Sales & PartnershipsFree
Deal Desk Approval Routing Template
The request splits into items, each routed to the one person who can decide it, with precedent that includes the deals you turned down.
A deal desk request usually arrives as one object. Here is the deal, here are the six things about it that are not standard, please approve. It goes to whoever the highest of those six items requires, plus everyone else in the chain for visibility, and then five people read six items each looking for the one that concerns them. Three days later somebody finally notices that two of the six were never deviations in the first place.
Splitting the request fixes most of that arithmetic before anything clever happens. Six flags become four real deviations across three approvers, thirty approver-items become four, and a 4.6 day serial chain becomes 2.6 days in parallel. The 2.6 belongs to the General Counsel, who owns two contract items that were both sitting in the buyer's term sheet three weeks earlier, and that is the finding worth more than the deal.
Then the part that actually decides it. Fourteen comparable requests were approved in this segment and nine of them closed. Three were denied and two of those closed anyway, nine points cheaper, which makes denying worth about twenty-one thousand more per request on the evidence available. Term separated the winners from the losers, not discount. Built for whoever owns the desk. The proposal itself is the proposal pack, the quote that follows is quote generation, and the tool index has the rest.
An approval is a financial control, and controls have owners
This is not just an internal habit. For a public company, internal control over financial reporting is defined to include the policies and procedures that provide reasonable assurance that receipts and expenditures are being made only in accordance with authorizations of management and directors. A price concession is an expenditure in every sense that matters here, and the authorization behind it is either traceable to a named person holding the authority to give it or it is not.
That framing changes what a good request looks like. A memo addressed to five people and approved by whoever replies first satisfies nobody's definition of an authorization, because no individual has said yes to a specific thing within their own limit. Four separate items, each answered by the one person whose delegated authority covers it, produces a record that survives an audit, a dispute with the customer and an entire change of sales leadership.
The routing and the record turn out to be the same artifact, which is the part most desks miss entirely. Every decision gets written back against the item rather than against the opportunity, so a discount approval is filed as a discount precedent and a liability concession as a liability precedent. Next quarter's request in the same segment then opens with fifteen comparable decisions and their outcomes, rather than one person's memory of a call eighteen months ago.
How it works
Describe the deal
The terms you want approved, the size, the length and the date it has to land.
Paste the policy
The whole thing, including addenda, because that is where the false flags hide.
River splits and routes
Items, owners and turnaround times, then precedent from your own approvals and your denials.
Send three messages
One per approver, each containing only the item that person can actually decide.
What you get
- The request split into separate items, each with the policy line it actually breaches
- Items that turn out to be permitted already, which is usually two of every six flagged
- Each real deviation routed to the lowest authority whose delegation covers it
- Precedent from your own approvals, with the percentile this request sits at
- The denials, and what happened to those deals, which no desk normally keeps
- The variable that separated closed from lost, which is rarely the discount
Common questions
We do not have a written approval matrix.
Then the first output is one, reverse-engineered from whoever has actually been signing things off. Feed it a year of approvals and the pattern is usually clear within twenty decisions, including the two or three places where two people both think they own the same item. That ambiguity is worth more than the matrix. Its thresholds come from the standing policy in the concession log.
Why does splitting the request matter so much?
Because a bundled request is decided at the level of its most expensive item, by people reading five things that are none of their business. In the worked example that is thirty approver-items reduced to four. It also means one contentious line stops holding three uncontroversial ones hostage while everybody waits.
Our precedent data does not record whether deals closed.
Most do not, and it is the single most valuable column to add. Approval records without outcomes let a desk approve the same concession forever without ever learning whether it worked. Joining the approval log to closed-won and closed-lost is usually a one-time exercise and it changes the conversation permanently.
Two denials that closed anyway is not statistics.
It is not, and the brief prints the denominator next to it every time so nobody mistakes it for one. Three denials in eight quarters is a hint. The reason it still matters is that the alternative in the room is not a larger sample, it is somebody saying we always approve these, which has no denominator at all.
What if the deal genuinely needs the CFO?
Then it goes to the CFO, with one item in it and the precedent attached. Splitting is not a way of avoiding senior approval, it is a way of making sure the CFO reads one paragraph about payment terms instead of four pages about an SLA. The turnaround difference in the worked example is two full working days.
How does this connect to what we actually quote?
Approvals attach to items, and items become lines on a quote. Once the four decisions are back, the document that has to match them exactly is the quote and order form, where an approved concession that never made it onto the paper is one of the more expensive mistakes available.
Does this replace our deal desk?
It replaces the part of the desk that is transcription and lookup, which is most of the elapsed time and none of the judgement. Somebody still decides. They just decide one thing they own, with the fourteen comparable decisions and their outcomes in front of them, instead of a narrative from a rep who wants a yes. Whether the rival is actually cheaper is the three-year comparison.
Deal Desk Approval Routing Template
Fill in the form and your workspace opens with the work already underway.