Who Signs Off on RFQ Issue, and What They Look for
The second order is easier to manage than the first, so rfq issue signoff is treated here as a review rather than a pitch. An RFQ sign-off sounds like a single act. A name goes at the bottom of a form and the request goes out. In practice the decision is spread across six to nine named roles, and each of them is signing something different: the variant, the door, the evidence, the commercial structure, the contract form, the delivery position, the credential path, the spare-parts commitment. Only one of those signatures is a formality about the request itself. The rest are the real decisions, and the usual reason a fully signed RFQ still produces a poor programme is that the role who needed to look had no signature on it at all.
Direct answer. A sound RFQ sign-off map has three named parts: the initiator who issues the request, the concurring roles whose scope the request touches — each of whom signs one specific question rather than the request in general — and the record owner who keeps the pack after the responses arrive. The order in which those signatures are taken decides more than the number of them, because a signature taken against an unexamined specification is not a review. What follows runs four sign-off models against five buying situations, scores each round, and lands on the map you would build for your own organisation.
- Read the technical file — check what the first order assumed about rfq issue signoff
- Read the supplier record — check the delivery and support record behind rfq-issue
- Read the contract terms — check the price and lead-time drivers for three questions inside one again
- For rfq issue signoff, rfq issue is what each step has to answer to.
Two boundaries before the arena opens.First, the subject is the internal side of the motion: the request has not been sent yet, and the question is who inside the buying organisation has to look at it before it leaves.Second, the legal dimension of that internal review sits behind a line.
CONFIRM-WITH-LOCAL-LEGAL — which contract form your organisation is willing to sign, which clauses it must not sign without review, and who holds delegated authority to commit the organisation. This is a per-organisation and per-jurisdiction question, and it changes the map before a single RFQ is drafted.
A lock makes this harder than a generic purchase. A smart lock is not one line item. It is a mechanical body that has to fit a specific door, a credential model that has to interoperate with a platform somebody else operates, an electronic product with a support life, and an installed base that a technician will service years after the paperwork is closed. Each of those pulls a different reviewer into the room, and each of them is capable of stopping the request on their own authority.
The commercial structures belong to the same map but not to this page. The price basis, the order-size basis, production timing, the payment structure, the warranty period and the shipping position are all things you ask about in the request and receive back in a written offer. You can request those structures at /request-a-quote, and the sample path is set out at /order-samples.
Three Questions Inside One: Initiator, Concurrence, Record Owner
"Who signs off" is three questions wearing one coat. The first is who puts the request into the system and owns its coherence. The second is who must look before it leaves, and what each of them is looking at. The third is who keeps the pack afterwards, so that six months later somebody can reconstruct why this specification went out in this form to this shortlist.
| Element | The question it answers | Who normally holds it | What breaks when it is missing |
|---|---|---|---|
| Initiator | Who assembles the request and holds it together | Procurement or buyer, sometimes with a project engineer | Requests arrive from two teams with two different scopes and one shortlist nobody can rank |
| Concurrence | Which roles must look before release, and at what | Named role, not named person, with a stated trigger | A request circulates for a month because nobody could say who was holding it |
| Record owner | Who holds the issued pack and the responses | Procurement, or a buyer administrator | The programme starts and nobody can say which version of the specification the quotations answered |
The distinction between a role and a person is where most sign-off maps quietly fail. If the pack names an individual, the map breaks the day that individual leaves, takes leave, or changes job. If it names a role with a stated trigger, the map survives churn and a new signatory can be found by asking a question rather than by searching a phone list. Write the role, the trigger, and the delegation rule together. What can be supplied is listed in the smart door locks.
