Overhaul 18: W=X=Y=Z ... They're Just Numbers
Introduction
This post is based on my email to a friend about Navy ship maintenance problems. The email was almost the target length of a blog post so I am reproducing it here with slight modifications and more explanation.
Three warnings. First, the email was a little snarky, just below the level of disrespect, and I’m not bothering to reduce that. People in senior Navy leadership positions don’t care what I think anyway. Second, none of the snark should be interpreted as a lack of respect for the real heroes of Navy ship maintenance: the planners, engineers, and mechanics doing real work. They do great work despite the nonsense that goes on above and around them. Finally, the post contains “inside baseball” for maintenance planning that normal humans don’t care about. You have to be a little strange to like maintenance planning. Normal people probably won’t care any more than they already do after reading this post.
This post contains my reactions to a 2015 technical report of a Navy team convened by NAVSEA to analyze gaps (“chasms” might be more appropriate) in Naval shipyard performance.
The Readiness Kill Chain?
The term Kill Chain was tr’ifcally popular among fleet staffs in the first half of the 2010s. The basic idea came from putting ordnance on target to “kill” it (severe mission failure was probably an acceptable alternative). If you conceive of all the physics and material processes that have to go right from launching the ordnance to severely damaging the target as a chain, the links can be identified, analyzed and optimized to improve the odds of a successful outcome and bad day for the target.
Someone had the bright idea to apply this type of analysis to ship readiness and the Ship Readiness Kill Chain (RKC) was born. We’re going to kill readiness? Never mind. There were training, manning, supply, and material RKCs. It sounds a lot like the PESTO pillars: Personnel, Equipment, Supply, Training, and Ordnance. There are many ways to slice Navy ship readiness, but it always comes down to these.
Someone had the even brighter idea to use the RKC approach to perform a Root Cause Analysis (“there you go again”) of ship maintenance availabilities from budgeting (“programming”) through the doing the work (“execution”). I see two problems with this idea. First, ship maintenance isn’t the same as launching ordnance. If there were as much dependence on guessing “notional values” for launching a projectile as there is for ship maintenance, the safest place to be before launch would be in front of the launcher (this was the case in the age of smoothbore weapons). The only thing that gets launched in ship maintenance is the occasional Project Sup.
Second, the links in the ordnance kill chain don’t interact out of sequence as they do in the end-to-end process of ship maintenance. For example, someone launching something from another platform won’t impact your ordnance.
In the maintenance “kill chain,” your plan can be impacted by other ships. You can have a great plan, fully funded (I can dream), with priority for resources (another dream), and start on time. Great. Then an aircraft carrier collides with something (it happens). It will need urgent repairs and will get priority over your project and resources. Your plan and Resource Leveled Schedule become trash. Even if someone at the fleet or OPNAV level with stars cared (they don’t) or listened when you told them about the impact of the emergency work (they won’t), your ordnance just self-destructed before getting near the target. The kill chain metaphor is inapt for ship maintenance, but don’t let that stop you from convening a team and publishing a report.
What’s Wrong with Ship Maintenance?
Like other complex problems, what you think is wrong with ship maintenance depends on where you sit and whom you ask. I’ve covered the plethora of activities involved in the end-to-end process in other posts. All of the people in them have opinions about what the other guys are doing wrong.
As you know from this series on ship overhaul, the Government Accountability Office (GAO) has issued a series of reports about problems with ship overhaul. The GAO should use its old name, the General Accounting Office, since it can’t hold anyone accountable for anything.
The GAO reports referred to the 2015 NAVSEA RKC causal analysis (wait for it) … zero times. GAO insight came from analyzing letters SYs wrote explaining delays (which have a requirement NOT to be insightful) and asking people at HQ and SY, “What do you think makes ship overhauls late and over budget?” It speaks volumes that the NAVSEA analysis wasn't mentioned in those conversations, even in the Navy comments that have to be included in the GAO reports.
In several places, the NAVSEA report (it isn’t on the web, so I can’t link to it) declares with refreshing honesty, “Our metrics tell us nothing useful, so we put together focus groups and told them, ‘Draw some pictures about barriers.’” The technical report has good diagrams, but diagrams untethered from data are editorial cartoons.
The NAVSEA report was issued in 2015 followed by a decade-plus of unfulfilling ship overhaul performance. Why? The guesses behind the cartoons were wrong? The people responsible for the actions recommended didn’t believe them or couldn’t do anything? So many questions.
W = X = Y = Z It’s Just Math!
The real motivation for this post was the use of W, X, Y, and Z in the NAVSEA report. This is where the real inside baseball starts, but I’ve convinced myself that I can simplify it to make it interesting even if you’re not a ship maintenance geek like myself. Raise your tray tables because we’ll be going above and below 10,000 feet repeatedly.
