| name | deal-desk |
| description | Deal desk: reviews, approves, and structures non-standard sales deals. Use when standing up a deal desk, building approval-threshold matrices, designing deal-review packets, routing deals, or auditing deals for compliance.
|
| license | MIT + Commons Clause |
| metadata | {"version":"1.0.0","author":"borghei","category":"commercial","domain":"business-growth","updated":"2026-05-27T00:00:00.000Z","tags":["deal-desk","sales-operations","discount-approval","commercial-operations","contract-review","deal-velocity","gtm"]} |
Deal Desk
End-to-end deal-desk operational practice: charter, approval thresholds, deal-review packet design, routing automation, velocity analysis, and the governance that turns "every deal is a snowflake" into "we close non-standard deals in 48 hours predictably."
This skill is provider-agnostic: works whether your CRM is Salesforce, HubSpot, Pipedrive, or homegrown. The patterns and decisions transfer.
When to use this skill
| Situation | Skill applies |
|---|
| Starting a deal-desk function from scratch | Yes — start with charter design |
| Reviewing existing deal-desk for slowness / inconsistency | Yes — use scripts/deal_velocity_analyzer.py + bottleneck patterns |
| Defining who can approve what discount / term | Yes — use approval threshold matrix + scripts/discount_authority_router.py |
| Building the deal-review packet template | Yes — see deal-review packet section + scripts/deal_review_packet.py |
| Approving / declining a specific deal | Use the packet generator + approval router |
| Setting pricing strategy | Use business-growth/pricing-strategy first |
| Forecasting / measuring pipeline | Use business-growth/revenue-operations |
| Negotiating an individual contract | Pair with business-growth/contract-and-proposal-writer |
What deal desk does (and doesn't)
Does:
- Review non-standard deals: discounts beyond rep authority, custom legal terms, custom SLAs, multi-product bundles, payment terms outside policy
- Make the approval decision (or route to the right approver)
- Structure the deal: pricing, terms, ramp schedule, success criteria
- Maintain the deal-desk policy — what's standard, what needs approval
- Track deal velocity (time from request → decision → signature)
- Produce evidence for finance / audit (every concession traceable)
Doesn't:
- Set the published pricing (that's pricing strategy)
- Negotiate with the customer (that's the sales rep / AE)
- Close the sale (that's the rep + customer success)
- Run the order-to-cash workflow (that's billing / RevOps)
- Replace legal review (legal is one of the approvers, not the function itself)
A clean deal-desk = the lubricant. Without it, every non-standard deal turns into a multi-week negotiation among engineering / product / legal / finance / executive. With it, those people are consulted by deal desk as needed and the rep gets a yes/no in days.
Deal-desk charter (template)
Every deal desk needs a written charter. Use this template:
purpose:
Deal Desk reviews, approves, and structures non-standard deals to enable
sales to close faster while keeping commercial / legal / financial risk
within company tolerance.
scope:
In-scope:
- All deals > $X ARR
- All deals with discount > Y%
- All deals with non-standard terms (custom SLAs, custom legal language,
payment terms beyond Net 30, ramp
[ ]
See references/deal-desk-charter-and-process.md for the full charter template, including sub-charters per region, intake form spec, and the standard SLAs.
Approval threshold matrix
The matrix defines: for each deal characteristic (discount %, contract length, custom term type), who can approve it.
Standard matrix template
| Deal characteristic | Rep | Sales Manager | Director | VP Sales | CRO | CFO | CEO |
|---|
| Discount 0-10% | ✓ | | | | | | |
| Discount 10-20% | | ✓ | | | | | |
| Discount 20-30% | | | ✓ | | | | |
| Discount 30-40% | | | | ✓ | | | |
| Discount 40-50% | | | | | ✓ | | |
| Discount > 50% | | | | | | | ✓ |
| ACV > $250k | | ✓ | | | | | |
| ACV > $1M | | | | ✓ | | | |
| ACV > $5M | | | | | | | ✓ |
| Multi-year > 12mo standard | | ✓ | | | | | |
| Non-standard payment terms | | | | | | ✓ | |
| Custom SLA / penalties | | | | (with CCO) | | | |
| Custom legal language | | | | | | | (Legal must concur) |
| MSA red-line on liability cap | | | | | | | (Legal must concur) |
| Most-favored-nation clause | | | | | | ✓ | |
| Acceptance criteria / payment-on-acceptance | | | | | | ✓ | |
| Multi-product / cross-BU bundle | | | (each BU lead approves) | | | | |
| Whitelabel / OEM rights | | | | | | | ✓ |
Customize per company stage, ACV distribution, and authority preference (some orgs want CRO at 30%, others delegate further down).