The Arena: Four Sign-Off Models and What Each One Means
Buying organisations run one of four models, usually without naming it. Knowing which one you run is the difference between fixing a delay and assigning blame to a slow reviewer.
| Model | Who signs | What the signature means | How it fails | Where it holds |
|---|---|---|---|---|
| Sole Signatory | One named buyer, no concurrence | This request is mine and I have read it | Nobody with authority looked at the parts of the request the buyer cannot judge | Repeat orders of an already-approved variant |
| Full Council | Every function signs every request | No function is surprised by a decision | Queue time grows, later items get a nod, nobody is accountable for the outcome | First programmes with new door types or new platforms |
| Threshold Delegation | Procurement signs below stated limits, named functions concur above them | Within my delegated authority, this is decided | The thresholds are stale, so the wrong requests get escalated and the right ones do not | Steady-state buying with occasional exceptions |
| Evidence-Gated Review | A named reviewer signs only when a document pack is complete | The pack is complete enough to be decided | Reviewers become packet checkers and stop reading the specification | Programmes where evidence quality drives the outcome |
The four are not a maturity ladder. The Full Council is not more professional than the Sole Signatory; it is better suited to one situation and worse suited to another. The rounds change the winner every time, which is why the model is chosen per programme rather than declared once at company level.
Round Rules: How Each Model Is Judged
Each round scores the same four criteria, so the comparison stays honest across situations.
Speed to issue. How much calendar time passes between the first draft and the request leaving the building. This matters more than it sounds, because a request that goes out late changes which suppliers can answer properly.
Coverage of the risk that actually bites. Not the number of reviewers, but whether the specific failure that would hurt this programme has a named person responsible for catching it before the request is issued rather than after the contract is signed.
Defensibility of the record. Whether an auditor, a successor or an approving authority could reconstruct the reasoning later. A record that cannot be reconstructed is a memory, and memories fail in exactly the situations you need them.
Cost of a refusal. What it costs when a reviewer says no. A cheap refusal is a fast programme; an expensive one is a queue that people stop using.
Two rules of engagement. Every round is an illustrative scenario assembled from commonly reported buying experience, not a record of any real project, company or supplier. And the scores are about governance, never about the commercial outcome: the question in every round is who decides and on what question, not who bids lowest.
Round 1 — The Repeat Order You Already Buy Every Quarter
The situation: an illustrative mid-sized operator re-orders the same lock variant for a rolling refurbishment programme. The variant was approved earlier, the door is unchanged, the specification is already written, and the volumes are steady. What is being decided is a repeat purchase, so does the full sign-off map apply again?
Sole Signatory moves fastest here, and the temptation is real because there is genuinely little new to review. The danger is not incompetence; it is drift. Between the last order and this one, the credential platform you run may have changed, the plant may have moved a process step, the radio module inside the lock may have come from a different supplier, and the door profile may have been quietly standardised in a way nobody updated the specification for. A sole signatory reads the request against their memory of the last one, and memory does not update itself.
Full Council produces the worst outcome in this round. Nine roles reviewing an unchanged specification is nine people re-reading a document they approved last time, and the queue becomes the delivery risk. Council members learn to move items quickly, and fast-moving items are the ones that get the least attention.
Threshold Delegation handles this round correctly, provided the threshold is defined by the object and not by the value. "Routine re-order of an approved variant" is an object-based trigger. "Anything under a certain value" is a value-based one, and value-based triggers push ordinary repeat orders into a queue when the unit value moves.
Evidence-Gated Review is also strong, and for an obvious reason: the evidence already exists. The gate is cheap because the pack was built last time and can be refreshed rather than rebuilt. The gate must still be re-run, because "the evidence existed" and "the evidence still covers the variant being ordered" are different statements.
The round's lesson is narrow and useful: the correct model for a repeat order is the one with an explicit re-verification step, because a lock specification ages silently.
Round 2 — A First Programme on Doors Nobody Has Measured
The situation: an illustrative first programme across a mixed estate, where the doors come from an older building stock, the thickness and profile vary, and no survey has been done. Now the technical reviewer has something to block on, and the request is going out to suppliers who will each propose a different answer.
Sole Signatory fails hardest here, because the buyer is being asked to sign a specification containing a door description that nobody has verified. The request goes out, the responses come back with different assumptions about the door, and the reconciliation work returns to the buyer as a negotiation that now looks like the buyer's problem.
