Deal Desk Simulation: Cross-Functional Approval Process Roleplay

Two ways to run it: gate-sequential versus committee-simultaneous

Most teams building this exercise reach the same fork within ten minutes of planning: do you stage the approval as a sequential chain of gates, or as a simultaneous committee where every approver sits in the same room at once? Both are legitimate. They train different muscles, and choosing badly means you spend 90 minutes rehearsing a process your company does not actually run.

The gate-sequential format mirrors how most CRM-native approval flows behave. The rep submits, the first approver (usually a sales manager or VP) evaluates discount authority, and only on approval does the record advance to finance for terms, then to legal for contract language. Each approver sees the deal cold, with only what the rep wrote into the record. In simulation, you run this as three short rooms: 8 minutes with the VP, 8 with finance, 8 with legal, with a 2-minute rework window between each where the rep can fetch missing evidence. The instructional value is that it makes evidence gaps expensive in a way participants feel. A rep who never documented the economic buyer gets bounced at gate one, watches the clock burn, and arrives at legal with no time left. That is exactly what happens in production when a deal desk queue is serial and each hop adds a business day.

The committee-simultaneous format puts the rep in front of all three approvers together, typically for a single 20-25 minute negotiation. Nobody waits. Objections collide in real time: finance's cash-flow concern about extended terms runs directly into legal's liability cap and the VP's margin floor, and the rep has to trade across all three at once. This is closer to how escalated, strategic, quarter-end deals actually get resolved — someone pulls the approvers into a room because the serial queue would blow the close date. The instructional value is different: it teaches cross-functional trade-making, not documentation hygiene. The rep learns that legal will often flex a liability cap if the term length increases, and that finance will accept longer payment terms in exchange for a prepayment or an annual-in-advance structure.

There is a third variation worth naming, because teams invent it accidentally: the hybrid, where gates one and two run sequentially and gate three convenes everyone. This is the most realistic for companies over roughly 200 employees, where routine approvals flow through workflow automation and only the exceptions get a human huddle. It is also the hardest to facilitate, because you are managing two formats and two timing models in one session. Run it only after your team has done each pure format at least once.

The trade-off summarized: sequential exposes documentation debt and queue latency; simultaneous exposes negotiation skill and cross-functional priority conflicts. Sequential is better for a team whose approvals are slow. Simultaneous is better for a team whose approvals are fast but whose outcomes are bad — deals approved with terms nobody would defend in a renewal conversation.

Choosing the format for your organization

The decision is not a matter of taste. It follows from what your approval data already tells you, and from how your approval authority is actually delegated. Before you pick, pull three facts: the median elapsed time from approval request to final approval, the number of distinct approvers a non-standard deal touches, and the rework rate — the share of submitted approval requests that get returned to the rep for more information.

If your rework rate is high, sequential wins. Rework is a documentation problem, and the sequential format is the only one that punishes weak documentation visibly. If your elapsed time is long but rework is low, your problem is queue latency and approver availability, not evidence — and a simulation will not fix an availability problem, so run the simultaneous format to at least improve the quality of the trade the rep makes when they do get the room. If both are healthy but discounting is drifting upward quarter over quarter, run simultaneous with a hard margin floor and let the rep discover that the VP will not move on price without a corresponding concession on term, volume, or payment timing.

Also weigh how many people you actually have. Sequential needs a minimum of four roles per group plus a facilitator, and it scales awkwardly past three groups because you need enough approvers to staff parallel rooms. Simultaneous runs cleanly with a single group of five to six in a room while everyone else observes, which is why it is the default for a large all-hands training. If you have fewer than eight participants, run simultaneous and rotate the rep chair each round.

One more input to the decision: whether your legal team will participate. Legal counsel is the scarcest role in any deal desk simulation, and the sequential format demands their presence for a defined 8-minute window while the simultaneous format demands a full 25 minutes. If legal can only give you a short slot, sequential lets you schedule around them. If legal cannot attend at all, do not fake it with a sales manager reading a policy card — instead, pre-record legal's positions as a one-page policy sheet and have the facilitator voice them, and be explicit with participants that this is a stand-in. A roleplay that misrepresents legal's real posture trains reps to make promises the company will not keep.