Stacking rule
When multiple non-standard items apply, the highest required approver applies. A $1M deal at 25% discount with custom SLA needs VP Sales (ACV) AND Director (discount) AND VP Sales+CCO (custom SLA) → effectively requires VP Sales sign-off + CCO + Legal concurrence.
Use scripts/discount_authority_router.py --deal deal.yaml to compute the required approvers for any deal.
See references/approval-thresholds-and-routing.md for the full matrix design guide, regional variants, escalation paths, and routing automation patterns.
The deal-review packet
Every non-standard deal gets a packet. Without it, approvers ask the same questions repeatedly and decisions take days instead of hours.
Standard packet structure
# Deal Review: <Customer Name>
## Summary
- Customer: <name, size, industry>
- ACV: $<amount>
- Discount %: <%> (vs standard $<list-price>)
- Contract: <length>, <payment terms>
- Decision needed by: <date>
## Standard vs Requested
| Item | Standard | Requested | Delta |
|------|----------|-----------|-------|
| ACV | $X | $Y | -Z% |
| Term | 12mo | 36mo | +24mo |
| Payment | Net 30 | Net 60 | +30d |
| SLA | 99.5% | 99.9% | +0.4% |
| Liability cap | 1x fees | 2x fees | +1x |
| Termination for convenience | No | Yes (90d) | New |
## Justification
- Why customer wants this: <competitor situation, budget cycle, etc.>
- Why we're considering: <strategic value, logo, vertical>
- Customer leverage: <alternatives they have>
## Financial impact
- Standard ARR: $X
- Discounted ARR: $Y (Z% off)
- Net new gross margin: $A (with cost overlay)
Projected LTV with this discount: $B
Discount payback if customer renews:
Logo value:
Reference value:
Vertical foothold:
Competitive replacement:
Credit risk:
Compliance risk:
Technical fit risk:
Concession follow-through:
[ ] Director:
[ ] VP Sales:
[ ] CFO:
[ ] Legal:
— with reasoning
Discount expires
Customer must agree to:
Customer agrees this is single-instance (not precedent)
Payment must close by
Use scripts/deal_review_packet.py --deal deal.yaml to generate this packet from a deal spec.
Velocity analysis
A slow deal desk strangles sales. Measure and tune.
Key metrics
| Metric | Healthy | Warning |
|---|
| Median time-to-decision | < 1 business day | > 3 days |
| 90th percentile time-to-decision | < 3 business days | > 7 days |
| % of deals waiting on a single approver > 24h | < 10% | > 30% |
| Deals stuck > 7 days | 0 | > 5 |
| Sales rep satisfaction with deal desk (NPS) | > 50 | < 0 |
| % approved (high approval rate may mean threshold too low) | 60-80% | > 95% or < 40% |
| Discount-on-discount: deals where customer negotiated up after deal-desk approval | < 10% | > 30% |
Run scripts/deal_velocity_analyzer.py --deals deals.csv to compute these from a CRM export.
Common bottlenecks
| Bottleneck | Diagnosis | Fix |
|---|
| Single approver bottleneck (one person on everything) | Routing matrix concentrated authority | Delegate; add back-ups; raise thresholds |
| Legal review takes a week | Legal sees every deal | Standard MSA + pre-approved clause library; Legal only on deviations |
| Engineering needed for SLA review | Custom SLAs every time | Publish standard SLA tiers; only deviations route to eng |
| Approval cycle back-and-forth | Packet missing key info | Use the standard packet template; reject incomplete submissions |
| Long executive lag | Exec doesn't have context for every deal | Weekly deal review meeting for batch decisions on smaller items |
| Sales submits incomplete packets | Reps don't know what to include | Intake form that enforces required fields |
| No SLA enforcement | Deals sit in queue with no urgency | Publish + report SLA; aging dashboard visible to leadership |
See references/discount-and-concession-playbook.md for the discount/concession patterns: legitimate reasons for each concession type, how to evaluate, alternatives to discounting, and how to structure performance-based discounts.
Clarify First
Before generating the deal-desk artifact, confirm these inputs. If any is unknown or vague, ASK — do not assume:
Stop rule: ask only the 2-3 that most change the output. If the user says "just draft it," proceed and list your assumptions at the top of the artifact.