Full Council is at its best in this round, and the reason is specific rather than general: the technical reviewer and the operations reviewer each hold a veto that is real, and both are needed. But the council only works if it is convened as a review of the specification, not as a meeting where each function presents its area in turn. The former finds the variant mismatch; the latter produces minutes.
Threshold Delegation works if the trigger catches new door types. A threshold written around programme value will wave this through, because a single building may be small relative to the organisation's total purchasing. A threshold written around the object — new door profile, new variant, new platform — catches it, because the object is what changed.
Evidence-Gated Review is where this round is decided, because the pack is the door survey. If the request can state the door population, the measurement method and the acceptance rule for an unmeasured door, the pack is complete enough to be decided. If it cannot, the correct outcome is not a faster request; it is a survey, and the sign-off map should make that explicit rather than leaving it to a reviewer's discretion.
The published Smart Lock Engineering Drawing Checklist and the door measurement guidance in the Smart Lock Door Measurement Guide exist because this is where specification questions belong, and the survey side of it is in the Smart Lock Retrofit Site Survey.
Round 3 — A Private-Label Programme With the Brand Owner
The situation: an illustrative distributor or channel partner is launching its own brand of locks and issues an RFQ that will define how the product is marked, packaged and presented in the market for years.
Sole Signatory is the riskiest model in this round, because a private-label programme contains questions that procurement cannot answer at all. Who may use the buyer's name and marks, on which packaging, for how long, and what happens to the artwork and the approvals when the relationship ends are not procurement decisions, and they are not price decisions either.
Full Council pulls the brand owner in, but it also pulls in people the brand owner has never met, which slows the pack down with comments that have nothing to do with the marks. A better structure is a narrow concurrence: the brand owner signs the naming, marking and asset questions, while the commercial reviewer signs the volume and cost questions.
Threshold Delegation is nearly useless in this round unless the private-label route is treated as its own trigger rather than as a normal order. A first private-label programme is a first-of-kind object, and first-of-kind objects should not sit inside a routine band.
Evidence-Gated Review has a specific shape here: the gate is a set of approvals rather than a set of test documents, and the reviewer is the person who owns the brand. The published Private Label Smart Lock Supplier Checklist: A Weighted Tear-Off for Brand Owners and Smart Lock OEM/ODM Project Checklist are the closest published equivalents for this shape of programme.
The round's lesson: some roles sign questions that only exist in certain programmes, and a sign-off map written once for all buying will either ignore them or apply them everywhere.
Round 4 — An RFQ That Crosses a Border for the First Time
The situation: an illustrative first import into a market the buying organisation has never supplied, with a new delivery destination, a new set of market-facing documentation, and a contract form nobody in the team has used before.
Sole Signatory is dangerous in a way that is easy to miss, because the request looks like a normal request with a different address on it. The cross-border review is not about the address; it is about which document form governs, which party bears risk between production and arrival, what the import side needs from the supplier, and which local requirements the finished product must meet.
Full Council is slow but it is the right instinct, provided the legal and compliance reviewer is given a scoped question rather than an open one. An open question produces a memo about everything. A scoped question produces a decision.
Threshold Delegation needs two separate triggers here, and this is the round where a single threshold fails hardest. One trigger covers first supply into a new market. The other covers first use of a contract form, because that is a different risk with a different reviewer.
Evidence-Gated Review is where the documentation bundle and the market-facing evidence meet, and where the buyer must be careful to ask rather than to assume. Which authority signs off in the destination market, which documents the finished product needs there, and how those documents must be named and dated are questions to put to the supplier and to local advisers, with the answers confirmed rather than inferred from another market's rules.
CONFIRM-WITH-LOCAL-LEGAL — the contract form, the governing rules, the enforceability of the terms you intend to sign, and which market-specific requirements apply to this product and this destination. Confirm the import and classification position separately with your customs broker rather than inferring it from the supplier's description.
The round's lesson: a new market is a new trigger for legal review, and the trigger is easier to defend in the record than the outcome ever is.
Round 5 — A Framework RFQ With No Firm Volume Behind It
The situation: an illustrative buyer wants a multi-year arrangement or a standing offer, with a price structure and a service position, without being able to promise a volume yet.
