Inside a Deal Desk: A Conversation With Todd Johns, Former Head of Deal Desk at Salesforce

Most of what gets written about deal desk is secondhand — synthesized from surveys and framed by people who have never run the function under quota pressure. This post is different. It is a lightly edited transcript of a roughly fifty-minute conversation with Todd Johns, who built and ran deal desk at Salesforce for twelve years, scaling it from an informal function into an organization supporting an enormous global sales team. We have edited only for length, filler, and readability — the substance of every answer is unchanged from what Todd actually said.

Disclosure: Todd is an advisor to Revolear, and this conversation is part of why — it gave us the standing to ask him to go on the record at this level of detail. We think that is worth knowing as you read it, not something to obscure.

Idea in Brief

The Problem. Most deal desk guidance is written by people who have never run one — it reads like theory because it is.

The Instinct That's Wrong. Assuming deal desk exists mainly to say no to discounts, when its actual job — as Todd Johns describes it — is balancing short-term growth against long-term business health across pricing, deal structure, and every non-legal term in the agreement.

The Fix. Go straight to a practitioner who ran the function at scale, and let the specifics — the "good, better, best" pyramid, the one-minute request template, the price curve as a starting point rather than a ceiling — do the teaching.

What a Deal Desk Actually Does

Raja Singh: You've run a deal desk for over twelve years at a very large company. What's your definition of what a deal desk actually does?

Todd Johns: That can look very different from company to company — I've seen it done differently everywhere. The way we ran it was as an organization looking out for balancing short-term growth with the long-term health of the business. The main lever was pricing at the deal level, but it also extended into deal structuring: the business terms, and working with everyone else around the deal to make sure the whole thing was well structured, not just the price.

Raja Singh: Within that, what are some examples of terms that aren't discounts that deal desk might structure, advise on, or approve?

Todd Johns: Rights for future purchases. Volume purchase agreements — anything where you're setting up the relationship over time and the customer wants some security about what they'll be paying going forward. Any other business terms that aren't purely legal in nature: adding new products into an agreement or an existing MSA, adding or subtracting licenses, and what price you pay for those.

When Does a Company Need a Deal Desk?

Raja Singh: You started at Salesforce when it was already well over a billion dollars in ARR. What's a good telltale sign that a company needs to start thinking about adding a deal desk, or at least deal desk functions?

Todd Johns: That's really two different questions. Thinking through the questions a deal desk answers, and building those into your company — whether digitally or through people — comes first. Then it's a question of headcount. If you're getting to a point where you've got a big enough book of business that it's hard to control everything going into every deal, and you're spending too much time trying to keep visibility and control over every negotiation, you need to think about a deal desk to keep pricing and deal structuring consistent with how you position your product in the market, whether that's premium or discount. Consistency protects the individual customer relationship and the market as a whole.

Diagnosing the "Wild West" Problem

Raja Singh: I'll share something from my own experience as an approver at Salesforce. We had an insurance product priced as a percentage of premium — a bit of a non-standard approach — and we'd done about ten deals in Latin America. Deal desk kept flagging the discounts, which ran around 85%. When we plotted deal size against discount level, it looked like a shotgun pattern — no relationship at all, nothing like the upward curve you'd expect. That struck me as a good example of where deal desk needed to step in and add some rationale. Does that match what you've seen?

Todd Johns: Absolutely. Without any controls, it turns into the wild, wild west — salespeople will sell at whatever price they think it takes to close the deal. You get that wide scatterplot with no real pattern between volume and price. Instituting that discipline around how you think about pricing, and how deal desk can help sales price consistently, is one of the biggest things I've had to build into a company.

Building Pricing Guidance Sales Will Actually Use

Raja Singh: Everybody knows that in B2B, you don't actually pay list price. Does deal desk have a role in proactively communicating a guidance price — what the market price should be?

