Webinar Details

Do we even need variable compensation anymore? In a world of PLG, AI SDRs, and hyper-skeptical buyers, it might be time to rethink how GTM comp actually works ...

Session Highlights

We started with a simple question, do we even need variable compensation anymore? Then we spent an hour proving that question is way less ridiculous than it sounds.

This article is my attempt to codify what we talked through live. The mess, the tradeoffs, and where I’m personally starting to land on GTM comp in a world of PLG, AI SDRs, and hyper-skeptical buyers.

Nobody likes their comp plan

I’ve never worked anywhere that people genuinely loved the comp model.

“At least everybody complains about it equally, right? Like salespeople don't like it, the CFO doesn't like it. Like everybody kind of doesn't like what they have, but they live with it because it's sort of like the best that anybody can accept.”

That’s… insane when you think about it.

Compensation is the primary lever we use to drive behavior in go-to-market roles, and yet:

• Reps think it’s unfair or unachievable

• SEs think it under-values their contribution

• Finance thinks it’s too generous or too unpredictable

• Leadership is constantly fiddling with it like a broken thermostat

So maybe the right question isn’t “How do we tweak the comp plan this year?” but:

Is the whole variable-comp-as-default model past its sell-by date?

Variable comp: sacred cow or legacy artifact?

Sales is one of the last true “pay-for-performance” roles in tech.

We tell sellers: “You take more risk, you get more upside.” We tell SEs: “You’re crucial, but a little less risky, so here’s 70/30 instead of 50/50.” We tell everyone else: “Here’s your base salary and maybe a bonus. We’ll figure it out at promo time.”

Over time, I’ve gotten less religious about variable comp.

On the webinar I said:

“Quota is like a made up number… you’re kind of like finger to the wind going, okay, you know, we can generate this many leads… and the average seller should hit this.”

We pretend this is an exact science. It’s not. It’s a forecast, politics, optimism, and vibes.

And the randomness is real. You can have:

• A great rep with bad luck

• A bad rep with great luck

• Or somebody who happened to be sitting on the right patch of territory when a seven-figure deal wandered in.

As I put it during the session:

“If you have like a thousand people together and you have like a paper, scissors, rocks tournament… somebody has to win. That doesn’t mean they're the best in the world.”

Yet our comp plans treat that person as objectively “the best.”

The SE problem: three systems, zero fairness

I’ve built and led solutions teams at Cloudinary, Moveworks, Copy.ai, Adobe, Sitan, and others. Over and over, I’ve seen the same pattern: SEs are critical to the sale, under-recognized in comp, and structurally misaligned.

I joked that there are three SE compensation “political systems”:

The Communist Model All SEs share a global/team number. Top performers carry the team but get paid the same as the folks cruising.
The Socialist Model Half your variable tied to your own deals, half to team/global. Gives room for collaboration and backup coverage, but still rewards impact.
The Capitalist Model You carry a number like an AE. Maximum individual accountability; minimum safety net.

Personally, I’ve always disliked the pure “communist” version:

“You have like top performers… they're carrying the team and they're making the same amount. I didn't like that.”

It’s how you burn out your best SEs and eventually lose them.

Also: why are SEs 70/30 or 80/20 when AEs are 50/50 lone wolves?

“Why would SCS be like 80/20, 70/30, and then salespeople are 50/50? And then one team works on this communist style and the other team is just lone wolf capitalist. How did that get started at all?”

Nobody ever gives a good answer. “Because that’s how it was when I got here” is not a strategy.

Compensation dictates behavior (whether you like it or not)

This is the one thing I feel extremely strongly about:

Compensation always dictates behavior.

We talked about a real example from a prior company:

Reps were only compensated on the initial land, not upsells.
The product naturally expanded over time once adopted.
But comp drove reps to maximize the first deal at all costs.

As I explained:

“If I have four deals and I'm gonna go for the $75K, I'm gonna try to 3x that to 75,000. Even if that decreases my win rate by 50%, I'm still making more money. But that's way worse for the business.”

Then you get churn, misfit customers, strained implementations, and “WTF is this?” handoffs to delivery.

That’s not bad people. That’s bad systems.

We also touched on another classic behavior leak: end-of-quarter desperation. If someone’s mortgage payment depends on closing a certain number, they will:

Over-discount
Over-promise
Push bad-fit deals over the line

That’s not a moral failing. That’s rational behavior in a poorly designed game.

Spiffs and micro-incentives: way more powerful than we admit

One area I do get excited about is behavioral spiffs.