End-to-end workflows
Workflow: A rep submits a non-standard deal
- Rep submits via intake form: customer + ACV + requested deviation + justification
- Deal desk triages within 4h: assigns analyst, validates packet completeness, requests missing info
- Deal desk reviews within 1 business day: financial impact, strategic value, risk
- Deal desk recommends approve / counter / decline
- Route to approver(s) per matrix (auto via
scripts/discount_authority_router.py)
- Approver decides within SLA
- If approved: packet signed off, conditions sent to rep with expiration
- If countered: deal desk works with rep on alternative structure
- If declined: clear reason + alternatives sent to rep + customer
Workflow: Stand up a deal desk from scratch
- Draft charter with sales, finance, legal sign-off
- Build the approval matrix — interview key stakeholders, document existing tribal knowledge
- Design intake form — CRM-integrated or Slack-bot
- Hire / appoint deal desk lead + analyst(s)
- Train sales — what triggers deal desk, what info is needed, what to expect
- Soft launch — manual operation for 1 month; track metrics
- Iterate — refine thresholds, automate routing, publish SLAs
- Quarterly review — metrics, threshold adjustments, charter updates
Workflow: Audit deal-desk performance
- Export deals from CRM for the period (CSV with deal IDs, stages, approval timestamps, discounts)
- Run velocity analyzer — compute medians, percentiles, aging, approver bottlenecks
- Sample 10-20 deals for qualitative review (was the packet complete? were conditions met?)
- Identify patterns — are certain reps over-discounting? are certain customers getting MFN clauses inappropriately?
- Propose adjustments — to charter, thresholds, intake form, training
- Present to leadership with metrics + recommendations
Workflow: Quarterly threshold review
Thresholds drift. Quarterly:
- Pull discount distribution for the quarter
- Identify outliers — deals where discount % was anomalous for ACV / segment
- Compare approval rates by threshold — if 30%+ discount deals get approved 95%+ of the time, the threshold is too low
- Compare win rates by discount band — does deeper discount actually improve win rate, or does it just give up margin?
- Adjust thresholds based on data + market shift
- Publish new matrix with effective date; train sales
Anti-patterns
- Deal desk as bottleneck. SLAs published but ignored; deals stack up; sales builds workarounds. Measure + enforce SLAs.
- Deal desk that always says yes. Approval rate > 95% means thresholds are too low — you're rubber-stamping. Tighten or raise thresholds.
- Deal desk that always says no. Approval rate < 40% means policy is too strict OR sales doesn't understand it. Investigate root cause.
- No deal-desk policy. Every deal evaluated case-by-case. Inconsistent decisions; legal exposure; reps gaming the system.
- Concentrated authority. One person approves everything → bottleneck + bus factor. Delegate.
- Pricing strategy disguised as deal-desk policy. If 80% of deals need discounting, the published price is wrong. Fix pricing.
- Discount creep. Each deal raises the bar for the next; eventually published price is irrelevant. Track + reset.
- Concession with no quid pro quo. Customer asks for 20% discount; you give 20% discount. Always trade: 20% for case study, 20% for 3yr contract, etc.
- No expiration on quotes. Customer can come back in 6 months and demand the same terms. Always time-box (typically 30-60 days).
- Single-instance language never enforced. "This is a one-time exception" → next year the customer cites it as precedent.
Tooling outputs
| Script | Input | Output |
|---|
scripts/deal_review_packet.py | Deal spec YAML | Markdown deal-review packet with summary, financials, strategic value, risk, approver list, recommendation template |
scripts/discount_authority_router.py | Deal spec YAML + approval matrix YAML | Required approver(s), routing order, escalation path, SLA-aware ordering |
scripts/deal_velocity_analyzer.py | CSV of deals from CRM export | Median / p90 time-to-decision, aging dashboard, approver bottleneck identification, discount-on-discount analysis |
All scripts: stdlib only, argparse CLI, JSON or markdown output.
References
Related skills
business-growth/pricing-strategy — sets the prices that deal desk enforces deviations frombusiness-growth/revenue-operations — measures the pipeline; deal-desk metrics flow into RevOps dashboardsbusiness-growth/contract-and-proposal-writer — drafts the final contract once deal desk approvesbusiness-growth/channel-economics — channel deals have their own deal-desk patternsbusiness-growth/partnerships-architect — partner-mediated deals route through both deal desk + partnershipsbusiness-growth/commercial-policy — the broader governance framework deal desk enforcessales-success/sales-engineer — provides technical validation in packetsales-success/sales-operations — owns CRM / forecast accuracy that deal desk feeds