Todd Johns: Yes. What we did at Salesforce, and what I think is a good approach generally, is take all your deal data, plot it the way you're describing, and think about how you want to position price. You might have a high list price and a certain amount of discounting you consider acceptable, with detail built in around region and product. We built pricing guidance — not necessarily what we were getting in the market, but what we thought we should be getting. So you have your actual price curve, and a target curve of what you think you should be getting. That's the guidance we gave sales: if you're selling this product at this volume, factoring in these variables, here's what we think you should be selling at.

Todd Johns: It's a useful tool at the end of a negotiation, but with more data it can also work as a starting point. Sometimes list prices are so high that reps go in and get laughed out of the room because they don't know where else to start. I had a sales leader tell me, "Todd, my guys are getting laughed out of the room because they're going in at list price — help me figure out where to start." That's where multifaceted guidance you can automate over time is extremely valuable for sellers.

Raja Singh: Was that guidance primarily market-data driven, or did you also factor in ROI and customer-benefit calculations?

Todd Johns: Primarily market data. Being a SaaS company, margins were high across most of the business. As we added products with margins that mattered on a one-off basis — some telecom products where we were using a third party, for example — we built more complexity into the model. But most of the time, market dynamics were the main driver.

Who Gets the Pricing Data — and Who Doesn't

Raja Singh: What did the communication process look like — big calls with the sales team, regional meetings, briefing management? Is there anything you'd withhold from frontline reps and only give to leadership?

Todd Johns: A good example is the price curves themselves. Frontline reps aren't motivated to care about the curve — if you show them every deal ever done, they'll cherry-pick the ones that support what they want to do. There's an argument for giving reps as much transparency as possible, and I understand it, but what we did was focus that data on sales leaders: here's what the market looks like, here's a dashboard showing where your team is pricing relative to the market. For example, the gap between where a given VP's team is pricing versus what the market is bearing, and how many additional licenses that gap would require the team to sell just to make up for it. Getting the data to the right people at the right point in the hierarchy mattered more than blanket transparency.

Raja Singh: How often did you update it?

Todd Johns: It was dynamic — a dashboard you could check any time. But because salespeople are busy, we also built the relationship piece in: quarterly meetings, at minimum, with senior VP-level sales leaders, where an analyst or frontline manager would walk through their deal data — what's working, what isn't, how their leaders are performing on pricing — and then use that same meeting to talk about what's coming in the pipeline.

Todd Johns: One of the things a deal desk needs to do is be proactive, and that's a two-way street with sales. Too often a rep comes in and says, "I need this price, deal desk, just say yes," when if it had come to us a month earlier — or even sooner for a big deal — we could have spent that time on structure and negotiation strategy instead, and the deal would have come out better.

Raja Singh: I've seen a related dynamic: when a rep doesn't have a good feel for market pricing but the customer does, the customer anchors at their preferred number, and the rep ends up internally selling deal desk on a price the customer picked. If reps have the right information earlier, they can anchor the conversation themselves instead.

Todd Johns: Absolutely.

Which Deals Actually Go to Deal Desk?

Raja Singh: Let's get into the operational details. Which deals go to deal desk, and when?

Todd Johns: The answer really depends on the size of the business — it changed a lot over my time at Salesforce as we went from a good-sized company to an enormous one. At a high level: think about the risk in your business, whatever that's measured through — discounting, or risk embedded in certain business terms. However you quantify it, you push the highest-risk slice of deals through deal desk. You can do that through institutional knowledge, where reps just know that an enterprise license agreement or a discount over some threshold needs to go to deal desk, or — preferably — through an automated system that flags it mid-deal. The best reps learn to recognize in advance which deals will need support and start that conversation early, logging the request before it becomes urgent.

Raja Singh: Let me push on that. Say a business is booking ten million dollars a quarter. Is there a rule of thumb — if a deal is some percentage of that number, or deal desk handles the top third of quarterly volume?

Todd Johns: I wouldn't put a specific number on it — that's for each business to decide based on its own risk tolerance. But you're thinking about it the right way: there's a real trade-off, because routing deals through a check-and-balance process, especially a manual one, will slow your business down. You don't want too much going through that gate. At the same time, especially as a smaller company, you're worried about the long-term impact of the pricing you're putting into the market. I don't know the right percentage — but that's the right way to think about the trade-off.

