Inheriting Struggling Projects

In 2019, the country of Turkey took delivery of S-400 air-defense systems from Russia. The US government then took long-threatened action and forced Turkey off the F-35 fighter jet program. The six F-35s that Turkey had received were impounded in the US (there for training) and Turkey was banned from manufacturing F-35 components.  Enter me.

Turkey was an "International Partner" on the F-35:  they had paid for part of the development costs in exchange for industrial participation

Geopolitical background

In the late 2010s, Turkey was drifting away from the United States and aligning more towards Russia. In response to a coup attempt in Turkey in 2006, Prime Minister Recep Tayyip Erdoğan took authoritarian action to repress challenges to his rule, which hurt his relationship with the US. In addition, the conflicts in the Syrian Civil War led to the US supporting Kurdish militias in Syria; Erdoğan feared a Kurdish national movement in Turkey and Russia feared the fall of Bashar al-Assad, so Russian and Turkish interests were aligned on at least one point.

Turkey chose to purchase S-400 air-defense systems from Russia. At the time, the S-400 was considered one of the best in the world, although their battle-testing in Ukraine has been mixed. This purchase infuriated the US, which feared that the Russian-manufactured radars would collect and share extensive data on the US (and its allies') premiere low-observable "stealth" jet and would make Russian and Russian-allied air defenses more effective against it. The US made Turkey an ultimatum: Turkey could have S-400s or F-35s, but not both. Turkey officially took delivery of the S-400 in 2019 and was kicked off the F-35 program. F-35 manufacturers were told to purchase parts from non-Turkish sources, and the initial commit to have this done was March 2020 (hah!).

Initial project

I was not present for the initial phases of the project, but because of the geopolitical risk (no Turkey allowed by government fiat) and urgency (project started in 2019 for completion in 2020) the work transfer started with an ad-hoc team. The company's work-transfer procedures were less mature,  dedicated work-transfer teams has only begun to stand up, and knowledge throughout the organization was patchy. Roughly 200 part numbers had to move to non-Turkish sources, passing through phase gates like project approval, first-article approval (can the new supplier make the part?), and production approval (can the new supplier consistently make the part and we can safely shut off the incumbent?). The ad-hoc team had identified new suppliers, gotten project approvals, and parts were on order with new suppliers. In some cases, first-article approval was complete. And then the project head (CE) left for military leave, most of the rest of the team disbanded, and the remaining one employee (KL, part-assigned to this project) and one contractor (GD) were struggling.

When my manager told me to take on the project, I began to look for the various planning documents. When I arrived, the following was true:

1. The project scope workbook was split into three files (corresponding with "A", "B", and "C" priorities from early on, no longer relevant) which included some parts duplicated across the workbooks.
2. The team had no systematic way to pull and verify that first articles were completed and approved
3. The quality plan was split across three workbooks, had no revision control, obscure notes, and didn't match any existing quality plan format
4. Because of the above, the team didn't have a part-by-part, action-by-action burndown plan to complete the project
5. We had a large inventory of parts purchased from Turkey before the March 2020 deadline, but no tracking on how close we were to running out of those parts to prioritize the various transfers
6. We were already past the initial "March 2020" plan (which was extremely optimistic for landing gear parts, which can have lead times greater than one year)

Side note: how do I know the project started no earlier than 2019? I still remember the project numbers with their "19" prefix. Scar tissue!

Attempt #1: manage the project through the people

Management was looking for better status updates and burndown charts, so I set up daily conversations with the contractor GD to coach her through what I needed. This included putting together the approval presentations and workbooks to close out first-article and production approvals in batches as parts became ready. I developed a template for the first set and then worked with her on part numbers ready for the next push, including collecting signatures on the various pre-approval documents.

While working with GD, I was also trying to wrap my head around the project with the help of the employee KL. KL was also trying to follow the clues left by the project manager CE, who was unavailable during his military service. As is my wont, I started trying to understand the project by making spreadsheets. Unlike my later VBA-enabled monstrosities, these were

• A combined project scope workbook including all three (later four) project numbers in one place, including their approval status
• A combined quality plan workbook across all of the project numbers
• Revision control by saving a "main" copy of the two files in the top-level project folder (as working copies) and saving "as-presented" and "as-approved" (if different) copies in individual folders for each of the various approval presentations

I also built an "actions burndown" tracker for all parts, but that quickly became "one more thing to update" and so I retired it in favor of adding some additional tracking and automated summaries within the working files.

After a few weeks or months, GD had helped assemble some presentations and track approvals, but we had run out of budget and time to keep KL and GD working on what had become "my" project, so GD's contract was reassigned and KL was available only to answer questions...until she left the company shortly thereafter. Scarily enough, I was now the person with the most knowledge of the project in the entire company.