The numbers each format needs to make the scenario bite

A deal desk Simulation only works if the scenario carries real constraints. Vague scenarios produce vague roleplay: the rep asks for a discount, the approver says yes, everyone learns nothing. Build the packet with hard numbers that create genuine conflict between functions.

Start with a deal that sits clearly outside standard terms on at least three dimensions at once. A single exception is a one-conversation approval. Three simultaneous exceptions force prioritization — the rep cannot win all three and must decide which matters most to the customer. A workable shape: a multi-year subscription in the low-to-mid six figures of annual contract value, a requested discount roughly 50% above the rep's delegated authority, payment terms two to three times longer than your standard net terms, and a liability cap request well above policy. Use your own real numbers rather than the ones in any template, because participants will immediately spot figures that do not match their world and will stop taking the exercise seriously.

Give each approver a written authority threshold and a written hard limit, and make sure they differ. The VP might hold authority up to a certain discount percentage, with anything beyond requiring finance concurrence — that is the threshold. The hard limit is the margin floor below which the VP will not go regardless of who concurs. Finance gets a threshold on payment terms and a hard limit expressed as a cash-flow cost: extending payment terms on a large contract for an extra 60 days ties up working capital, and finance should be able to state roughly what that costs at your company's cost of capital. Legal gets a standard liability cap tied to contract value, a stretch position for strategic accounts, and a category that never flexes — typically indemnification for intellectual property infringement, data breach obligations, or anything touching regulatory compliance.

The scarcity is the point. If every approver's hard limit is soft, the rep learns that pushing always works. If every limit is absolute, the rep learns nothing except that approval is a wall. The right calibration is roughly one flexible dimension per approver and one immovable one, so the negotiation has both give and consequence.

Then equip the rep with evidence artifacts, or deliberately withhold them. A qualification framework such as MEDDPICC gives you a natural checklist of what evidence should exist: quantified metrics on the customer's current spend and expected outcome, the identified economic buyer, documented decision criteria and decision process, the mapped paper process including procurement and purchase-order timing, a named champion with confirmed access to budget, and the competition. Populate four of those eight and leave four empty. In round one, the rep discovers which gaps stall the approval. In round two, they present with the gaps filled and feel the difference in approval speed directly.

Finally, put a clock on it and enforce it. The two most common facilitation failures are letting a gate run long because the conversation is interesting, and letting the rep negotiate without ever stating a specific ask. Cut off at the stated time even mid-sentence. The abruptness is instructive: real approval windows close, and a rep who spends four minutes on context and thirty seconds on the ask has just modeled the behavior that loses deals at quarter-end.

Building the packet, running the rounds, and sequencing the debrief

Sequencing matters more than content here. A deal desk Roleplay that opens cold with the negotiation produces stilted performances, because participants have not yet accepted the fiction. A session that spends too long on setup never reaches the second round, which is where the actual learning lives — round one surfaces problems, round two proves the fix works.

Open with a short friction story, roughly 10 minutes. Have participants pair off and each describe one real approval that either saved or killed a deal, naming the deal size, the function that caused the delay, and the outcome. Then take two examples aloud. This does two things: it establishes that approval friction is a shared, named problem rather than a criticism of any function, and it seeds the room with concrete language you can call back to during the debrief. Facilitators should listen for whether the stories are about missing evidence or unclear authority, because that tells you which lever to emphasize.

Spend about 15 minutes on scenario setup. Distribute the packet — the opportunity record with the requested terms, a call excerpt or note capturing what the customer actually said about competitive pressure and budget, the qualification field set with its deliberate gaps, and the legal exception request. Assign roles: rep, sales leader, finance, legal. Then require each role to write down three questions they will ask before the meeting starts. This single step is the difference between a roleplay and a conversation. Approvers who have written questions in advance interrogate; approvers who have not simply react to whatever the rep says.