Raja Singh: It sounds like it may be more a function of the deal's impact on your overall book than on the specific quarter.

Todd Johns: Yeah, absolutely.

The Deal Support Request: A Three-Field Template

Raja Singh: Walk through the submission process. Say your criteria is every deal over some dollar threshold. What data does the rep submit, and how does deal desk field it?

Todd Johns: At Salesforce we used our own tools — we built a custom object that let salespeople submit a request linked to their actual opportunity, so the request lived in the context of the deal. We didn't want eighteen thousand pieces of disconnected information. We called it a Deal Support Request, and my goal was to keep it as easy as possible: three fields. What do you want, why do you want it, and what supporting information can you add. I marketed it internally as a one-minute request — you didn't even have to talk to a person unless the analyst decided the deal needed a conversation. Because it was tied to the opportunity and account records, we already had the customer name, industry, and region pulled in automatically. What we needed from the rep was the product, quantity, and price point — or sometimes just the open question, "how should I price this?" — plus anything strategic about the account or any non-standard business terms they were considering.

Raja Singh: More generically — how did that flow interact with the quoting system? Was the quote built before or after the deal desk conversation?

Todd Johns: The ideal state, which we never fully reached, is a deal playground that flows straight through to a quote, approvals included, so you never re-enter the same information twice. In reality, for smaller deals reps would often use the CPQ tool as a playground first, put together a quote, and use that as the basis to start the conversation — a bit of a chicken-and-egg problem. For anything larger, reps would open an opportunity for forecasting purposes, then come to deal desk before finalizing a quote, essentially asking, "here's what I'm looking to do — is this okay?"

From Chatter to Slack

Raja Singh: Before you started, was there a messaging platform built into the workflow?

Todd Johns: I started right as Chatter was taking off — within about a year we had the Super Bowl commercial.

Raja Singh: So you went from no in-context messaging to Chatter, and later picked up Slack. How did those tools change the workflow?

Todd Johns: The ability to communicate in the context of the deal was essential to getting deals done efficiently. Chatter was fantastic for that early on — you could stay off email, stay inside the deal record, write a message, and never lose the history. Slack cranked that up another notch. We built the ability to request deal desk support directly through Slack while I was there, and I expect that kind of integration keeps deepening until Slack becomes the home for how deals actually get worked.

Hiring Deal Desk Analysts

Raja Singh: What did you look for when building the team?

Todd Johns: Early on, problem solvers — we were building our process from scratch, figuring out what worked for our size and growth rate, and there weren't many people with SaaS deal desk experience to hire from. I pulled mostly from consulting and finance backgrounds, which, not coincidentally, is also my own background. We built more diversity into that mix over time — you don't want everyone thinking the same way. As we scaled, I started looking for more directly relevant experience, but you have to define "relevant" carefully. I had plenty of applicants who'd done deal desk work at companies where the whole job was checking whether a spreadsheet showed a three percent margin — not much critical thinking involved. That wasn't what a SaaS business needed. I looked for people who'd actually done the back-and-forth with sales, could communicate clearly, and could use the tools well.

Saying No to Sales Without Losing the Relationship

Raja Singh: Tell me about the times you had to say no, and how you handled the escalations and politics that came with it.

Todd Johns: A lot of it comes down to having relationships at the right level so people know where the conversation stops. I dealt with plenty of reps who would keep pushing until they got the answer they wanted. It comes down to the brand you build across the organization, not just within sales, and having trust with senior executives so that when you say a deal shouldn't happen, they know there's a real reason behind it. You might still come together and decide, for some specific reason, to do it as a one-off. But ninety-nine percent of the time, once you've built that trust and backed your position with data, sales will get on board, even if reluctantly. And it's important never to just say no and hang up the phone — it's, here's why this is a bad idea, and here's what we should do instead. Sometimes that pushes a rep toward a structure that's actually better for them and their customer.

Raja Singh: Like trading a lower discount for additional future-purchase terms.

Todd Johns: Exactly.