Sole Signatory has the strongest apparent case and the weakest actual one. A framework looks like a light document, so it gets a light review. But the question at sign-off is who may commit the organisation to a volume band, a price-adjustment mechanism and a service level, and that is a question about authority rather than about the request.
Full Council is the right model, because this round contains the commercial reviewer's hardest question and the operations reviewer's most specific one. Volume bands and adjustment mechanisms are not decorative, and the reviewer who signs them should be someone who can see the downside of the band they are agreeing.
Threshold Delegation is where this round is usually mishandled. A framework that carries volume exposure is not a routine order, and a threshold written around order value will not see it, because the first order under a framework is small.
Evidence-Gated Review needs a gate that is not only about documents. The completeness test for a framework includes whether the volume structure has been named as a band rather than a point, whether the adjustment mechanism identifies what moves the price and how a change is communicated, and whether the payment structure has been set out as a choice rather than as an assumption.
CONFIRM-PER-SUPPLIER — the commercial structures named in a response, including the price basis, the order-size basis, production timing, the payment structure and the warranty period, are the supplier's own and are confirmed only when they arrive in a written offer letter against this specification.
The round's lesson: the strongest reason to build a formal sign-off map is the request that is small today and binds the organisation for years.
The Scoreboard: Where Each Model Wins and Where It Breaks
| Criterion | Sole Signatory | Full Council | Threshold Delegation | Evidence-Gated Review |
|---|---|---|---|---|
| Speed to issue | Fastest | Slowest | Fast for routine, slow for exceptions | Fast once the pack exists |
| Coverage of real risk | Weak on anything the buyer cannot judge | Strong, if convened properly | Strong when triggers are object-based | Strong on document risks, weaker on organisational ones |
| Defensibility of the record | Weak, it is a memory | Strong, if minutes are kept | Strong, if the threshold log is kept | Strong by construction |
| Cost of a refusal | Low, one person can overrule | High, a meeting to renegotiate | Medium, escalations are formal | Medium, a refusal is usually a missing document |
| Worst failure mode | Silent drift | Council fatigue | Stale thresholds | Packet chasing |
| Rounds it wins | 1 | 2 | 1, 5 | 1, 2, 4, 5 |
Read the bottom row as the decision aid rather than the top. No model wins every round, and the row that matters most is the fourth: a refusal is a design feature. The model that makes refusal expensive is the model that suppresses the information you most need, because people route around a queue instead of pushing a document back through it.
Two hybrid patterns are common in practice, because they are what most mature organisations end up with. The first is threshold delegation with an evidence gate: procurement decides routine requests, and everything above the trigger is released only when the pack is complete. The second is a two-stage council, where a small group of roles sets the specification once per programme and the full council signs only the exceptions.
The Sign-Off Map: Nine Roles, Nine Vetoes, Nine Questions
Each role signs one question. A role that cannot state its question in a single sentence usually does not need to be on the list yet.
