← All essays

The comp plan change that fixed our win rate

One structural tweak to quota and accelerators that realigned the whole team in a quarter.

We fixed a sliding win rate by changing one thing, and it was not the playbook, the enablement, or the manager scorecard. It was the comp plan. We re-pointed accelerators and quota credit at the deals we actually wanted, and the team realigned inside a quarter.

If your numbers are drifting the wrong way while activity looks healthy, the problem is probably not your reps. It is what you are paying them to do.

The symptom: a busy team and a falling win rate

On paper, everything looked fine. Activity was up. Pipeline coverage was north of 3x. Reps were hitting their meeting numbers and their outbound numbers. The leading indicators were green.

The lagging indicator was not. Win rate had slipped several points over a couple of quarters, and average deal cycles were getting longer. We were generating more pipeline and closing a smaller share of it.

When I pulled the deals apart, the pattern was obvious. Reps were spraying. They were creating single-threaded opportunities off one lukewarm contact, marking them qualified, and parking them in the funnel to pad coverage. A lot of these were not real. Some were sandbagged into next quarter. The forecast looked thick and weighed nothing.

None of this was malicious. Every rep was doing exactly what the plan rewarded. We paid for bookings volume and we paid for activity, so we got volume and activity. Quality was nobody's incentive.

The change: pay for the deal you want, not the deal you can stuff

We made one structural change to the plan. We did not touch base salary or on-target earnings. We re-pointed where the money concentrated.

Three moves, all in the same direction:

• Gated the accelerator on deal quality, not just attainment. Reps earned the over-quota accelerator only on deals that met a tighter bar: multi-threaded (more than one real stakeholder), with a documented business case, and net-new logo or genuine expansion. Single-threaded one-call closes still paid the base commission. They just did not reach the rich part of the curve.
• Made quota credit follow qualified pipeline, not raw creation. We stopped celebrating opportunity count. Pipeline that did not clear a real qualification stage carried no credit and no recognition. Creating junk stopped being a way to look productive.
• Put a small kicker on win-rate quality. A modest bonus for reps who stayed above a clean win-rate threshold across the quarter. Not huge. Enough to make closing well, not just closing often, worth something.

That is it. We did not rebuild the motion or rewrite the methodology. We moved the money to sit on top of the behavior we wanted and removed the money sitting on top of the behavior we did not.

Why it worked: comp is the real strategy doc

You can write the strategy on a slide, repeat it in every QBR, and pin it to the wall. Reps will still read the comp plan instead. The plan is the only strategy document they trust, because it is the one that pays them.

Tell a rep to focus on quality while you pay them for volume, and they will hear the pay. Every time. That is not a discipline problem. It is rational behavior responding to the actual signal.

So when we moved the incentive, the behavior moved with it, fast. Within a few weeks reps were qualifying harder before creating an opportunity, because a weak opp no longer helped them. They were multi-threading earlier, because the accelerator depended on it. The junk pipeline thinned out on its own. Nobody had to police it.

By the end of the quarter, win rate had moved up about 6 points and the forecast got more honest, because there was less garbage in it to begin with. Total pipeline created actually went down. That was the point. We traded volume we were never going to close for a funnel that converted.

Reps optimize exactly what you pay for, down to the decimal. If the output is wrong, look at the incentive before you look at the person.

The tradeoffs: every plan creates a new distortion

There is no comp plan that only produces good behavior. You are always choosing which distortion you can live with, so go in clear-eyed.

Here is what we had to watch:

• Coverage anxiety. When you stop paying for raw pipeline, total coverage drops, and that spooks people who grew up reading 3x as safety. You have to be willing to defend a thinner, truer number to your own board.
• Gaming the quality bar. Define "multi-threaded" and reps will find the cheapest way to technically clear it. Two contacts, one of them a junior nobody. We had to keep the qualification definition tight and audit a sample of deals, not just trust the stage field.
• Punishing legitimate fast deals. Some genuinely good deals are single-threaded and quick. We kept base commission whole on those so we were never penalizing a clean win. The gate shaped the accelerator, not the floor.
• Plan complexity. Every gate you add is one more thing a rep has to model in their head before they act. Past a point, complexity kills the behavior change, because nobody can compute it. Keep it to a few rules a rep can hold in their head between calls.

The discipline is to pick one quality behavior that matters most right now and pay for that. Not five. One plan cannot fix everything, and a plan that tries fixes nothing.

Why this matters more in the AI era

This used to be a slow-moving problem. A rep could only spray so much junk by hand. That ceiling is gone.

When an AI agent sits on the top of your funnel, it can manufacture activity and create pipeline at a volume no human team ever could. If your comp plan still pays for activity and raw creation, you have now wired a firehose to the exact incentive that was already distorting behavior. You will drown in qualified-looking pipeline that converts at nothing, and the metrics will look better than ever while the business gets worse.

The defense is not in the tooling. It is in the plan. When volume becomes nearly free to produce, the only thing worth paying for is quality and outcomes. Pay for the deal you actually want closed, and it stops mattering how much noise the top of the funnel can generate.

Before you add an agent to your pipeline, read your comp plan as if you were a rep being paid by it. If the honest answer is "this pays me to create volume," fix that first. Otherwise you are about to scale the wrong behavior at machine speed.

The cheapest, fastest lever you have on win rate is not a new tool or a new process. It is moving where the money sits. Reps will follow it there. They always do.