SaaS Metrics and the Broader Book of Business

Raja Singh: SaaS has developed very specific metrics — net retention, churn, LTV to CAC. Did deal desk bring those macro metrics into individual deal decisions?

Todd Johns: Definitely. We had a tight partnership with the renewals organization to think about churn, and to think about pricing not just for a single deal but across a customer's whole book of business — how one deal might affect future deals, or deals with other companies in a small, interconnected industry where procurement and executives move between companies and pricing precedent tends to get out despite NDAs. You have to be thoughtful about the ripple effects of any individual deal.

Raja Singh: Would finance come to deal desk and say, this ratio is off, we need more deals that look a certain way?

Todd Johns: I'd see that in pockets. Margin usually wasn't the first concern in SaaS — willingness to pay was the bigger topic — but where margin did matter, you'd see patterns emerge geographically or otherwise: a sales team making its number but at a lower margin, which is dilutive to the broader book, not just regionally but globally. Those were the cases where we looked more closely at the macro numbers.

Working With Legal

Raja Singh: How did you liaise with legal? Some argue deal desk and legal are disconnected — commercial deals get handed off separately. Others argue the terms negotiated in the MSA have real economic impact and need to loop back to the commercial side.

Todd Johns: A tight relationship with legal is essential, because terms that some companies file under "legal" absolutely affect the financials of the deal and the customer relationship. I spent years working closely with legal counterparts, not just deal by deal but at the level of: what risks show up across deals, and who should own each type of risk? If it was fundamentally a business risk, deal desk owned it, but that didn't mean we operated in a silo. The conversation stayed open, and tools like Slack helped everyone see the holistic view of a deal.

Raja Singh: What's an example of something on the fence between legal and deal desk?

Todd Johns: Future-rights language is a good one. Once you're writing contract verbiage — "customer has the future right to do X, Y, or Z" — some people assume you need a lawyer to get the wording right. I think you can have people with both the business judgment around deal structuring and the ability to write that language sitting outside of legal. Anything that's contractually negotiated language about what a customer can or can't do with pricing or products in the future falls into that gray zone; people often default to using lawyers just to be safe on wording.

Fallback Clauses: The "Good, Better, Best" Pyramid

Raja Singh: What about fallback clauses — a preferred clause with pre-approved fallback positions sales can use without looping in legal?

Todd Johns: Internally, yes — the idea of good, better, best.

"I always formulated this as a pyramid: what are the flavors of a given term, and where does each one fall on the risk spectrum? That lets you arm sales to handle the low-risk version on their own. If the conversation turns riskier because of some variable, it escalates to deal desk. If it starts touching precedent-setting language with real legal exposure, it goes up to a dedicated lawyer. It's about arming the largest group — sales — to handle as much as possible at the low end, and routing only the riskier cases to the right expertise as risk grows. It's fundamentally about velocity: sales shouldn't have to go back and forth repeatedly if they already know their guardrails."

— Todd Johns, on the origin of the fallback-clause hierarchy used throughout this series

Product-Specific Terms and the Approval Matrix

Raja Singh: With a large, complex product suite, a lot of product-specific terms come up. How did you manage who the experts were for approving exceptions on an individual product?

Todd Johns: If it was about how the product functions, we leaned on relationships with our product counterparts, sometimes including pricing strategy people who knew the products well. If it was about how the business accessed the product — user counts, where it could be used — deal desk could usually handle that directly. Once you needed deep product knowledge, we relied on those relationships to make sure contractual rights matched actual product functionality; it was too much to hold in any one person's head.

Raja Singh: Did you have a directory or knowledge base of approvers and checkpoints?

Todd Johns: We tried to make it as system-driven as possible. We had what we called the Worldwide Approval Matrix — if you're doing X, Y, or Z, here's where to go, accessible in the context of the deal. It wasn't fully automated; at times you still had to check a knowledge base. Over time, reps just learned the paths by repetition.

Ramping New Analysts and Documenting the Work

Raja Singh: If a new analyst gets a request for a 70% discount, was there a prescribed process to evaluate it?

