Coupon and Cashback Placement with Incrementality Terms

Start with the invoice, not the partner list. The first test we run on any coupon or cashback relationship: take every conversion the partner claimed last quarter and measure the time between their click and the order. Where that gap is under two minutes and the click landed on a cart or checkout URL, the buyer was already buying and a browser extension took the last click on the way through. Across the accounts we have opened, that pattern accounts for fifteen to forty percent of the invoice. Nobody should agree a placement fee before knowing which end of that range they sit at. Partners we place with in this region: ShopBack across Singapore, Malaysia, Taiwan and the Philippines. Fave where redemption happens in a physical store. iPrice and Picodi for code discovery. Shopee and Lazada in-app voucher inventory, where the buyer never leaves the marketplace and attribution is the platform's rather than yours. Global Savings Group properties when the market in question is Europe. For a dealer group or a distributor the useful surface is narrower than the category suggests: service, parts and accessory commerce, and the finance-adjacent cashback offers. Not the vehicle. How we structure the terms: Unique non-public codes per partner, so leakage becomes measurable instead of arguable. Two commission rates, new customer and returning, priced off your actual repeat rate rather than a category benchmark. A deduplication clause in the insertion order that names the interception pattern above and pays it at a reduced rate or not at all. That clause is negotiated harder than anything else on the page and is most of the reason to have someone outside your team negotiate it. A trademark bidding clause with a stated monitoring cadence, because a partner sending genuinely incremental traffic will still bid your brand name if nothing stops them. The lift test: Before a paid placement renews, we suppress the code for a randomised share of sessions or in a matched market, for long enough to clear your purchase cycle. Two weeks clears nothing considered. The output is one number: incremental orders per dollar of placement fee and commission combined. We have used it to recommend cancelling placements the client's own team was fond of, and we will do that here. What we refuse: A homepage or newsletter feature fee on a partner never tested for lift. Public codes that undercut your own site price, which teach the buyer to leave, find a code and come back through someone who then invoices you for the return trip. Exit-intent code widgets on your own checkout, which is paying commission to intercept yourself. And partner-side reporting as settlement; commission is paid against your ledger, reconciled monthly, and disputes are resolved on order IDs. The wrong engagement if: Your margin cannot carry a discount and a commission on the same order. These partners are a volume instrument and are honest about being one. If the arithmetic only works at zero commission, a better partner is not the answer. It is also wrong if placements have to be live next week for a campaign already booked, because the measurement that makes any of this defensible takes longer than the placement itself. Delivered: Signed insertion orders carrying the deduplication and trademark clauses, code allocation per partner, a leakage monitor with a named owner, and the lift test design together with its result. On the retainer, the monthly output is a partner-level contribution statement your finance team can reconcile against invoices without calling us, and a recommendation on each partner that says renew, renegotiate or stop.