Run round one for 25 minutes: five minutes of rep presentation, ten of questioning, two for each approver to write a written decision plus one condition, and the balance for immediate debrief. The written decision is essential. Verbal approvals in roleplay are mushy; a written "approved conditional on X" forces the approver to name a specific, testable condition, and those conditions become the raw material for the checklist you build at the end.

Then change the facts and re-run. Between rounds, hand the rep new information that closes two of the four evidence gaps and delivers one customer concession — for instance, procurement agrees to a shorter payment term in exchange for holding the discount, and the champion's budget authority is now confirmed on a recorded call. Rotate roles one seat clockwise so everyone experiences at least two functions. Round two runs 20-25 minutes and should end in an approval with explicit written conditions. If it does not, your hard limits were set too tight; note that and recalibrate before the next session.

Close by building the checklist together, about 10 minutes. Ask each group to contribute one condition that must be satisfied before any deal above your exception threshold enters the approval queue. You will typically get six to eight items clustering into evidence, authority, and policy categories: recorded confirmation of budget from the economic buyer, qualification fields populated in the CRM, a confirmed purchase-order timeline, the named approver for each exception dimension, liability language within policy, and payment terms within the finance threshold. Write them on a board as the group states them. Do not pre-write the list and reveal it — a checklist the room authored gets used, and one the facilitator handed down gets filed.

End with a 5-minute commitment round: each participant names one specific change to their next real approval submission. Specificity is enforced — "document better" does not count, "attach the recording where the CFO confirms the budget line" does. Follow up in writing within 48 hours with the checklist, and check the same approval metrics you pulled at the start after two full cycles.

Facilitation traps and how to run it remotely

Several failure modes recur reliably enough to plan against. The most common is approvers playing nice. Peers do not enjoy blocking peers, and without explicit permission to hold the line, finance and legal will approve everything in round one, which collapses the exercise. Fix it by giving each approver a written card with their hard limit and a single instruction: you may not cross this, regardless of the argument. Frame it in the room as "your job today is to be accurate, not agreeable."

The second trap is the rep who negotiates against themselves. Under time pressure, reps often open by pre-conceding — "I know the discount is aggressive, and I'm sure we can't do the payment terms." Watch for it and name it in the debrief. It is the single most transferable observation the simulation produces, because it happens constantly in live approval conversations and reps rarely notice they do it.

The third is facilitator drift into subject-matter argument. When legal and finance disagree about a liability cap, the facilitator's instinct is to adjudicate. Do not. Let the disagreement stand and surface it in the debrief as a real policy ambiguity that needs resolution outside the room — you have just found a genuine gap in your company's approval framework, which is worth more than the training itself.

The fourth is over-discounting without legal review, which shows up almost every time in round one. The rep wins the price argument, everyone feels good, and nobody notices the indemnification request is still sitting unaddressed in the packet. Let it happen. The moment in the debrief where someone realizes an approved deal still contains an unreviewed liability exposure is worth more than any slide.

For distributed teams, the format adapts cleanly but needs tighter structure. Use breakout rooms per function during preparation so each role can caucus, then bring everyone into the main room for the negotiation itself. Assign a timekeeper separate from the facilitator, because remote facilitators lose track of the clock while managing breakout mechanics. Use a shared document for real-time evidence logging so approvers can see what the rep is claiming as it is claimed — this substitutes for the physical packet and, if anything, works better, since it makes the gap between what the rep asserts and what is documented immediately visible to everyone.

Two remote-specific adjustments: add roughly five minutes to every segment, because turn-taking is slower over video, and require cameras on during the negotiation rounds. Approval Approval conversations turn on reading hesitation, and a black tile removes the signal the exercise is meant to train. Recording the session is worth doing if your participants consent — the round-one/round-two contrast is more persuasive on replay than in memory, and the recording becomes a reusable onboarding asset for new hires joining sales, finance, or the deal desk.