Todd Johns: Early on, we'd always pair a new analyst with someone more experienced — sometimes their manager, sometimes a peer doing similar work. We also built guardrails, similar to what we did for sales, in something I called the Price Escalation Matrix: categories of decisions analysts could make on their own, and where to go for a second opinion outside those bounds. It was mostly about repetition — get in there, use the guardrails, and know you had ready access to a "buddy" to think things through.

Raja Singh: Were analysts expected to document their reasoning — I approved 70%, here's the analysis and comps I used?

Todd Johns: Always, for a couple of reasons. First, quality control — frontline managers did quarterly reviews checking whether the right process was followed, for both new and experienced analysts. Second, documentation protected against a rep hearing what they wanted to hear on a call instead of what was actually approved — that can go sideways fast. Having a documented, referenceable history mattered both for that deal and the next one with the same customer.

Raja Singh: Did you have formal authority tiers — more senior people handling bigger deals?

Todd Johns: Yes, and it evolved with the company. We eventually built what we called a big deal team — more experienced people for larger, more complex transactions, since those deals are harder and those analysts could be customer-facing with real confidence. At the lower end, it was more about efficiency: as analysts got more reps, we looked for what could be offshored or automated. I used offshoring heavily as a stepping stone toward automation — moving something offshore with the explicit plan to eventually automate it out of an analyst's queue entirely.

Where AI Fits In

Raja Singh: The automation question is interesting to me as a product person. If you can express something as a clean rule, it's fairly easy to automate — the challenge is that most of this is fuzzier, with a lot of if-then logic and gray area. AI potentially handles that fuzziness and more variables at once. Any thoughts on where that goes?

"I think that's exactly the direction we were headed as a business, and I think that's the future of how a deal desk works — AI-driven, and getting more of that knowledge into the hands of the people closest to the deal, the sales reps, earlier, so they don't have to go searching for answers or loop in someone else. I do think there will always be a need for people in this role — it's never going to be a hundred percent automated in my mind, because there's complexity and nuance that won't fully automate away. But there is a massive opportunity for automation in this space that AI is going to help drive, especially for keeping your cost-of-selling and administrative ratios down as the business scales into new regions and languages."

— Todd Johns, on the future of deal desk

Diagnosing Bottlenecks

Raja Singh: Last question — you mentioned metrics like requests supported per analyst, or number of VPs supported per analyst. When one of those ratios looked out of whack, how did you diagnose it?

Todd Johns: You use the data as the canary in the coal mine — it's a flag that something's going on, not a diagnosis by itself. Sometimes it means an analyst is buried because of something happening on the sales side, sometimes it's about their own working style. You dig in with the manager, understand the relationship with the sales team — maybe discounting is unusually hard in a particular region or product line right now, which explains the volume. Or the manager finds something personal, or a breakdown in trust with a sales partner. You use the metric as the flag, then diagnose whether it's market-related, personnel-related, or something else entirely.

A Note From Revolear

Todd Johns is an advisor to Revolear, and conversations like this one are part of why: it gives us standing to sit down with the people who actually built the function at scale and ask the unglamorous operational questions. Revolear builds software for pieces of the mechanics Todd describes above — the discount and approval matrices, the deal support request, contract assembly — for a portfolio of usage-based B2B sales teams today. None of that changes anything he told us here. It's the reason we could ask.

The Takeaway

The through-line in Todd's answers isn't a specific tool or template — it's that deal desk earns its authority the same way any internal function does: by building trust with data, being proactive instead of reactive, and giving sales just enough guardrail to move fast without creating a shotgun-pattern scatter plot of your own pricing. The good, better, best pyramid he describes is the direct ancestor of the Preferred, Fallback, and Approval-Required framework used throughout this series — which tells you it has held up for over a decade for good reason.

Related in this series: this post is part of Revolear's The Deal Desk Handbook. Read more from the series:

What a Deal Desk Actually Does (and Why Most Companies Eventually Build One)

The Deal Desk Glossary (capstone) — read the capstone

Raja Singh, Founder & CEO, Revolear — in conversation with Todd Johns, former Head of Deal Desk, Salesforce.

Sources: Todd Johns, LinkedIn · Revolear