THREE QUESTIONS EVERY BUYER ASKS A CONTRACT MANUFACTURER. ANSWERED FROM YOUR OWN RECORDS.
A buyer evaluating a contract development and manufacturing organization gets a schedule with an open slot, a flowchart of the deviation procedure, and a list of qualified people. Assure replaces the slides with the answers the records give: the capacity that is real, how a deviation actually moves through the building, and what would still run if the most senior engineers were gone tomorrow. It is optional. An enterprise that does not make or buy contract manufacturing leaves it off and nothing else changes.
1. What is your real capacity margin, not your quoted capacity?
Anyone can show an open slot on a schedule. The question that matters is what happens to that slot when the largest customer on the same equipment and operators has a demand spike, and the program has no leverage in that conversation.
| Assure shows | How |
|---|---|
| Quoted against available | The hours quoted for each resource, and the mean of the hours that were actually available each week over the history you recorded. Derived. |
| Margin that is left | Available hours less the hours committed to every customer, as a share of the available hours. |
| The spike | The largest customer on each resource, and how far its own recorded highest week rose above its average week. Where there is too little history, a named assumption is used and labeled as one. |
| What the program loses | The hours a program is displaced in two cases: under the allocation you recorded (reserved blocks first, then priority), and with the largest customer served first, which is the case a buyer without leverage has to price. Both are scenarios. |
2. How does a deviation actually move through your building?
Not the flowchart: the real turnaround and the real number of signatures. In steady state this is occasional. At scale-up it is weekly, and a change-control process built for the occasional case becomes the bottleneck for the frequent one.
| Assure shows | How |
|---|---|
| Turnaround and signatures | Median and nine-in-ten days to close, and the signatures collected, from the closed records. Not stated until there are enough closed records, and set against the turnaround your procedure states if you record it. |
| Where it waits | The wait at each approving role, the role that holds the largest share of all the waiting, and how many signatures are asked one after another. |
| Open records | How many are open, how old, and which role each is waiting on. |
| At scale-up | At a volume multiple you choose, a role at or above full load is shown as saturated. A scenario on named assumptions. |
3. If you pulled your three most senior process engineers out tomorrow, what would still run correctly?
The answer says whether the knowledge that keeps the process in specification lives in a document or in a person, and whether that is a risk the buyer buys along with the capacity.
| State after the removal | Meaning |
|---|---|
| Runs | Two or more qualified people remain. |
| Thin | One qualified person remains. |
| Documented, untested | A current, verified procedure exists, but nobody qualified would remain. It shows where the knowledge is written down, not that anyone has run it. |
| At risk | Nobody qualified remains and the procedure is partial, missing or out of date. The knowledge lives in the people who left. |
The number removed, and how many months a procedure counts as current, are named assumptions you can change. The concentration of qualifications in three people is shown beside the result.
What it proposes, and who decides
Six kinds of proposal, each from the records: put the program's capacity in writing, quote capacity on the hours that were available, route the signatures of a deviation at the same time, add approver capacity before scale-up, cross-qualify the steps that hang on one person, and write down and verify the procedures that live only in people. Each carries its insight with its evidence, the expected outcome, the impacts, and alternatives that each carry the same, including doing nothing.
The outcomes are measures of the thing itself (hours a program is displaced, days of turnaround, steps with two or more qualified people), not dollars. A named owner, admin or controller decides, with a reason. An agent never does. The outcome is reported later by a person, with its evidence, and compared with the target.
What it needs, and what it does not have
Eight kinds of record: resources, loads, weekly history, deviations, approving roles, people, steps and laboratory sample results. They arrive by spreadsheet template in the platform, by a warehouse query, or by a signed push from your systems. Two systems can be read directly: a quality management system with a REST list endpoint feeds deviations, and a laboratory system with one feeds sample results, each through a field map you supply, conformed against a scripted vendor and not yet run on a customer tenant. There is no native connector to a manufacturing execution or scheduling system; where those hold the data, a person or a scheduled export carries it in. A record that fails its checks is named with the reason and the rest are recorded.
See it. Open the platform and choose Assure to read a recorded run on the synthetic manufacturer, or sign in to run it on your own records.
Open Assure in the platform · Optional modules · Optimize · Fleet