A good SDR commission plan in 2026 pays 70% fixed and 30% variable, puts most of the variable on meetings held and accepted by the account executive, adds a small component on pipeline or closed-won, and uses one accelerator above target. It is simple enough that a rep can compute next month's pay in under a minute, and it is written down before the period starts.

Here is how we would build one, with the ratios, the units, three grids and the numbers from our 1,000,000-call benchmark that tell you what an SDR can realistically book.

Start with the ratio: why 70/30 is the European default

Public benchmarks agree on the shape. The Bridge Group's 2025 SDR report puts the US split at 68:32 (median OTE about $80k). In France, Germany, the Netherlands and Spain, job listings and salary guides converge on 70/30; UK and SMB teams drift to 60/40; enterprise teams go to 80/20.

The logic is simple. SDRs control activity and qualification. They do not control pricing, product, the AE's closing skill or the buyer's budget cycle. Put more than 35% of their pay on outcomes they do not control and you get stress, churn and gaming, not pipeline.

In France, there is a legal floor under this logic. Variable objectives must be communicated at the start of the period and be achievable, or the variable is fully due (Cour de cassation, 2011, confirmed 2024). And you cannot shift pay from fixed to variable without the rep's written consent.

Pick the unit: booked, held, accepted, closed

The unit you pay on is the behaviour you will get. Choose with that in mind.

• Meetings booked. Easiest to count, worst to pay on. It buys you calendar invites, no-shows and AE complaints. Use it only in the first 30 days of ramp.
• Meetings held. The meeting happened. This should carry 50-60% of the variable. It aligns the SDR with the confirmation and reminder work that lifts show rate.
• Opportunities accepted by the AE. The AE confirms it fits the ICP and has a next step. 25-40% of the variable. This is the quality check.
• Closed-won or pipeline value. A small kicker, 10-15%, so the SDR cares what happens after the handoff without betting their rent on a six-month cycle.

Two or three units, never five. The fastest test of a plan is to ask a rep what they earned last month and watch how long they take to answer.

Size the quota with real conversation math

Most SDR quotas are inherited numbers nobody can explain. Build yours from the funnel instead.

Our analysis of 1,000,000 B2B cold calls in France gives the inputs. Average connect rate: 11.3%, up to 15.8% in the 4pm slot. Conversation-to-meeting rate: 2-3% with pitch-first openers, 7-10% with trigger-based openers. An SDR on a parallel dialer holds about 20 conversations a day; a hand-dialing SDR holds 6-8.

• Hand-dialing, pitch-first: 7 conversations x 20 days x 3% = roughly 4 meetings a month.
• Parallel dialer, pitch-first: 20 x 20 x 3% = roughly 12 meetings a month.
• Parallel dialer, trigger openers on CRM signals: 20 x 20 x 7% = roughly 28 meetings booked a month, before no-shows.

Apply a realistic show rate (in our experience 65-85% depending on how fast the meeting is scheduled) and a realistic AE acceptance rate, and you get a held-and-accepted quota of roughly 8-15 a month for a ramped rep with proper tooling. A quota of 15 on a hand-dialing team is a lottery ticket, not a target.

Three grids you can copy

All three assume 70/30, paid monthly, objectives written and signed before the period.

• Seed-stage grid (1-3 SDRs, no AE bench). Variable 100% on meetings held, flat amount per meeting, quota 8-10 a month. One accelerator: from the 11th meeting, the per-meeting amount is multiplied by 1.5. No clawback; a meeting that does not happen is simply not paid.
• Scale-up grid (4-15 SDRs, dedicated AEs). 55% on meetings held, 35% on opportunities accepted, 10% on closed-won from SDR-sourced meetings in the quarter. Threshold at 70% of quota before the variable starts paying, then linear, then 1.5x above 100%, uncapped. Quota 10-15 held-and-accepted a month by segment.
• Enterprise grid (long cycles, 80/20 ratio). 40% on opportunities accepted, 40% on qualified pipeline value created, 20% on a quarterly quality score (ICP fit, next step booked, notes complete). Quota 4-6 opportunities a month. Paid quarterly to smooth the lumpiness.

Accelerators, thresholds and no-shows

An accelerator is the signal that you want top performers to keep going in the last week of the month rather than sandbag into the next one. 1.5x above 100% is enough; 2x makes finance nervous and rarely changes behaviour.

A threshold (nothing paid below 60-70% of quota) protects the budget from reps who coast. It also protects the plan from being read as a per-meeting bounty.

No-shows: pay on held, not booked, and give the rep a 14-day rescheduling window. Then there is nothing to claw back. Reward show rate explicitly in the quality component if it is a recurring problem. The reps who confirm by phone the day before have higher show rates than the reps who rely on the calendar invite, and the plan should notice.

How Sonik approaches commission plans

We do not design comp plans for customers, but most Sonik teams run a version of the scale-up grid above, and the dialer data is what makes the quota defensible. Managers see conversations per day, connect rate per persona and slot, meeting rate per opener and show rate per rep, so the quota is built from the team's own funnel rather than a number inherited from the last VP.

The parallel dialer is what makes 20 conversations a day routine, and the CRM signals (website visitors, job changes, closed-lost, dormant leads) are what push the meeting rate from the 2-3% band to the 7-10% band. Half of Sonik's own pipeline comes from signals that were already in the CRM. In plan terms: the same 30% variable is reachable on Sonik and theoretical on a softphone. See how signals work.

Conclusion

A 2026 SDR plan is 70/30, paid on meetings held and accepted with a small outcome kicker, one 1.5x accelerator, no clawback, written before January. The quota comes from your own conversation math, not from a template. This week, compute what your median rep actually earned last quarter against OTE. If it is under 70%, the plan is the problem, not the rep.