In short: A transition stalls at the first physical task, which is getting a tool out of somebody else's building, and that task has a legal question underneath it that ownership on paper does not settle. Build the qualification calendar backwards from the date the old agreement ends, and include one tool correction loop, because a plan without one is a plan you will miss. The bridge buy is the qualification calendar plus contingency minus the cover you already hold, bought from the supplier you are leaving, which is why the sequence puts the purchase before the notice. The drawing pack will not make good parts on its own: the control plan, the process window and eight years of approved deviations are what actually travel.
The sourcing decision was signed in March. It is now July, the incumbent is still shipping every unit, the new supplier has quoted and holds no tooling, and the saving in the business case has earned nothing for four months. No one cancelled the project. It stalled on the first task in the sequence, which is getting a tool out of somebody else's building, because that task has an owner in procurement, a dependency in engineering and a legal question nobody wants to be the one to ask.
Transitions fail on sequencing far more often than on supplier selection. The selection work is visible, it produces a decision paper, and it ends. What follows is a chain of dependencies where a two week slip at the front pushes the crossover into your peak season.
Tooling title is a document, possession is a fact
The supply agreement says title in the tooling vests in the buyer on payment. That clause is necessary and it does not get the tool onto a lorry.
What matters when the tool needs to move is whether your interest is perfected and identifiable. In the US that usually means a UCC-1 financing statement filed against the supplier describing the tooling, so that a trustee cannot fold it into the estate. In England and Wales it means a clearly drafted bailment with asset-level identification, and registration at Companies House where a charge arises. This gets done at purchase or it gets done badly under pressure, and if it was never done, it should happen before you announce anything. Take that to counsel rather than to a template, because the mechanics differ by jurisdiction and by whether the supplier is solvent when you ask.
Then verify what exists. The most common discovery at the start of a transition is that the tooling schedule lists 47 tools and the supplier can locate 39. Three were scrapped after a design change and nobody amended the schedule, two were cannibalised for spares, and three are at a subcontractor you have never heard of. Asset tags and dated photographs are the cheap fix, and an annual physical verification is cheaper than the alternative.
Transferability is the part that breaks business cases. A tool carries setup that belongs to the machine it has run on: clamp and bolster pattern, ejector layout, hot runner controller mapping, shot size and a cycle window tuned to one screw and one chiller. Moving a mould from a 400 tonne press to a 500 tonne press means a fresh tryout and a dimensional sign-off, and transfer dies tuned to a specific press line may not move at all. Get a refurbishment quote from a toolmaker before you commit to a transfer date. The figure is large enough to change the answer and it cannot be read off a drawing.
Check who owns it at all. Many contract manufacturers fund tooling and amortise it into the piece price, which makes the tool theirs until the amortisation completes and sometimes after. Find out whether the piece price contained tooling recovery and whether it was ever paid off, because the answer decides whether you are transferring an asset or buying a new one.
The transfer package that is not the drawing pack
A drawing and a specification will get you a part that measures correctly and a process that does not perform. Eight years of production knowledge sits with the incumbent, and most of it is written down somewhere in their quality system.
What has to travel: the control plan, the process FMEA, the process parameter sheets with the approved operating window, the gauge list with calibration records and the fixture designs, the packaging specification, and the identity of the incumbent's own tier two for any special process such as plating, heat treatment, anodising or a welding certification.
The deviation and concession log is the item people forget and the one that causes the most damage. Eight years of approved concessions is eight years of quietly agreed reality. A new supplier held to the nominal drawing will either fail parts that were always slightly out of print, or produce parts to print that your assembly line has been compensating for since the tool was new. Both discoveries happen at first article, which is the worst point in the calendar to find them.
Where the qualification runs through a customer's approved source list, or a special process approval such as Nadcap, the transition calendar belongs to somebody outside your company. Specialty chemicals has the sharpest version of this problem and Q2 covers it.
Building the qualification calendar backwards
Work back from the date the incumbent's agreement ends, rather than forward from today. The forward version always produces a plan that finishes comfortably before a date that has already moved.
A realistic set of stages for a moulded or machined part, with durations that hold up: purchase order and refurbishment quote, two weeks. Tool removal, refurbishment and transport, six. First article tryout and dimensional report, four. One tool correction loop, four. Measurement systems analysis and gauge repeatability, two. Run at rate and the significant production run, two. PPAP submission and review, three. Customer sign-off where it applies, four. Pipeline fill and transit, four. That is 31 weeks including one correction loop.
The middle of that is incompressible for a reason. AIAG's PPAP requires the significant production run to come from production tooling, at production rate, in the production location, with a minimum of 300 consecutive parts unless otherwise agreed, and a Level 3 submission is the usual default. You cannot buy your way past it by paying for overtime, because the point of the requirement is that the process demonstrates capability rather than that the parts get made.
