Most RevOps teams built their planning muscles around subscription SaaS: predictable bookings, clean quotas, and a renewal date you could set your watch to. Consumption and usage-based models blow up that playbook entirely. The signals are different, the talent requirements shift, and the forecasting math gets uncomfortable fast.
In a recent RevOps Co-op webinar, Werner Schmidt, CEO and Co-founder of Lative, joined Eddie Parkansky, Director of Sales Strategy and Planning at Intercom, and Sunil Mathew, Global Head of Sales Compensation at LinkedIn, to share hard-won lessons from the trenches. Moderated by RevOps Co-op CEO Matt Volm, the session delivered a practical toolkit for operators navigating this shift — from quota architecture to change management to forecasting in a world where revenue is earned after the signature, not at it.
The move from subscription to consumption isn't happening because someone read a trend report. There are structural forces at work on both sides of the transaction.
Werner Schmidt framed it plainly: there's a growing fatigue — on the customer side — with paying for software that doesn't get used.
"I think there's under-utilization that happened within SaaS. Customers are demanding and feel, hey, there should be more fairness and I should be paying for what I'm consuming." — Werner Schmidt, CEO & Co-founder at Lative
The under-utilization problem is compounded by what Werner described as a misalignment between what software vendors build and what customers actually use. And on top of that, the economics of AI-driven products simply don't map cleanly onto per-seat pricing. When you're paying your LLM provider by the token, selling access to AI products by the license is structurally awkward — and customers know it.
There's also a retention angle. Consumption models, when executed well, tend to produce stronger customer retention. If a customer is actively using your product and deriving value from it, they stay. That's a fundamentally different relationship than a signed contract that sits in a drawer until renewal.
Werner also acknowledged the comp plan angle directly: subscription models rewarded reps for booking seats — whether those seats got used or not. Consumption models force sellers to care about what happens after the signature.
Eddie Parkansky has lived through this transition at Intercom, which moved from a classic SaaS model to a hybrid consumption-based architecture as AI products reshaped its pricing. He named two challenges that organizations consistently underestimate.
The first is change management — and not just the communications kind. The mental model shift required from sellers, finance teams, and even product teams is significant. Under SaaS, a closed deal generates a booking. Under consumption, a closed deal is the beginning of a process that may not generate meaningful revenue for months.
"SaaS, as Sunil said, is a great model. It's predictable, it's simple. And a lot of our industry — sellers, finance folks, pricing folks, product folks, everybody — were used to a SaaS model. And now we're moving to a much more complex, a lot of unknowns." — Eddie Parkansky, Director of Sales Strategy and Planning at Intercom
This has real implications for incentive design. The behaviors that made someone a great SaaS seller — hunting for large bookings, closing fast, moving on — can actually work against a consumption model where adoption and ramp-up are what generate revenue.
The second underestimated challenge is data infrastructure. Subscription businesses can often get away with a BI tool sitting on top of Snowflake, pulling from Salesforce. Consumption requires a more granular layer — usage by product, by account, by rep, by region, and critically, by whether consumption is under contract, an overage, or pay-as-you-go.
"You need an additional data layer, more granular data. Ability to pinpoint usage by product, by account, by rep, by region, by — is it under a contract? Is it overages? Is it as-you-go? So there's a lot of complexity that you need to start building." — Eddie Parkansky
As Eddie noted, even the best-planned transitions will surface edge cases that nobody anticipated. The ability to iterate quickly on that data model matters as much as getting it right the first time. This is also directly relevant to fixing the broader data foundation challenges that RevOps teams carry regardless of pricing model.
One of the sharpest threads in the session was the question of who owns what in a consumption world — and whether existing talent can make the transition.
Sunil Mathew's headline was clear: a seller trained on SaaS needs to become significantly more customer-centric and product-centric to succeed in consumption.
"A person who's traditionally sold SaaS needs to become much more customer centric and product centric. They have to have a much deeper understanding of how the customer uses the product, why they're using the product, what they can do more with the product they're not doing." — Sunil Mathew, Global Head of Sales Compensation at LinkedIn
Sunil's recommended approach: run a pilot before rolling out the new model to the full organization. This gives you champions, a proof of concept, and the opportunity to validate your assumptions before they become someone's quota. He also stressed the value of having frontline leaders who have actually sold consumption-based products — not just executives who understand the theory.
On the question of who owns post-close adoption, Sunil was unequivocal: the rep that acquires a logo needs to keep that logo through implementation and early consumption ramp before any handoff happens. This changes the hunter-farmer dynamic meaningfully, because hunters are now partially accountable for time-to-value, not just time-to-close.
The incentive architecture follows from the role architecture. Hunters might be compensated on new logos plus consumption ramp milestones. CSMs might be measured on pure consumption or expansion metrics. The key is getting the role definition right first, then building the comp plan around the behaviors those roles need to exhibit. For more on building compensation structures that actually drive the right behaviors, this deep dive on comp plan design covers many of the same tensions from a different angle.
Under subscription SaaS, quota math is relatively straightforward: take your revenue target, divide by the number of sellers (with appropriate over-assignment), and you have a framework. Consumption breaks that model.
Werner described the shift: you're no longer planning against bookings. You're planning against adoption. And adoption introduces variables that don't exist in a subscription world — organic consumption growth, account ramp timelines, what happens to a book of business when a rep leaves mid-year.
At Intercom, the response was to bring in data science — for the first time — as a core part of the planning process.
