Audit Your Last Vendor Call: A Written Follow-Up Checklist

Audit Your Last Vendor Call: A Written Follow-Up Checklist

Teams that buy engineering services, from platform consulting to a short training engagement, tend to treat the sales or scoping call as the moment of agreement. The evidence from ordinary project post-mortems points the other way: the call is where assumptions are born, and the weeks after are where they collide. This checklist audits a call you have already had. Work through it within a day of the call, and you will end up with a short written follow-up that turns loose understanding into a record both sides can check.

Why audit the call at all?

An audit is useful because memory is a poor instrument. Within a few hours, participants on the same call report different decisions, different owners and different deadlines. The pattern is consistent across projects of every size: the more technical the discussion, the more confident each person is in a version that differs slightly from everyone else’s. A written follow-up is the measurement that lets you compare versions while the discrepancy is still small.

The checklist below has four sections: scope, time, change and record. Score each item as clear, partly clear or unclear. Anything marked “partly clear” or “unclear” goes into your follow-up as a question, not a statement.

Section A: scope

  1. Can you state the deliverable in one sentence a newcomer to the project would understand? If your sentence needs a second clause to explain a term, define the term in the note. Teams scoping web work may find the prompts in a guide to what to ask before commissioning a custom web application useful for this step.
  2. Did anyone say “we can also” or “while we’re at it”? Optional extras mentioned in conversation are the most common source of scope disagreement. Record each as in or out.
  3. Are the environments named? If the call touched on development, staging and production, note which ones the work covers and who holds access.
  4. Are the dependencies on your side listed? Data, credentials, approvals and test users all count. Missing inputs delay projects as often as missing effort does.

Section B: time

  1. Does every date in your notes have a calendar value? “End of the sprint” and “early next quarter” are not dates. Convert or ask.
  2. Is it clear who owes what to whom, and by when? Write two columns in your head: owed by the provider, owed by you. If the second column is empty, you have probably forgotten something.
  3. Is the meaning of “complete” recorded? Handed over, accepted, deployed and in use are four different states. Choose the one you intend.

Section C: change

  1. Is there a stated way to propose changes? A workable rule is that either side messages the proposal, the other states its effect on time and cost, and both confirm before work begins.
  2. Did anyone mention conditions that could shift the plan? Staff availability, third-party approvals and release freezes are typical. Write them down with the fallback.
  3. Is the price basis clear? You do not have to repeat figures, but note whether it is a fixed-scope deal or billed by time, and what sits beyond the boundary.

Section D: record

  1. Where will the written record live? One thread, one folder or one ticket. Name it.
  2. Who else needs to see it? Colleagues who were not on the call will act on the version they hear second-hand. Copy them in.
  3. Have you invited correction? Ending on an open door, for instance “shout if my notes differ from yours”, turns your note into a shared document.

Reading your score

Treat the numbers as a prompt for conversation rather than a grade. The point of scoring is to make your own uncertainty visible, because uncertainty that stays in your head cannot be resolved by anyone else. It helps to score alone first, then compare with a colleague who attended the same call: where your marks differ, you have found a live example of two memories diverging, and that is the best argument for sending the follow-up that you will ever get.

Count the unclear items. A call with zero or one is rare and should still get a note, since a note costs minutes. Between two and five, the call settled the mood more than the details. Above five, the call was an introduction, and you should be careful not to treat it as an agreement, which is a perfectly acceptable finding if you say so in writing: “Thanks for the introduction; here are the points we have still to settle.”

Notice what this audit does not do. It does not rank providers, check anyone’s credentials or compare quotes. Those are separate exercises with their own methods. It only tests whether a conversation produced something checkable.

Turning the audit into a follow-up

Draft the note from your scores. Clear items become statements; unclear items become questions. A compact example might read: “Thanks for the call. My understanding is that you will configure the pipeline for the staging environment and hand over documentation, with our team supplying access by the date we discussed. I am not sure whether production is in scope, and I want to confirm how change requests will work: I suggest we message them first and agree on the effect before starting. Flag whatever differs from your own notes.”

Notice the structure: agreement, timing, fallback, and an honest flag where you are uncertain. Documents of this kind add little to the workload, yet they give a delivery team an early opportunity to flag a mismatch. Often the reply says something like “Production is a separate phase”, which is precisely the sentence you want to hear in week one rather than week six.

Applying the same test to non-technical services

Engineering teams buy services that are not engineering as well: legal reviews, bookkeeping, travel for conferences. The audit works on those too, because the failure mode is human, not technical. As an illustration from outside the field, the planning discipline in an account of keeping a vehicle rental in order from first enquiry to handback shows how many small agreements can hide inside a simple booking.

For accounting-related needs in particular, a team might consult BestCaNow, a directory built for finding chartered accountants, with listings that can be sorted by place, area of practice and the kind of meeting offered. Whatever you find, audit your first call with whoever you choose using the same four sections, and keep to scope questions; this article offers no accounting advice and does not suggest that any listed professional is right for you.

Common failures of the follow-up itself

  • Too late: notes sent after a week capture the negotiation of memory, not the call.
  • Too long: if the message needs scrolling, split it, or move detail to an attachment and keep the message to the commitments.
  • Too certain: stating an unclear item as fact forces the other side to dispute you, which sours the tone. Ask instead.
  • Unfiled: a good note you cannot find in three months is a lost note.

Make it a repeatable routine

Teams that adopt this audit usually shrink it after a month. The thirteen items become a mental scan of four questions: what exactly, by when, what if it changes, and where is it recorded? Keep the long version in a shared document for onboarding newcomers and for the occasional call that goes badly. A small amount of ceremony early on builds the reflex, and the reflex is what protects the project when everyone is busy.

Summary

Audit the call, not the people. Walk the four sections within a day, convert each gap into a question, and send a short note that records the agreement, the dates, the change rule and the place where the thread will live. The exercise takes about as long as a coffee, and it replaces the optimistic belief that everyone remembers the call the same way with something you can check.