The anatomy of schedule quality: what the DCMA check reveals across a portfolio

Run the DCMA 14-point check across enough programmes and a striking thing emerges: bad schedules don't fail at random. They fail in a recognisable shape — the same handful of checks, in the same order — and those structural failures quietly predict which projects will slip months later. This is a model of that shape, and what the best schedules look like instead.

The uncomfortable baseline: most schedules fail

Start with the headline every reviewer learns early and every contractor hates to hear: the typical first-submission CPM schedule fails several of the 14 checks. Not because planners are careless, but because the checks measure something most schedules were never built to satisfy — structural integrity strong enough that the CPM engine can calculate honest dates. A schedule built to look right on a wall chart and a schedule built to survive forensic scrutiny are different artefacts, and the gap between them is what the 14 points measure.

The distribution of quality scores across a mixed portfolio is not a bell curve centred on "good". It's a lump sitting low, with a long thin tail reaching up toward excellence that very few schedules occupy.

The failure fingerprint

Here is the part that genuinely surprises people the first time they see it laid out. The fourteen checks do not fail with equal frequency. A small number fail constantly; others almost never do. And the order is remarkably stable from one programme to the next, across sectors. This is the "fingerprint" of schedule failure — and because it's consistent, it tells you where to look first on any file you're handed.

Three things fall out of this fingerprint. First, the top of the list is all structure — missing logic, lags, durations, relationship types, constraints. These are the checks that determine whether the network can compute honest dates at all. Second, the zero-tolerance hygiene checks (leads, invalid dates) are near the bottom — easy to keep clean once you know they exist. Third, and most usefully: if you only have ten minutes with an unfamiliar schedule, you now know the order to look in. Missing predecessors and successors first, every time.

The insight that should change how you review: structure predicts slippage

Now the part that elevates this from a checklist to a forecast. It is tempting to treat the 14 points as a tidiness audit — a way to make a schedule "neat". But the structural checks are leading indicators of delivery performance. Plot a portfolio with structural quality on one axis and actual on-time delivery on the other, and the schedules don't scatter randomly. They fall into two families.

Why would the tidiness of a network predict whether the project finishes on time? Because the structural failures aren't cosmetic — each one is a small lie the schedule tells about how the work connects. An activity with no successor can slip without moving the finish date, so the slip goes unmanaged. A lag hides a real duration nobody owns or updates. A hard constraint forces a date the logic can't support, masking negative float. Individually small; collectively, they blind the team to the very slippage the schedule exists to surface. A structurally weak schedule isn't just untidy — it's an early-warning system with the wires cut. The project doesn't slip because the schedule is messy; it slips because nobody could see the slip coming.

Quality is a trend, not a snapshot

The single most diagnostic thing you can do with the score is not measure it once — it's watch it across revisions. Healthy programmes hold their quality roughly steady update to update; the planner maintains the logic as the work evolves. Troubled programmes show a quietly declining score: open ends creep in, constraints get added to force dates, durations balloon as detail is abandoned. The score decays months before the completion date moves — which is precisely why the trend is worth more than any single reading.

What good actually looks like

It's easy to dwell on failure, so here is the other end. A top-decile schedule — the thin tail of Fig 1 — has a recognisable profile, and none of it is exotic. Logic is complete (open ends only on the genuine start and finish). Lags are rare and named; leads are absent. Over 90% of relationships are finish-to-start. Hard constraints are counted on one hand and each has a documented reason. Activity durations sit comfortably under the two-month line, so progress can be measured honestly. The critical path responds when you test it. And — the tell that separates the genuinely good from the merely tidy — the score holds steady across updates, because someone is maintaining it.

None of that requires talent, only discipline and the habit of checking. Which is the quietly hopeful conclusion buried in all this: schedule quality is almost entirely a maintained property, not an innate one. The portfolios that score well aren't staffed by better planners — they're staffed by planners who run the check every period and fix what it flags.

Methodology — so you can argue with it

In the spirit of the schedules this site exists to scrutinise, here is how the model was built, openly. The failure fingerprint (Fig 2) follows the long-reported pattern in the DCMA 14-point assessment and the GAO Schedule Assessment Guide, in which logic, lag and duration checks fail far more often than the zero-tolerance hygiene checks — the ordering shown is representative, not measured from a named sample. The distribution (Fig 1) reflects the widely-observed reality that first-submission schedules cluster below the threshold most reviewers trust. The structure-vs-delivery relationship (Fig 3) is directional: the GAO guide and NDIA's Planning & Scheduling Excellence Guide both treat structural soundness as a precondition for a schedule being a usable forecast, and the diagonal expresses correlation, not a fitted regression. The decay curve (Fig 4) illustrates the well-known lead/lag between quality erosion and date movement. Every figure is labelled indicative for exactly this reason: the shapes are robust and worth acting on; the specific percentages are illustrative. The most honest thing about schedule quality is that you can measure your own in seconds and stop relying on anyone's model — including this one.