The report defines the variables thusly (with some additional notes from me):
W is the Navy part of the DoD budget value notional cost of the availability (set years in advance),
X is the budget execution year guidance (what OPNAV authorizes the fleet to spend),
Y is the Final Review Estimate (FRE, what the project team and SY leaders say the work in the Availability Work Package (AWP) will cost, only tangentially related to the notional cost W) and
Z is the actual cost required to do the work
W = X = Y = Z is a great engineering way to conceptualize values that have to exist at points in time for specific reasons, but don’t have to have any connection to reality (actual cost Z is the only one). That’s DoD budgeting for you. Another advantage of the equation is that it has visceral appeal for the engineers who dominate the end-to-end maintenance process after budget shenanigans processes stop and the real work begins.
Engineers love equations. All engineers know that they must balance. That’s our problem, HA! We’re out of balance! We just need to balance those numbers like an oxidation-reduction formula!
Problem 1: OPNAV needs a budget number for the overhaul at least five years in advance before anyone knows what’s in the AWP. People in the maintenance community are not allowed to say, “We have no idea” so they say something that means the same thing, “Here’s the notional value W from the Class Maintenance Plan” (I’ve written about CMPs before). It’s nice, but entirely optional, for the CMP to have a strong factual connection with what work on the ship actually costs. Sometimes, a NAVSEA program office will insist that the SY and planning activity “sharpen their pencils” (code for “give us the number we want) to reduce the notional value of the availability. The people with the newly sharpened pencils aren’t allowed to say, “But we have lots of data and that’s what it actually costs to do the maintenance that is necessary.” That’s called “not being a team player.” Whatever it takes, the program office gets the number they need. They speed read past the unrealistic assumptions necessary to produce it. They have to or risk being accused by OPNAV staff of “not being team players.”
Problem 2: No matter what the budget execution value X is, it is no more grounded in reality than W because the availability hasn’t been planned, and the SY hasn’t visited the ship to learn how much worse its condition is than assumed by the notionals (it’s never better).
Logic Problem 1 (these are different from “problems” because they’re based on bad thinking): why would you expect W=X=Y (Notional budget = Execution Year Budget = FRE)? EVER? They are different quantities based on different things for different purposes.
The only truth in W=X=Y, if it can be called that, is Y because it is based on known work and the resources for accomplishing it. Its validity depends on SY leaders not lying to themselves about committing to specific job man-day estimates. Technically, it’s not lying to others if you convince yourself it’s true.
Of course, the validity of the FRE is buried in SY performance factors (history for recent, similar work) and the tug-of-war between SY actual performance and “corporate notionals.” There is always a struggle over SY performance factors (what they estimated it would cost them to do specific jobs divided by what it really costs them). Performance factors are usually less than one and sometimes much less for new and really hard work.
Sometimes the struggle over performance factors is public, with bill-payers, and sometimes it is just within NAVSEA. If customers knew that the last three times with work was done, it cost x amount more than the notionals, they might get angry. Yup, but they pay the bill no matter what.
Logic Problem 2: Y = Z (FRE and actual cost to do real work). If those are ever equal it is a miracle because Z depends on skills, actual shipboard conditions, new work, resource numbers, other conflicting work, and SY support (oh, and work stoppages due to bad behavior). There are many more logic problems, but I’m struggling to keep the post under 2000 words.
Conclusion
The way the Navy budgets and plans for work on ships is orthogonal to the physical work (hardware, conditions, resources, people). Trying harder won’t help. No one can say that aloud, of course.
A different, “let’s accept reality” approach to the Navy depot maintenance end-to-end process might be:
Let OPNAV staff have notionals W for the budget process, but update them as often as allowed based on what we learn when we do maintenance. This may involve beatings from staff members of Congressional committees, but they keep siccing the GAO on the Navy, so perhaps they would be interested in the facts. Maybe this has been tried and dead bodies resulted. I don’t know.
Raise X, the budgeted value for the overhaul in the year of execution, as necessary for the long periods when W is out of date. This will likely lead to force structure cuts because the Navy can’t afford to do the maintenance on a ship it has when it is lying about hiding the actual costs. The Navy already does this when it parks ships at the pier and doesn’t maintain or modernize them.
Make Y equal to the cost of doing the actual work at the SY with actual performance factors. Saying anything else is lying a dream.
When Y doesn’t match X (brace yourself, it won’t), remove work from the AWP until it does. It might be sacrilege, but it is the usual practice (after lots of yelling).
Transfer what you can’t afford to future maintenance opportunities. Of course, this builds a “maintenance backlog,” but the Navy already does this when it defers availabilities because W, X, and Y are always out of balance. This would at least be more transparent than shrugging shoulders when called to testify about decommissioning ships early. It also wouldn’t waste the time of focus group members who could be doing real work instead of writing reports with recommendations that senior leaders don’t implement.