| Role | What the signature covers | The question they are answering | What they want to see attached | What a refusal sounds like |
|---|---|---|---|---|
| Procurement manager | That the request is coherent, complete and worth sending | Is this request fit to go out | Scope, door population, destination, response date, the commercial questions to be asked | "I will not issue this until the scope names the doors in scope" |
| Technical and specification reviewer | That the product described is the product needed | Which variant, on which door, in which environment | Variant, finish, dimensions, door profile and thickness, power, environment, credential model | "Two of these variants fit and only one of them fits, and the request does not say which" |
| Quality and inspection reviewer | That acceptance is defined and judged the same way by everyone | Who judges the batch, against what, and what happens if it fails | Acceptance standard, sampling approach, inspection point, retained sample, the record kept | "There is no agreed definition of pass, so I cannot sign an inspection point" |
| Commercial and finance reviewer | The shape of the cost and the risk, not the number itself | What moves this cost, and how would we find out first | Price structure and what adjusts it, order-size basis, payment structure, shipping position, tooling cost treatment | "I can approve the structure but not a request that leaves the adjustment mechanism unstated" |
| Legal and compliance reviewer | The form that governs and the boundaries inside it | Which document are we signing, and what does it leave out | Contract form, confidentiality scope, liability position, warranty and remedy structure, data terms, exit and termination | "I need the intended form named before I can review it, and I am not the approver of the commercial numbers" |
| Operations and project owner | That the thing can arrive, be installed and be handed over | When does it need to be on site and who installs it | Required-on-site date, packing and marking, document bundle, installation and commissioning scope | "The delivery position does not match the programme, and the gap is on the request" |
| Brand or marketing owner | Naming, marking, artwork and asset use | Whose name goes on it, and who else may use it | Mark licence, artwork and packaging approvals, marketing asset permissions, exit position | "Nobody may approve artwork that I have not seen, and the artwork is not in the pack" |
| IT, security or credential reviewer | How the lock connects and who owns the event data | Who issues the credential and where does the log live | Platform and protocol choice confirmed on the product page for the specific SKU, provisioning path, log ownership, data handling | "I cannot sign a request that leaves the credential model unnamed, because the integration depends on it" |
| Service and spares reviewer | What is stocked, replaced and retired | What do we hold, and what do we tell the customer | Spare-parts list, failure and return path, end-of-life notification, last-time-buy practice, training scope | "There is no stated position on parts, so I cannot sign the service side of this" |
Two columns do the work. The third is the question that makes the signature meaningful; a reviewer who cannot answer it has no veto to exercise. The fourth overlaps with the evidence work, covered in detail in Evidence to Attach to RFQ Issue.
What Each Reviewer Is Actually Asking
The table above compresses nine reviewers into nine rows. Underneath, each of them is asking for a small number of specific things, and the buyer who knows them can put them in the request rather than waiting for them to come back as refusals.
The technical reviewer asks which variant, on which door, with which power source, in which environment, and what happens at the edge of the specification. A request that names a finish but not the exposure environment, or a thickness but not the door profile, has pushed an unanswered question into the responses. Where the lock connects to the building system, the credential model has to be named rather than implied; whether a given model uses a Matter or Thread based ecosystem, or a proprietary one, is confirmed on the product page for the specific SKU rather than assumed from a brochure.
The quality reviewer asks what "good" means measurably, who judges the batch, at what point in the programme the judgement happens, and what a failure permits the buyer to do. That last question is the one most often left unanswered, and it is the one that matters when goods arrive short of the standard. The published Smart Lock Pre-Shipment Inspection China and Smart Lock Quality Inspection China cover the inspection side for a cross-border programme.
The commercial reviewer asks what the price is made of, what moves it, what the order-size basis is, what the payment structure does to risk between production and delivery, and what happens to the tooling if the programme ends. This reviewer is not asking for a lower number. They are asking for a structure, because a structure is what can be compared across responses and audited later. The published Smart Lock Supplier Evaluation Scorecard and Smart Lock MOQ Negotiation Guide address the same terrain from the evaluation side.
The legal and compliance reviewer asks which form is intended, what is protected and for how long, what is excluded from liability, what the remedy is when something fails, what happens to personal data about door events, and what happens to the installed base if the programme ends. This is the reviewer whose scope changes with jurisdiction and with the market, and it is the reviewer whose answer belongs in front of counsel rather than in a procurement policy.
The operations reviewer asks when the goods are needed on site, what packing and marking look like, what documents must arrive with the container, who installs, who commissions, and who signs the handover. The published Smart Lock Purchase Order Checklist and Smart Lock Commissioning Checklist are the closest published equivalents for the handover half of the question.
The brand owner asks what name and artwork appear on the product, who may use them and in which markets, who approves a change, and what happens to all of it if the relationship ends. The IT and credential reviewer asks who issues the credentials, how the lock is enrolled, where the audit trail lives, who operates the platform, and how personal data about who opened which door is handled. The service and spares reviewer asks what is held on the shelf, what is replaced and what is returned, how end of life is communicated, and who trains the technicians who will do the work after the project's own staff have moved on.
The Sequence: Why Order Decides More Than Count
Given the same nine roles, the order in which they are asked changes the outcome more than the number of them does. Three principles do most of the work.