As part of my initial learning, I found that some decisions from the previous year were outdated: selected suppliers were now banned for future orders or strategies had changed. For those I drafted and presented "delta project approval" reviews to ensure that we had the paperwork closed out that matched the as-planned project to the end result.

The big push

In the months that followed, I worked with the various people in commodity management, procurement, supplier quality, etc. and the suppliers themselves to understand the roadblocks to first-article delivery and getting through them. Sometimes this was just reminding suppliers that we needed the parts, and working with procurement to not push out delivery dates (remember we had a bunch of Turkish parts in-stock?) I developed a method of pulling all first article inspection reports in our system, parsing out the ones I needed, and then using that to validate that we had an actually approved part (rather than just the supplier saying "finished!"). I had to chase quality tags to get parts moving again after manufacturing issues, and do all of the other not-fun-but-important expediting things.

Once the parts were in the door, many required additional quality checks. Because the original plan was for these parts to be brought in and then sit on a shelf for a while (because we had bought out Turkey's inventory to be consumed first), we had to avoid the risk that when we finally pulled the parts out of the box they wouldn't be compliant. So the quality plan specified many parts for a "fit check" (test-install the part on its next-level assembly) and side-by-side inspection or "tin soldier" inspection (put a Turkish part and a new part next to each other, "lined up like tin soldiers": if they look different, there may be a problem).

Fit-checks required pulling a new part from controlled inventory (while it's still not approved for use on the assembly line!), carry it to the assembly line, and catching the interfacing part at the right time in assembly to test-fit it. Coordinating the timing here was key so that a next-level assembly was ready and the new part could be returned to controlled inventory as quickly as possible.

Side-by-side inspections required pulling two parts simultaneously (hopefully there was still a "known" part in inventory!) and avoiding any potential for mix-ups where the new part went back into production inventory. Because these inspections were unique to the F-35 transfer, we had to explain ourselves constantly on what was needed. Some suppliers had delivered parts months before, but because of the shambolic management up to that point, we had to enlist forklift operators to pull hundred-pound crates from top shelves ("DO NOT USE - CONTROLLED INVENTORY FOR [no-longer-relevant person]") so that we could count the parts inside and pull some for inspections.

Once the first article inspection report was approved, inspections passed, and PPAP interim approvals collected, I had to present those parts for work-transfer approval, after which we could finally release them to production. When I was lucky, the new supplier had already made a statistically-valid quantity of parts and I could get first-article approval and production approval in one request. Often, I could only get first-article approval, which at least meant I could release the parts from controlled inventory.

Approvals trickled in and I was able to mark the majority of part numbers closed. Eventually, I could even mark some entire project numbers closed. But then there were the troublesome parts.

The stragglers

Eventually I had all of the parts through first-article approvals, but many still lacked production approvals. This might be because the parts were re-re-sourced (i.e. sourced to a third supplier after the first batch of parts came in), a lack of demand meant that the required batch size hadn't been hit, or maybe the supplier just couldn't make good parts consistently yet. Attention from management had waned: first articles were all complete and no one was complaining about the production line being short, so they went to put out other fires. I still cared because I refused to mark the project "closed" on the divisional work-transfer tracker until the project got final production-level approval and any post-approval actions were closed

Re-re-sourced parts were easy enough: during the production-approval review, I'd present the review as simultaneous approval of scope-reduction for those parts to remove the statistical requirement. Some parts didn't have that requirement anyway, and we didn't need production approval for suppliers who weren't going to be making parts.

The last half-dozen or so parts were ones where we had the ability to hit production approvals except for lack of demand. Almost all of those closed out after weekly calls with Supplier Quality and the supplier to keep things at the forefront of people's minds when demand trickled in and then push through any obstacles. Using my follow-up tools was essential here, because at this point the parts took up maybe 45 minutes of brain-space per week so I needed to rely heavily on my notes. Slowly we worked our way through the various PPAP approvals and revisions until we could get everything approved.

Conclusion

After 4-5 years of pain, I finally closed out the entire project. We had not affected deliveries for the F-35 despite uprooting a substantial fraction of the bill of material. I spent some extra time organizing the evidence of approval so that a future person had a chance of figuring out that the project was approved and marked it closed in the divisional tracker. Like many projects that I've taken over midstream, this one reminded me of the value of well-organized project documentation rather than trying to figure it all out later; you might not be in that job anymore, and you don't want to get the call to help once you've moved on to bigger and better things.

Probably the hardest part of this project was inconsistent management attention: after the first burst when I was assigned, it became long silences punctuated by my manager (or his manager) remembering this existed and asking for detailed updates. And why wasn't it done yet already? I could have closed it a full year or two earlier than I did and no one would have noticed until a paperwork discrepancy showed up in an audit, but I kept at the project until I could close it in good faith.