Failed DCMA? Fix These Three First

Your schedule came back red on eight of the fourteen checks. The owner wants it fixed by Friday.

Don't fix eight things. Fix three.

Most of the DCMA 14-point failures aren't separate problems. They're the same two or three mistakes showing up in different columns. Fix the causes and the rest clear on their own.

Here's the order.

1. Open ends

   Any activity missing a successor. That's the first one, always.

   An activity with nothing after it has nowhere to push. P6 gives it enormous float and takes it off the critical path.

   One cause	An activity with no successor
   Three red checks	High float · critical path test · a CPLI number that means nothing

   Filter for activities with no successor. Give each one a real successor, or write down why it doesn't need one. That's the whole fix.

   This pass alone usually clears half your failures.

2. Hard constraints

   Mandatory Start, Mandatory Finish, Must Finish On.

   These override your logic. The schedule stops calculating and starts obeying a date somebody typed in two years ago. Slip happens and the schedule doesn't show it, because the constraint is holding the date in place.

   Then it shows up as negative float, which everyone treats as a scheduling problem. It isn't. It's a constraint problem.

   Pull every hard constraint. Keep the ones the contract actually requires. Convert the rest to soft ones — Start On or After — and let the logic do its job.

3. Float, last

   Not before

   This is the part most people get backwards. They open the schedule, sort by total float, and start "fixing" the float numbers.

   You can't fix float directly. Float is output, not input. It's what the logic produces.

   So after steps one and two, re-run and look again:

   Still high float on a big chunk?	You have more open logic than you found. Go back to step one.
   Seeing negative float?	Now it's real. Now it means something. Now you can talk about recovery.

Then re-run the check

Run the DCMA check again and see where you actually stand. In most schedules I've reviewed, two passes on logic and constraints takes a failing schedule to a passing quality score without touching a single duration.

That matters, because changing durations to make numbers look better is how you lose credibility with an owner. Fixing logic is defensible. Shortening activities to hit a metric is not.

One more thing

A schedule that passes all fourteen checks is not automatically a good schedule. It means the network is built correctly. It doesn't mean the plan is right.

Those are different problems. Get the first one out of the way so you can argue about the second one.

For the full list, see the quality checklist.