Nobody signs against an unexamined specification. The technical review has to come first, because every other reviewer's signature depends on what the specification actually says. A commercial reviewer signing against a request that has not had its variant settled is signing a question mark.
Nobody signs against a question they cannot see. A request that does not state the commercial questions is a request the commercial reviewer must invent, and an invented question is rarely the right one. Put the questions in the request as questions, and let the response answer them.
The record owner signs last and then issues. The last signature is not ceremonial: it is the point at which the version is frozen and the request goes out. After that point, changes go out as a revision with a new number and a new signature, which is a discipline that pays for itself at the second order of the programme rather than the first.
stateDiagram-v2
[*] --> Drafting
Drafting --> TechnicalReview : Scope, variant and doors written in
TechnicalReview --> Rework : Ambiguity or conflict found
Rework --> Drafting : Questions returned to the author
TechnicalReview --> CommercialReview : Specification accepted as the object of review
CommercialReview --> LegalReview : Commercial questions stated in the request
LegalReview --> Release : Contract form and review scope agreed
Release --> Issued : Version frozen, signature recorded, request sent
Issued --> [*]
The rework loop is the part most organisations remove for speed and need later. The question it answers is simple: how many issued requests came back with a change that was never re-reviewed by whoever made it.
The Two Axes Every Sign-Off Policy Has to Choose
Stripped of the round-by-round detail, a sign-off map sits on two axes. The horizontal one is speed to issue against coverage of the risk. The vertical one is how weak the record is against how defensible it is. Where each model lands is worth knowing before you argue about it.
quadrantChart
title Where the four sign-off models sit
x-axis Slow to issue --> Fast to issue
y-axis Weak record --> Defensible record
quadrant-1 Fast and fragile
quadrant-2 Fast and defensible
quadrant-3 Slow and fragile
quadrant-4 Slow and defensible
Sole Signatory: [0.86, 0.16]
Full Council: [0.14, 0.86]
Threshold Delegation: [0.70, 0.72]
Evidence Gated Review: [0.50, 0.92]
The lower-left quadrant is where a policy should not sit for long. The upper-right quadrant is occupied mostly by the hybrid models, which get there by making routine requests fast rather than by making every request slow. The interesting question is not where a model sits but how far it can be moved: a Full Council that convenes once per programme rather than once per request moves rightwards, and an Evidence-Gated Review whose gate is automated for the repeat case moves leftwards.
The horizontal axis gets argued about most, usually as a trade-off between speed and rigour. It is not quite that. Coverage and speed are only in tension when the review is aimed at the wrong object. Reviews aimed at the specification rather than at the request can be both fast and wide, because a well-built request has already answered most of the reviewer's questions.
Where Sign-Off Meets Evidence, Timing and the Record
A sign-off map is a decision structure, and the other parts of this motion are its inputs and outputs. The pieces below are worth reading alongside this one rather than instead of it.
Evidence to Attach to RFQ Issue covers the document side: what each reviewer needs to see in the pack, and how to ask for it so that it arrives attached rather than chased.
Timing Dependencies Around RFQ Issue covers the calendar, including the dependencies between your internal review steps and the time a supplier needs to answer properly. A review that ends after the realistic answer window has changed what a response can contain.
Internal Approval Pack for RFQ Issue covers the artefact that travels between the roles in this map: what it contains, who builds it, and how it is versioned.
Records to Keep After RFQ Issue covers what survives after the signatures are collected, which is the half of the record that the sign-off map depends on later.
RFQ Issue: Reasonable Asks from Both Sides covers the requests a supplier makes of you during clarification, and how to tell a reasonable request from one that will move your timeline.
What Kills a RFQ Issue and How to Recover covers the failure modes on the receiving side, when the request has left and the responses never arrive.
For a published starting point, the Smart Lock RFQ Guide for International Buyers sets out the request as a whole, and the Smart Lock Procurement Checklist for B2B Buyers covers the surrounding programme.
Three Ways Sign-Off Fails That Have Nothing to Do With the Supplier
The failures worth planning for are internal, because the supplier is usually not the reason an RFQ stalls. These three account for most of them.