"For the first time when we moved to a more usage-based consumption base, we started to leverage our data science team, where something that we haven't used heavily before for setting targets and quotas. And we had data science that worked very closely with us for the entire planning process — a very granular model based on more than 50 different variables." — Eddie Parkansky
The model runs experiments against historical results to validate its predictions, accounts for seasonality, and is treated as a live system that gets updated after every planning period with new variables and new insights. Crucially, it distinguishes between contracted amounts (which are knowable), overage patterns (which can be modeled), and pure unknowns (which require assumption-based ranges). Lative's platform is part of how Intercom operationalizes this — replacing the spreadsheets that simply can't handle the complexity.
Sunil added a behavioral dimension to quota setting that often gets overlooked: when launching into a new model, err on the side of achievable targets.
"I generally like setting more achievable targets. So if our model or plan said we can do maybe a hundred thousand per rep per month, I'd prefer to set that at 80,000, 85,000, 90,000 so that we have that culture of success in our locker room." — Sunil Mathew
The reasoning isn't about being soft on targets. It's about keeping reps in a growth mindset during the transition — always looking for ways to do better — rather than a deficit mindset where they've already convinced themselves the new model is unwinnable.
If quota setting is hard, forecasting is harder. Consumption revenue is, by definition, less predictable than subscription revenue. One of the honest admissions from this session was that you have to accept some of that unpredictability as part of the trade-off — and communicate it clearly to stakeholders up the chain.
But there are meaningful ways to improve the precision of your forecast. Eddie's framework: break the forecast into what you know versus what you're estimating. Contracted amounts are knowable. Seasonal patterns for known customer types are modelable (tax software spikes in March and April; retail customers spike around major holidays). From there, you layer in rep-provided signals through tools like Gong or Clari, and you reconcile those against the data science model's output.
The result isn't perfect predictability — but it narrows the range between best case and worst case. That's the goal. This mirrors the broader challenge of building accurate sales forecasts in Salesforce, where the discipline of assumption-tracking makes the difference between forecasts that inform decisions and ones that don't.
Werner emphasized the need to close the loop between planning and execution in ways that most organizations simply don't do today:
"We would tend to do plans and then we leave the plan and then we go into our execution. We don't ever go back and look at the plan to say, how did that actually track? With consumption, we really do need to know that because we need to make sure we understand the execution against where we said we expect to be." — Werner Schmidt
This requires real-time visibility into account-level consumption — tracking month-over-month trends, flagging when accounts are drifting from projected ramp, and building operating cadences around that data. Excel and Google Sheets are not up to this task. Lative's approach to connecting sales planning to execution is designed specifically for this gap — making the plan a living document rather than a quarterly artifact.
Sunil laid out a three-part framework for tracking performance in a consumption model:
Build a granular plan with assumption-level tracking. Your forecast needs to go below segment and region — down to buckets of customers with similar seasonality profiles. Every assumption in the model should have a corresponding tracking mechanism.
Invest in real-time, shared visibility. The operating model needs dashboards that everyone can see, with triggers built in for when consumption is trending off-plan. Without this, you're flying blind between quarterly reviews.
Stand up the operating model before day one. Don't wait until after launch to decide how you'll review performance. Define the cadence — what you'll look at, who will be in the room, how quickly you can course-correct — before you go live.
He also recommended starting with quarterly planning cycles rather than semi-annual or annual, at least during the transition period. The faster feedback loop allows you to test whether your assumptions are actually holding up in the market.
Both Eddie and Sunil, independently, gave the same piece of advice when asked what they'd tell their pre-transition selves: you will underestimate how much change management this requires. Pay the tax willingly — and then pay more.
Eddie's specific recommendations:
Sunil added the co-creation angle: in his second transition, he spent a month meeting with frontline sales leaders four times a week to co-design the new go-to-market motion with them. The result was a model that was both better designed and had built-in advocates.
Werner tied it back to the knowledge-sharing dimension that's easy to miss: in a consumption model, what works for one rep on one account is genuinely useful to other reps. Building systems for that knowledge to flow matters more than in a SaaS environment where winning a deal is largely an individual act. This connects to broader principles around sales enablement that actually gets used — the design question isn't just what to teach, but how to make the right behavior visible and repeatable.
"What did I do to help increase consumption where I saw it go up? That could help another salesperson. Sharing that knowledge is absolutely key in consumption models." — Werner Schmidt
The closing question was deliberately provocative: is this a permanent shift, or will we look back in five years and see that SaaS subscription came back into fashion the way '90s fashion is currently reappearing on high schoolers?
The panel's consensus: it's not going away, but it's also not replacing everything.
Werner's take was that consumption will exist alongside subscription — the determining factor is whether the product and billing infrastructure can actually support a consumption model. If it can't, don't force it. Eddie framed it as a pendulum swing: the market has moved deep into consumption, but the desire for predictability will pull some of that back toward committed contracts and hybrid structures. Sunil pointed to economics as the long-run driver — if AI inference costs come down significantly, the unit economics of consumption-based AI pricing may start to look more like subscription again.
The more durable insight is that the operators who build the planning, forecasting, and incentive infrastructure to handle consumption will be ahead regardless of which direction the pendulum swings next.
The message from this panel is direct: the shift to consumption isn't a trend to wait out. It's an operational discipline to build — and the teams that invest in the planning infrastructure now will have a structural advantage when the model is fully mainstream.
For more on how Lative approaches sales planning and quota-setting for modern revenue teams, explore their resources for consumption-model operators.
Check out our blog, join our community and subscribe to our YouTube Channel for more insights.
And be sure to check out Lative for more resources on AI-powered sales planning for consumption and usage-based models.
