicitly: the shared data dictionary, the integrated schedule, the end-to-end quality check. Those are the deliverables that a split structure leaves unowned.
One award does not have to mean one opaque price. Structure the line items by phase or by discipline so the program office keeps cost visibility: fieldwork, system build, analysis and integration can each be its own line, priced and reported separately, while the lead remains responsible for all of them. That gives budget staff the breakdown they need and keeps the accountability in one place.
Give the specialists a direct line to the integration lead and to each other. The point of the structure is that a field coordinator who notices a recurring data problem can raise it with the data engineer the same day, without waiting for a government coordination meeting. The lead sets the rules for those conversations and records the decisions that come out of them.
What an integrated plan looks like
The plan below shows how a lead might sequence a mixed project that combines community outreach, a survey, a data system and an analysis. The durations are illustrative; the order and the overlaps are the point.
An Integrated Plan for Mixed Work (Illustrative)
Notice step four. In a split structure, the analysis team often waits for fieldwork to close, and the dashboard team waits for the analysis. An integrated plan overlaps them, because the same lead can see that early data is good enough to start and can absorb the rework if it is not. That overlap is often where the schedule savings come from.
How to evaluate an offeror's integration approach
Every proposal for mixed work will claim integration. The difference between a real plan and a diagram shows up in specifics. Ask for them, and score them.
- A named integration lead with hours on the staffing matrix that reflect leading the whole, not occasional oversight
- Named specialists and partners for each discipline, with signed commitments and hours
- An integrated schedule showing cross-discipline dependencies, not separate schedules placed side by side
- A plan for shared definitions, with the glossary or data dictionary as an early deliverable
- A description of how data moves from the field to the system to the analysis, with the checks at each step
- An end-to-end quality step that verifies the final outputs against source records
- A clear statement that the lead firm is responsible for the whole result and for every specialist's work
- One example, from the offeror's own experience, of a seam that broke and how it was fixed
The last item is the most revealing. A team that has actually integrated mixed work can describe, without hesitation, the time the survey codes did not match the database or the model inputs drifted from the field definitions. A team that has not will describe its process instead.
Questions to ask in an oral presentation
If the evaluation includes orals or a technical interview, three questions separate integrators from assemblers.
"Walk us through what happens when the fieldwork runs three weeks late." A real integration lead answers with the downstream effects on each piece, what can be resequenced and what cannot, and what they would tell the COR in the first week.
"Who decides how a term is defined when two of your specialists disagree?" The answer should name a person and a process, and that person should be in the room.
"How will you know the number on the final dashboard is right?" Listen for tracing from output back to source record, reconciliation between the survey and administrative data, and a named owner for the check.
When splitting the work makes sense
One accountable lead is not always the right call. Split the work when the pieces are genuinely independent, with no shared data, definitions or schedule dependencies. Split it when one piece needs independence by design, such as an evaluation of a program or an independent review of a system, where the reviewer should not report to the team being reviewed. And split it when a specific piece requires a rare capability that no integrated team can credibly offer.
When you do split, budget for the integration work on the government side. Name a government integration lead with the time and technical range to do it, require each vendor to support a shared schedule and a shared glossary, and write cooperation obligations into each statement of work. The seams will still exist. The plan is to own them on purpose.
What to watch after award
In the first two months of an integrated award, look for the shared glossary delivered on time and used, a schedule that changes as one document when something slips, and status reports that discuss the seams, not just each discipline in turn. The integration lead should be the person in every coordination conversation, and specialists should refer cross-discipline questions to them rather than to the COR.
The clearest signal is how the first cross-discipline problem is handled. On a working integrated team, the COR hears about it from the lead, with a fix already in motion. If the COR finds out by noticing two specialists giving different answers, the integration exists on paper only, and the time to say so is then.
Frequently asked questions
A single firm that holds the award and is responsible for the whole result, with a named integration lead who directs every specialist and partner on the team. It provides one point of accountability, one integrated schedule, shared definitions, one status report and one invoice.
The lead's price includes integration work that a split structure leaves to the government. Compare the integrated price against the sum of the separate awards plus the government time spent coordinating them, and the cost of schedule slips at the seams.
Look for a named integration lead with real hours, committed specialists, a single schedule with cross-discipline dependencies, a shared data dictionary as an early deliverable, an end-to-end quality check, and a specific example of a seam the offeror has fixed before.
When the pieces are genuinely independent, when one piece needs independence by design such as an evaluation or an independent review, or when a piece requires a rare capability no integrated team offers. Budget for government-side integration when you split.