Plan the correction loop explicitly. First articles rarely pass on a transferred tool, because the tool has moved to a different machine with different thermal behaviour. A calendar with no loop in it is a calendar you will miss, and the slip lands at the crossover where you have the least room.
The bridge buy and who you sell it to internally
Bridge weeks equal the qualification calendar plus contingency, minus the weeks of cover you already hold. Multiply by weekly demand and you have the cheque.
Take demand of 4,000 units a week, the 31 week calendar above, and 5 weeks of existing cover including pipeline. That is 26 weeks of bridge, or 104,000 units. At 11.40 a unit the bridge is 1.19 million of stock. It depletes roughly linearly, so average holding is about 13 weeks, and at a 20 percent annual carrying rate the carry costs about 59,000.
Set that against the case. The new supplier quotes 10.50, saving 0.90 on 208,000 units a year, which is 187,000 a year. The carry is worth about four months of the first year's saving, which sounds comfortable until the one-off costs go in: tool refurbishment, engineering and quality time, travel, first article scrap, dual running, and the elevated defect rate while the new line works down its learning curve. Wright described that curve in 1936 and it has been rediscovered on every transferred process since. Those one-offs are what turns a stated two year payback into four, and none of them sit in the procurement budget that owns the saving.
Two features of the bridge buy deserve to be argued out before the cheque is signed.
You are buying it from the supplier you are leaving. Their negotiating position is strongest the moment they understand what a 26 week order is for. Sequence the purchase before the notice, or accept the price and the lead time you get afterwards. A bridge buy large enough to be obvious will be read correctly, so consider staging it across the normal release pattern.
A 26 week bridge is a 26 week design freeze on that part. If engineering has a change queued, it either goes into the bridge quantity or it waits for the new supplier, and both options cost something. The same logic applies to demand: run the downside case from the demand plan and ask what the bridge is worth if programme volume falls 20 percent. If the answer is uncomfortable, shorten the bridge by shortening the calendar, which usually means paying for tool refurbishment to start before the commercial paperwork completes.
Winding down the old agreement
Three clauses decide most of the cost, and they should be read before the transition plan is written.
Notice period and its trigger. Whether notice runs from any date or only from a contract anniversary changes the earliest possible crossover by up to a year.
Raw material and work in progress liability. Most direct material agreements oblige the buyer to take material the supplier committed against a released forecast inside the lead time fence. Stop releasing without unwinding the fence and you owe for weeks of steel, resin or components that will never become your parts.
Exit obligations. Tooling return, transfer of process documentation, a defined support period, and who pays for what. Where the agreement is silent, the negotiation happens at the worst possible moment.
Residual volume commitments deserve their own arithmetic. A take-or-pay or minimum volume clause can make finishing the term cheaper than breaking it, particularly once the bridge carry goes in. Compute both, and include a realistic view of how the incumbent behaves during a notice period, because delivery and quality performance from a supplier who knows they are leaving is a live risk. Raise incoming inspection on the final releases and avoid scheduling the last shipments into your peak.
Do not switch the incumbent off at PPAP approval. Run both sources on a defined split for a defined period, keep the incumbent's tooling available until the new source has produced at rate for a stated number of weeks, and write the ramp split into the transition plan rather than leaving it to whoever is running the shortage call that week.
Where this stops
Some transitions should not be attempted, and recognising them early saves more than executing them well.
If the incumbent developed the process and holds the know-how as their own intellectual property, the drawing gets you geometry and nothing else, and you are funding a development programme with a sourcing budget. If the tooling cannot physically move, you are buying new tools and the business case needs rebuilding at that number rather than adjusted at the margin.
If the trigger for the move was the incumbent's financial condition, the transition calendar and their failure calendar are racing each other, and 31 weeks is a long time to be exposed. In that situation the first actions are inventory position and physical recovery of tooling, with the sourcing work running behind them (YY1).
The systematic error is on cost rather than on time. The saving is a per-unit figure anyone can compute and defend. The cost is a set of one-offs spread across engineering, quality, logistics and working capital, sitting in budgets that did not approve the project and will not volunteer for it. Ask for the transition cost as a single number with named owners before the decision paper goes up, and expect the first version to be low.
The sequence also assumes the crossover can be held. Where demand is growing faster than the new supplier can ramp, or the part is on allocation across the market, this is a capacity problem wearing a sourcing problem's clothes, and no amount of bridge buying resolves it.
Pull the incumbent's supply agreement this week and find three things: the notice period and the date it runs from, the raw material liability window, and the exit obligations clause. Then send the tooling schedule to the supplier and ask for a photograph of each tool with its asset tag visible. The gap between the schedule and the photographs is the first real task in the sequence, and it takes about a fortnight to close.