Chained referral: how one order pays more than one person

How does a chained referral pay out?

In a single-level program, Asha refers Bhavna, Bhavna buys, Asha gets a reward. That is the entire transaction, and it produces one ledger entry against one account.

In a chained program the same purchase also pays whoever referred Asha, and whoever referred that person, until the configured depth runs out. One referred order can therefore create three separate reward records against three separate people who have never met.

Reward rates almost always decrease as the chain goes up. A common shape is 10% of order value to the direct referrer, 5% to the level above, and 2% above that, on the reasoning that influence weakens with distance from the sale.

The important structural point is that the chain is a property of the customer graph, not of the order. The same buyer placing the same basket produces a different payout depending on how they arrived at the brand months earlier, which means chain cost is a function of history rather than of cart contents.

That also makes chained payouts harder to reverse. A refund has to unwind rewards across several accounts, some of which may already have spent the credit, which is why the ledger design underneath matters more here than in a single-level program.

What does each additional level actually cost?

The headline cost of a three-level chain at 10, 5 and 2 percent is 17% of order value, against 10% for the same program run at a single level. Stated that way, most founders stop at one level and move on.

The headline number is misleading because levels only pay when somebody is sitting in them. A customer who arrived through a paid ad has nobody above them, so the second and third levels never fire on their referred orders.

Model it explicitly. On stated assumptions of a 40% independent probability that each upstream level exists, the second level fires on about 40% of referred orders and the third on about 16%.

Level	Reward rate	Probability populated	Blended cost
1	10%	100%	10.0%
2	5%	40%	2.0%
3	2%	16%	0.3%
Total	17%		12.3%

Blended cost lands near 12% rather than 17%, and whether that survives contact with your gross margin is the question worked through in the referral maths that decides whether you make or lose money.

Depth is a behavioural question before it is a financial one. A third level paying someone ₹18 on a ₹900 order does not change that person’s behaviour, which means you are carrying accounting complexity for no return. Where the useful ceiling actually sits is argued in how deep a referral chain should go.

How is a chained referral different from multi-level marketing?

Two things separate them, and both are about what triggers money moving. A chained referral pays only when a real product is sold to a real customer at a real price. Recruitment on its own pays nothing.

MLM structures typically reward recruitment itself, often require the recruit to purchase inventory or pay a joining fee, and pay out in cash. Remove the product sale trigger and add a recruitment payment, and the structure changes category entirely.

The reward currency reinforces the distinction. Capped store credit that can only be spent on the brand’s own catalogue is a discount mechanism, not an income stream, and nobody joins expecting to earn a living from it.

None of that makes the boundary automatic or self-enforcing. Chain depth, how rewards are described in marketing, and whether anyone is encouraged to recruit rather than recommend all move a program towards the line, and the specifics for India are set out in the legal line on purchase-gated referrals.

A live three-level payout, including what each participant actually sees, is walked through in what happens when one order pays three people.

Where SB&R fits

SB&R is a Shopify app for chained referral rewards. Every referral link belongs to someone who has already bought. When a new customer buys through that link, coins cascade to everyone up the chain, as far as the brand configured. Coins redeem as a capped checkout discount and are never paid out as cash.

Payout logic runs inside Postgres functions rather than application code, so an order interrupted midway through a cascade leaves either a complete result across every level or nothing at all. Partial chains, where two of three people got paid and nobody can say why, are the specific failure that design exists to prevent.

SB&R is not the right tool for cash-payout affiliate programs or for B2B and wholesale referral. If the reward has to leave the business as money, or the referrer is a purchasing manager rather than a customer, a chained store-credit model is the wrong instrument regardless of how the cascade is configured.