Question and scope

This protocol measures how many live vendor commitments have no record outside an email thread. A live commitment is an obligation the buyer still expects the vendor to perform, such as a delivery, a repair, a document, or a scheduled service. A tracked record is an entry in the vendor tracker, procurement system, or other agreed system of record that a coordinator can find without searching a mailbox. The result is a share, a count, and a list of commitments that need to be moved into the tracker.

The measure matters because email is a weak place to hold an obligation. Messages get buried, threads fork, and the person who received the confirmation may leave the team. A commitment that exists only in email can still be real to the vendor while being invisible to the coordinator who plans around it. Moving it into the system of record makes the obligation visible to everyone who depends on it.

Scope is one coordination team, a defined vendor population, and a stated observation window. Include commitments created through email, chat, phone follow-up, or a vendor portal, provided the team still expects performance. Exclude completed work, cancelled orders, and one-off requests that were declined. Write down the systems, mailboxes, and dates so another analyst can rebuild the exact same population later.

What counts as living only in email

A commitment lives only in email when the evidence of the obligation appears in a message and nowhere in the agreed tracking system. The message may be a vendor confirmation, a buyer's purchase instruction, or a thread where both sides accept a date and a scope. The test is whether a coordinator who searched only the tracker would know the commitment exists. If the answer is no, the unit belongs in the email-only group.

A message that merely mentions a commitment is not enough. A forwarded note saying the vendor will handle a repair is a lead, not proof that the vendor accepted the work. Look for an explicit exchange: the buyer asks for a specific deliverable by a date, and the vendor agrees or proposes a date the buyer accepts. A message that only records an internal plan stays internal until the vendor responds.

Unit of analysis and exclusions

The unit is one vendor commitment with one deliverable, one vendor, and one expected completion date. Split a thread that covers several deliverables when each has its own date and acceptance. Keep one unit when a single exchange covers a bundle that ships or completes together. Count the commitment once even when several messages discuss it, because the question is about the obligation rather than the volume of correspondence.

Exclude quotes that were never accepted, internal requests that never reached a vendor, and completed work with no future obligation. Exclude duplicate units created when two people in the same thread describe the same promise. Record each exclusion with a count so the denominator is defensible. Keep a short note of what was removed and why, especially when a thread is ambiguous.

Building the population

Start with the tracker and the open purchase orders that still carry a vendor obligation. Then search the shared mailbox for vendor confirmations, revised dates, and service instructions that may never have reached the tracker. Include the personal mailboxes of coordinators only when the team agrees that those messages are part of the vendor record. Add commitments discussed in the recurring vendor review but missing from both sources.

Give every unit a stable identifier before coding so the same promise can be matched across the tracker and the mailbox. Record the vendor name, the requester, the date the message was sent, and the owner of the mailbox. Mark units found only in email, because those are the ones the measure is trying to count. Freeze the list before deciding which units are tracked, so the coding cannot drift toward an expected answer.

Separating email from other untracked channels

Email is not the only place a commitment can hide. A promise may live in a chat thread, a phone note, a portal message, or a meeting summary. Code the channel where the strongest evidence sits, and report chat and portal commitments separately. Combining every untracked channel into one email category would overstate the email problem and hide where the next fix belongs.

If a commitment appears in email and in chat but not in the tracker, count it once and name both channels. The useful question is which channel the coordinator would have to search to find it. That answer points to the intake change: a mailbox rule, a chat export, or a portal check. Report the channel mix beside the headline share so the team can act on the largest gap first.

Evidence standards for coding

Accept a unit as email-only when the message is attributable to the vendor and specific enough to act on. A vendor's own message naming a deliverable and a date is strong evidence. An internal note that summarizes a call is weaker, and it should be coded as unconfirmed until someone writes to the vendor. Record the evidence type so the analysis can distinguish a documented promise from a remembered one.

Do not upgrade a unit because a coordinator recalls the conversation. Ask the coordinator to send a short confirming message and attach the reply, then recode the unit. If the vendor will not confirm in writing, that refusal is itself worth reporting. The goal is not to punish informal communication but to make the obligation visible to the people who plan around it.

Sampling and verification

Use a census when the team can review every live commitment in the window. That is common for a small vendor population and gives the clearest result. Above roughly one hundred active commitments, draw a random sample and publish the method. Stratify by vendor and by receiving team so one large supplier does not dominate the share, and state the confidence range beside the result.

Have a second reviewer recode at least ten units and compare the results. Resolve disagreements by returning to the original message and the tracker record. Record the resolution so a later reviewer can see how the definition was applied. If disagreement is high, rewrite the coding rules and recode the sample before reporting. A stable definition matters more than a large sample.

Decision rules before coding

Decide before coding whether a commitment counts as tracked when it appears in a personal spreadsheet that the team does not treat as authoritative. The answer should follow the team's actual practice, not an ideal process. If the spreadsheet is the only place a coordinator would look, it may function as a system of record even when it was never designed as one. State the choice and apply it consistently.

Decide how to treat a commitment that was moved into the tracker after the observation date. Code it as untracked at the cut-off, because the measure describes the state at that moment. Note the later move in a separate field rather than backdating the record. The same rule applies to a commitment that was closed during the review. Freeze the state before reading the tracker so the result reflects one point in time.

Analysis and reporting

Report the share of live commitments that live only in email, the share that live in another untracked channel, and the share that are tracked. Show raw counts beside every percentage, especially when the population is small. Add the count of email-only commitments with a date in the next thirty days, because those carry the most immediate risk. A single headline number hides which commitments need attention first.

Break the result down by vendor, receiving team, and commitment age. Suppress groups with fewer than five units so one unusual promise does not look like a pattern. Pair the share with the time each commitment has been open, because an untracked promise near its deadline is more urgent than one with months to go. The vendor commitment date protocol covers whether a date is confirmed; this measure covers whether the commitment is visible at all.

Limits and sources of error

The sources cited here describe internal control and supply chain practice in general terms; they do not provide a benchmark for email-only commitments. A commitment may be missing from the tracker because the coordinator did not save the message, or because the tracker was never designed to hold that kind of work. A low tracked share is a prompt to inspect intake, not proof that the vendor relationship is failing. Report the measure with that distinction so the reader does not mistake a record gap for a performance problem.

Email-only commitments are also undercounted when coordinators delete messages or use personal accounts the review cannot access. Check a sample of recent threads and meeting notes to estimate the size of that gap, and report it as a limitation. If a large share of vendor work never reaches the tracker, fix the intake path before setting a target. A precise share built on an incomplete list is less useful than a range with a clear caveat.

Review cadence

Review the result at a fixed point and end with one change, one owner, and one check date. A practical first change is to require a tracker entry within one business day of any vendor commitment made by email. Repeat the measurement after two full cycles before setting a target, because a single period may reflect a seasonal peak or a supplier change. Keep the definitions and the sample method with the result so the next reviewer can compare like periods.

Retire the measure when every live commitment is visible in the agreed system and the email-only share stops changing. If the team keeps it, publish the email and other-channel shares together, because a narrow email headline can hide a larger chat or portal problem. Use the cross-team commitment log to give each moved commitment an owner and a next check. The purpose is an obligation the whole team can see, not a clean inbox.

Sources