The circular reference. Two roles each need the other's output before they can sign, and neither is willing to move first. Technical waits for the commercial position, commercial waits for the technical specification, and the programme waits for both. The fix is not persuasion; it is a written order of review with an explicit tie-break, so that one role signs provisionally against a stated assumption and both assumptions are recorded.
The rubber stamp. Everyone "approves" in a single meeting without reading the pack, because the pack is long and the meeting is short. The signature count looks healthy and the record contains nothing. This is the most common failure in a Full Council model and the least visible, because the approval rate looks excellent. The countermeasure is boring and effective: each signatory records the answer to their own question in their own words, in the pack, before the meeting.
The ghost approval. The person who signed has left, the programme has paused, and nobody can reconstruct why the specification said what it said. The signature was valid when it was given and useless afterwards. A map that names roles instead of people, keeps the issued version frozen, and states where the pack lives survives this; a map with signatures on scattered emails does not.
The Mistakes in RFQ Issue that Cost Money Later piece covers the commercial consequences of the same three failures, and RFQ Issue: How to Run it for a Smart Lock Purchase covers the meeting mechanics that make the map real on the day.
Building the Sign-Off Map for Your Own Organisation
The map is a document. It is short, it is versioned, and it names roles rather than people. The table below is the shape of it, with the cells left blank because the answers are yours to fill.
| Field | What goes in it | Who completes it | Why it exists |
|---|---|---|---|
| Role | The role title, not the name | Procurement, with operations | Survives churn and delegation |
| Trigger | The condition that brings this role in | Procurement | Stops every request reaching every reviewer |
| The question | One sentence, in the reviewer's own words | The reviewer | Makes a signature mean something |
| Evidence needed | The specific documents or facts this role requires | The reviewer | Turns a refusal into a missing item |
| Veto scope | What this role can hold the request for | The reviewer | Stops a small concern becoming a programme stop |
| Delegation | Who signs when the role holder is unavailable | Procurement, with the role holder | Removes the queue in a holiday week |
| Escalation | When a disagreement goes above the role | Procurement | Prevents a deadlock with no tie-break |
Three rules make it work. The trigger must be written against the object of the request, not against its value, because value-based triggers fail precisely when an object changes. The question must be answerable by the person who wrote it, in the form of one sentence. The veto scope must be stated, because a reviewer with an unlimited veto will use it on something small, and a reviewer with no stated veto will not use it on something serious.
CONFIRM-WITH-LOCAL-LEGAL — this map is a working structure to discuss with your counsel, not a document ready to sign. Which roles must review, what they may refuse, and how a disagreement is escalated are governance decisions with legal consequences in your jurisdiction, and they belong in front of the people responsible for them.
The map also needs an owner and a review date, and it needs to be checked against the last issued RFQ: pick a request that has already gone out and ask each role whether the question in the map is the question they actually asked. Where the answers differ, the map is wrong, and the map is the cheaper thing to fix.
The Decision Sitting On Your Desk
The question behind the title is not which role signs. It is which roles can hold the request, and what each of them is holding it for. An organisation that knows the answer has a map; an organisation that does not has a signature on a form, and the difference shows up months later, when a specification has to be defended, a change has to be approved, or a programme has to be restarted with the same mistakes.
Three things to do before the next request leaves. Name the roles that hold a veto on your next programme, and write each one's question in one sentence. Write the trigger that brings each of them in, against the object rather than the value. Then take the pack you built last time, find the gap that cost you a week, and decide which role should have caught it and where that role's signature belongs.
The commercial side of the request follows its own path. The price basis, the order-size basis, production timing, the payment structure, the warranty period and the shipping position come back in a written offer, and the structures you ask for are the ones you can compare across responses. You can request those structures, and any sample, at /request-a-quote and /order-samples.
What remains is the discipline of freezing the version at the last signature, issuing it, and keeping it. That single habit is what turns a set of approvals into a record you can still use when the programme changes, the team moves, or the question comes back two years later from someone you have never met.

