How to Fix DCMA #2: Leads (Negative Lags) in P6

A negative-10-day lag isn't a head start. It's a lie about how the work actually happens.

I see leads more often than any other DCMA finding except negative float, and they're almost always cosmetic — someone shaved a few days off a critical path by telling P6 that activity B is allowed to begin ten days before activity A finishes. The math works. The construction doesn't. Check #2 exists because owners and reviewers got tired of being handed schedules where the dates only added up if you ignored gravity.

What Check #2 Actually Measures

The DCMA 14-Point Assessment is a schedule-health framework developed by the Defense Contract Management Agency. Most owners, prime contractors, and federal agencies now use it as the baseline standard for evaluating contractor schedules.

Check #2 is binary, just like Check #7: no incomplete activity should have a negative lag on any relationship. Zero tolerance. The DCMA guidance uses the term "leads" because in scheduling terminology, a negative lag is a lead — it lets a successor start before its predecessor is complete, in defiance of the finish-to-start convention.

A single negative lag in a 500-activity schedule fails the check. The threshold isn't 5%. It isn't 1%. It's zero.

There's a reason it's that strict. Leads are the single easiest schedule-shaping trick to abuse. Drop a -5d lag on a Finish-to-Start relationship and the successor magically begins five workdays earlier. Apply that to the right activity and you've moved your finish date a week to the left without changing a single duration, removing a single relationship, or actually re-planning the work. Reviewers know this. Owners know this. So DCMA decided that any negative lag is a finding.

The Four Reasons Schedulers Give for Leads — and Why Three Are Wrong

In the past decade I've heard every justification for a negative lag. They fall into four buckets.

1. "It's a known overlap — the steel detailing actually starts before engineering is 100% done." The most common defense and the easiest to refute. If the work genuinely overlaps, the relationship is wrong. Replace the FS with -5d lag with a Start-to-Start (SS) relationship and a positive lag, or split the predecessor into two activities — Engineering 1-80% and Engineering 80-100% — and tie steel detailing to the first one. Now the overlap is modeled, not faked. The driving logic is auditable. Check #2 passes.
2. "We can't change the relationship type because the corporate template forces FS." Real but solvable. If your template only allows FS, then your template doesn't accurately model construction. Get permission to use SS/FF/SF where appropriate, or split activities. Either way, the fix isn't a negative lag.
3. "It buys back the schedule — the owner needs a six-month date." The dishonest one. Negative lags used to recover a slipping date are exactly what DCMA is screening for. If you can't make the date with real logic, you cannot make the date — the negative lag doesn't change the work, it just hides the problem until the field can't keep up.
4. "It's how the original baseline was built and nobody wants to touch it." Schedule technical debt. Leave it long enough and nobody remembers why it was there. Then you inherit it, the auditor finds it, and you're explaining a logic choice you didn't make. The fix is the same as #1 — refactor the relationship type, accept the cost of the conversation.

Only the first one is recoverable without changing the schedule's substance. The other three are findings.

How to Find Every Lead in P6

You don't need a custom report. P6's relationship view has everything you need.

• From any activity layout, right-click the columns header → Customize Columns
• From the Activity Relationships section (the predecessors/successors detail tab), display the Lag column
• Open View → Group and Sort → group by Lag → sort ascending
• Any relationship with a negative number is a lead

Or, more reliably for a large schedule:

• Tools → Schedule Log (after running F9) tells you total relationship count
• Run an SDK / Claim Digger export to spreadsheet, filter Lag < 0
• Or use Project Reports → Schedule → "Lag Report" if your install has it

For schedules with thousands of relationships this is the only way to be sure you've found them all. Eyeballing the Gantt won't catch leads that are buried deep in the WBS.

This is also where automation pays off. ArgusForge flags every lead in seconds along with the predecessor-successor pair, relationship type, and lag value — the same way it flags every other DCMA check. Spend five minutes there, or an hour scrolling the relationships tab. Either way, find them all before the reviewer does.

How to Fix Each Lead

Work through these steps in order for each negative-lag relationship.

1. Step 1 — Ask whether the overlap is real.If the predecessor and successor actually do run in parallel for some portion of the work, you have a modeling problem (not a finding), and Step 2 fixes it. If they don't overlap in the field, the lead was always cosmetic — go to Step 4.
2. Step 2 — Replace FS with SS (or FF).For genuine overlap, change the relationship type:
   • FS -5d → SS +X, where X is the duration of the predecessor's portion that must complete first
   • FS -Xd spanning fabrication and delivery → FF +X if the successor finishes some time after the predecessor finishes
   This preserves the timing intent but uses honest logic. Reschedule, confirm the dates still make sense, and confirm the critical path didn't shift unexpectedly.
3. Step 3 — If SS/FF doesn't fit, split the predecessor.Sometimes the "overlap" is really part of the predecessor finishing while the successor starts. The right model is two predecessor activities with a clear handoff:
   • Engineering Drawings — Issue for Procurement (allows long-lead material orders to begin)
   • Engineering Drawings — Issue for Construction (gates the field work)
   • The long-lead order ties to the first; the construction work ties to the second
   This is more work to set up than a negative lag, but it's auditable, defensible in claims, and trains the field to think in stage gates rather than fuzzy overlaps.
4. Step 4 — If the overlap isn't real, remove the lead and let the schedule slip.Cosmetic leads exist because removing them makes the date worse. Remove them anyway. The new dates are the actual dates. If the new dates fail the contract, you're in a Check #7 (negative float) situation — see the recovery playbook for that check.
5. Step 5 — Add a comment to any remaining justified overlap.Even after refactoring, you may have an unusual SS or FF relationship that a reviewer will question. Add a relationship note in P6 explaining why it's modeled that way. "SS+10: steel detailing starts after engineering is 80% complete per RFI-243 protocol." Documentation closes the audit loop before it opens.

What I Look For in My Own Schedules

When I'm reviewing a schedule — mine or a subcontractor's — I always run Check #2 first. It's the canary in the coal mine. A schedule with zero leads might still have problems, but a schedule with leads almost always has bigger problems hiding behind them. Wherever someone put a negative lag, somebody under pressure was unwilling to confront the math. That's a tell.

The schedulers who hand me clean Check #2 reports tend to also hand me clean Check #7, #10, and #1 reports. The discipline propagates. The schedulers who tell me "we'll fix the leads later" never fix the leads later. They just hand the problem to the next person who opens the file.

Make It a Pre-Submit Check

Run DCMA #2 on every schedule before it leaves your desk: bid schedules, baselines, monthly updates, claim narratives. Zero leads is achievable on any well-built schedule. It's not a stretch goal. It's the floor.

ArgusForge runs all 14 DCMA checks plus the Schedule Quality dimensions in about two minutes against a P6 .xer export. If you're already routinely opening Claim Digger or building manual lag filters, this turns it into a single import. However you check it, check it. Leads are the easiest finding to remove before submittal and the hardest to talk your way around once an owner has flagged them.

The schedules that hold up under scrutiny don't have leads. The schedules that don't… eventually get scrutinized.