We talked through a bunch of them from past companies:

Bonuses for press releases with customers
Bonuses for logo usage and case studies
Big bonuses for getting a badge to enter the client’s building (a huge trust marker)
Bonuses when customers speak on stage on your behalf

I framed it this way:

“I'm all for like interesting… one-off things that you can directly align to a behavior that you want. That’s way easier than trying to finagle some super complex math equation.”

Instead of saying “Here’s 12% of ARR, go do whatever,” we can say:

Get the press release → $X
Get the referenceable logo → $Y
Get them live and happy → $Z

These behaviors:

Make renewal and expansion dramatically easier
Increase marketing efficiency
Make fundraising easier (nothing like a steady drumbeat of logos and PR)

Yes, CFOs get nervous when a $25K deal pays out $12K in spiffs. But if that deal plus PR plus logo plus conference talk unlocks 10 more deals, the math is actually fantastic.

SDRs in the age of AI and PLG

I feel especially strongly that traditional SDR comp is broken.

Paying purely on meetings booked made sense in the “spray and pray” era. But when AI can send a thousand cold emails in an afternoon, the value of a human SDR is not “number of meetings.”

The value is:

Knowing who is actually worth going after
Figuring out how to break through the noise
Crafting sequences that resonate with real pain
Doing the in-person, high-intent things: conferences, meetups, actual conversations

That’s why, on the webinar, I said I feel even stronger now about tying SDR comp to closed revenue or at least qualified pipeline, not just meetings.

But there are practical problems:

Long sales cycles mean “money later” feels imaginary
SDR roles turn over quickly; many SDRs won’t be around when the opportunity closes

So I like hybrid structures:

Some comp on qualified opps (not just meetings)
Some kicker on closed-won so they care about quality
And clear feedback loops: “Deals that started from this motion close. These other motions just create noise.”

In a world where AI handles “busy work,” SDRs have to become strategic revenue creators. Their comp should reflect that.

PLG + sales: you can’t just bolt on an enterprise quota

We closed the session talking about PLG-heavy companies that want to “layer in enterprise sales.”

This is where comp goes from hard to dangerous.

If you’re doing, say, $40M in PLG revenue and want to see if there’s “enterprise upside,” you cannot just hire a few enterprise reps, hand them $1.5M quotas pulled from thin air, and expect it to work.

I said I’d start differently:

“I would probably either go hire two enterprise sellers or, more likely, take two of your existing sellers and promote them to an enterprise team… and give them time and a more experimental, entrepreneurial comp structure.”

Key points:

First, define what “enterprise” actually means (not “5K from JP Morgan on a credit card”).
Second, treat it as an experiment, not an immediately scalable machine.
Third, expect product gaps: if there’s nothing they can sell beyond the swipe-card PLG offering, no comp plan will save you.

PLG gives you a huge advantage: usage data, champions, in-product behavior. But if your comp model doesn’t reflect land-and-expand reality, you’ll train reps to chase monster one-shot deals rather than sustainable expansion.

So… should we kill variable comp?

I started this topic as a bit of a strawman. By the end of the conversation, I caught myself saying:

“I'm honestly not sure where I land… I started off as like, I brought this up almost like a straw man argument. Now I'm like, maybe this is how it should be done.”

Here’s where I am today:

I’m not ready to say “no variable comp ever again.”
I am ready to say we massively over-index on it and under-use other tools.
I think we dramatically underestimate:
The role of randomness
The damage of misaligned plans
The power of simple, fair, transparent systems

And I am absolutely convinced of this:

If your comp plan is a complex math puzzle nobody can calculate at a red light, it’s probably driving the wrong behavior.

A few practical principles I’d use if I were starting from scratch

If I were building a GTM org at a 50-person Series A company today, I’d bias toward:

Simplicity over cleverness
1. Plans people understand → plans people trust → behavior you can predict.
More base, targeted variable
1. Especially for SEs and SDRs. Reserve meaningful upside for clear, high-leverage behaviors, not vague “maybe this will drive ARR” assumptions.
Spiffs for business-critical behaviors
1. Customer stories, PR, references, go-lives, speaking slots, expansions.
2. Make those worth real money.
SEs treated as true revenue partners
1. No more “SEs carry everyone but get communist comp.”
2. At least partially deal-tied, with some team element for coverage.
SDRs measured on impact, not spam
1. Qualified opps, closed-won attribution, and learning loops.
2. Assume AI will handle the grunt work.
Experimental mindset for new motions
1. Especially for PLG → enterprise transitions.
2. Don’t pretend you can define a perfect 50/50, $1.5M quota plan on day one.