One last practical note on cadence. A single session builds awareness; the behavior change comes from repetition. Run it once per quarter with a different scenario shape each time — one quarter a discount-driven exception, the next a contract-language exception, the next a payment-terms and revenue-recognition exception. Rotate which function plays the rep, because a finance analyst who has spent 25 minutes trying to get a deal approved becomes a materially faster approver afterward. That role-reversal effect is the most durable outcome the exercise produces, and it is entirely free.

Related questions

Who should facilitate a deal desk simulation?

Ideally a RevOps or deal desk lead who owns the real approval process, not an external trainer. They can answer policy questions authoritatively and, more importantly, can convert the friction the room surfaces into actual process changes afterward.

How long should the whole session run?

Ninety minutes is the practical floor for two full rounds plus debrief. Sixty minutes forces you to cut round two, which removes the contrast that makes the lesson land. Two hours works for remote sessions or larger groups.

Should the scenario use a real past deal?

Use a lightly disguised real deal. Invented scenarios get dismissed as unrealistic; a recognizable one carries credibility. Change the account name and shift the numbers enough that nobody in the room feels their performance is being reviewed publicly.

What if we do not have a formal deal desk yet?

Run it anyway with whoever currently approves exceptions — often a sales leader and a controller. The checklist the session produces is a reasonable first draft of the approval policy you do not yet have.

How do we know the simulation worked?

Compare the same three metrics you pulled beforehand — rework rate, median approval elapsed time, and average discount on exception deals — after two full sales cycles. Attribute cautiously; other changes will be running concurrently.

FAQ

Does this work if we do not use a major CRM?

Yes. The exercise is tool-agnostic. What matters is that approvers can only see what the rep documented somewhere shared, and that the shared place is the same one your team uses in production. A spreadsheet works for a small team; the point is the discipline of documented evidence, not any particular platform.

How many people do we need to run it?

Four is the minimum: rep, sales leader, finance, and legal, plus a facilitator who can double as one of those roles if necessary. The upper bound is roughly twenty, at which point you should split into parallel groups with a facilitator each, or run one group live while others observe and score.

Should the rep role always be played by a real sales rep?

No, and deliberately rotating it is one of the highest-value adjustments available. Putting a finance or legal participant in the rep chair produces the sharpest behavior change, because they experience the cost of their own approval requirements from the other side of the table.

What if legal cannot attend?

Have legal write a one-page position sheet in advance covering the standard liability cap, the stretch position for strategic accounts, and the categories that never flex. The facilitator voices it verbatim. Tell participants it is a stand-in so nobody leaves believing they have negotiated real precedent.

Can the approval flow itself be automated instead?

Routing can and should be automated — threshold-based routing to the right approver removes queue latency and is straightforward in most modern CRMs. What automation cannot do is teach judgment about which concession to trade. Automate the routing, keep the roleplay for the trade-making.

How often should we re-run it?

Quarterly, with a different exception type each time, and always after a material change to your discount policy, pricing model, or contract templates. A team that trained against last year's liability cap will confidently negotiate against a policy that no longer exists.

Sources

• MEDDIC / MEDDPICC qualification methodology
• Salesforce: approval processes overview
• HubSpot: quote approvals
• Harvard Business Review: negotiation topic hub
• Program on Negotiation, Harvard Law School
• Gartner: sales practice and research
• Association of Corporate Counsel: contract resources
• Investopedia: days sales outstanding
• Investopedia: indemnity

Related on PULSE

• [The Deal Desk Operations Reboot — 60-Min Training](/knowledge/st207)
• [Objection Handling Roleplay Session Blueprint](/knowledge/st0752)
• [Upsell and Cross-Sell Scenarios: Roleplay for Existing Customer Growth](/knowledge/st0740)
• [Handling the 'Not Interested' Objection: A Ready-to-Run Roleplay Session](/knowledge/st0653)
• [60-Min Sales Training: Mapping the Decision Process](/knowledge/st0447)
• [The Sales Process Audit Reboot — 60-Min Training](/knowledge/st211)