Seller Incentives: Moving to Commissionable Revenue

1Problem statement

Our current incentive structure asks one metric to do too many jobs. It mixes what the seller controls (selling) with what they don't (lead mix, supply economics, discount capacity). Three problems follow:

1. Supply gains leak into discounts. When backend contracting improves, the improvement becomes visible discount room for the seller instead of margin for the company.
2. Conversion as a payroll metric creates volatility and gaming, not better selling. It punishes lead mix and month noise, and rewards subjective stage movement.
3. A common target across markets and tenures is structurally unfair. Europe and Bali are different businesses. An M2 seller and an M13 seller are different sellers.

We went to the data to answer one question first: what actually drives conversion — pricing or selling?

2What the data says

2.1 By tenure — experience compounds

SQL% actually falls with tenure (34% at M1 to 27% at M12+) — senior sellers are not getting better leads. Yet SCC% rises from 7.3% at M1 to over 12% by M8+. Same or slightly worse lead quality, materially better conversion. And M12+ sellers carry the lowest discount (4.3% vs 5.1% at M1) while converting the best.

2.2 By achievement — top performers sell better, they don't get better leads

SQL% is almost flat across achievement buckets (28–32%). The lead pipe is roughly the same for everyone. But once a seller believes "this is a genuine customer," the gap becomes enormous: GLOCC is 37% for 100%+ achievers vs 23% for the 50–60% bucket; SCC is 16.3% vs 8.7%. Nearly 2x on the same quality of input.

Discount shows no clean gradient across buckets (3.8%–5.2%, no pattern).

2.3 By discount — discounting does not buy conversion

SCC is 10.9% in the 0–2% discount bucket and falls to 9.3% at 8%+ discount. There is no correlation where more discount equals better conversion. If anything, the heaviest discounting sits with weaker outcomes.

3Key takeaways

• Better sellers create better outcomes through selling, not pricing. Skill → Conversion → Revenue → Achievement.
• Top performers don't get better leads. SQL% is nearly flat (28–32%) across achievement buckets. The lead pipe is the same; the selling isn't.
• Tenure compounds — on organic, not paid. SCC rises from 7.3% at M1 to 12%+ by M8, with the lowest discount. But POCC sits in a narrow 1.5–1.9% band throughout: paid conversion is very similar M1 to M4, and M12+ is actually the lowest at 1.5% vs the 1.9% peak at M8–M12. Tenure's advantage shows up in qualified and organic conversion — paid conversion is a lead-quality and speed problem, not a tenure problem.
• Discounting doesn't buy conversion. SCC falls from 10.9% (0–2% discount) to 9.3% (8%+). Price competitiveness matters; that's Supply's job.
• Sales' core KPI is conversion. Revenue is the payout metric, but conversion is what the target encodes — a seller cannot hit the revenue target without converting at the benchmark.
• How the target is set: tenure-based conversion benchmark × market average ticket size × market GM% (the fixed margin credit) = the commissionable revenue target — and therefore the commissions. The seller sees one number; the conversion expectation is inside it.
• Conversion must improve with tenure — and the data shows it does. So the fixed component improves with tenure, linked to demonstrated conversion. Tenure earns a higher fixed salary and a higher target; the variable stays on revenue. The earnings ladder: prove conversion → higher fixed → higher target → higher upside.

4Why this change

If selling skill is the highest-variance input we control, the incentive should pay for the revenue a seller creates — and nothing else.

1. Sellers earn only on value allocated to Sales. Marketing creates opportunity. Supply creates competitiveness. Sales converts and monetises. Paying sellers on live GM lets Supply improvements leak out as discounts.
2. Discount should not be an earning lever — because the data says it isn't a converting lever.
3. Targets, not commission rates, absorb market and tenure differences. Same rules for everyone; fair targets for each.
4. Conversion belongs in coaching and target-setting, not payroll. A seller cannot hit a revenue target without converting — it's already embedded.

5What we are doing

A single-metric incentive: commissionable revenue.

6Open decisions & path to launch

Decisions to close

• Which earning structure — floor or tiers? The ₹2.8L floor is unconfirmed, and the 80/90/100 tiered alternative changes payout distribution materially. Back-test both against 6–12 months of seller-month data (% earning zero, % crossing each tier, top-decile payout, variable cost as % of commissionable revenue) and pick one before the pilot.
• M1 economics vs the retention work. A ramped M1–M3 target with any floor means a new joiner earns nothing for months unless lead allocation is fixed first. If the new-joiner lead-starvation fix isn't live before this launches, this compounds early attrition rather than fixing it. Sequencing is the real decision.
• Europe and the discount lever. The flat discount curve carries a selection effect — sellers may discount because a deal is weak, not the other way around — and Europe hinted at a real relationship. Watch Europe conversion specifically once the discount pen is taken away.
• Checkout lag for long-haul. Europe books months ahead; Bali doesn't. Checkout-based payout pays long-haul sellers materially later for the same work. Does delayed cash change behaviour or attrition in long-haul pods?

Path to launch

1. Back-test both earning structures; confirm the floor / tier rates.
2. Freeze margin credits (Bali 15% / Maldives 12% / Europe 20%); pick the pilot pair — one short-haul, one long-haul.
3. Run one month of shadow payouts with dashboards; test seller comprehension.
4. Go live for two months. Readout cut by lead source and by market — blended conversion will mask the organic gain.
5. Refine and scale.