Coupon and Cashback Placement Priced on Incremental Lift

Across the cashback and coupon partners we have tested for service and lead-gen advertisers, the incremental share of what the partner invoiced has landed between 20 and 45 percent. The rest was traffic already bought and paid for elsewhere in the account. So we do not open a cashback partner before there is a way to switch it off in one market and read what happens to bookings. The test comes before the placement: A geo or time-based holdout, three to four weeks, measured on bookings or qualified leads rather than on attributed revenue, since attributed revenue is the thing in question. In Israel the split is usually Tel Aviv district against the rest of the country; internationally it is one market against a matched one. The stopping rule is written before the test starts and we hold to it when the number is inconvenient, which is the only reason to write it down first. What actually gets negotiated: Rate tiering that pays differently for a first-time customer and a returning one, and differently again for a session where no code was applied. Exclusion of sessions carrying a brand-search or direct touch inside the previous thirty minutes. A clause keeping the partner off extension and toolbar inventory, which is a separate placement with separate terms and should never ride along on a cashback insertion order. SubID structure, so that cashback, coupon pages and editorial placements from one partner arrive as three lines instead of one. And code leakage: partner-specific vanity codes, no public codes, and a takedown obligation with a deadline measured in hours. Israel is its own negotiation. The cashback layer here is thin and largely issuer-led, run as benefit clubs by the card companies rather than by an independent in the Rakuten shape. You are buying a slot in a bank's marketing calendar at a co-funded rate, not a CPA, and we price it as media with a performance floor and say plainly that it is media. Your own checkout is part of this: An open promo code field in a booking flow is an invitation to leave and go looking for a code. We put it behind a link, remove exit-intent code offers aimed at people already in the flow, and where the advertiser sells appointments we move the payable event from booked to attended. This is a smaller piece of work than the partner negotiation and usually a larger share of the result. What is not included: No code generation inside your commerce platform, no creative production, no network migration, no support for partners we did not place. Ongoing management is the monthly engagement, not the project. What we refuse: We do not onboard a coupon partner on default last-click terms. Where the network cannot express the tiering, we report that the placement is unavailable rather than run it anyway and reconcile later. We do not work with a partner who will not release SubID-level reporting. We do not put a public code into the world for an advertiser with a price ladder to protect. And we do not present a partner's own attribution report to you as evidence of anything. Who this is not for: Advertisers who need the partner live before a baseline exists; the placement is easy and the answer afterwards is unobtainable. Advertisers in categories with no discount mechanic and no cashback-eligible transaction, which includes most legal work, and we will say that on the first call rather than invent a route in. And advertisers who want to be told how many partners are now live rather than which ones are worth keeping. You end with a shorter partner list than you started with, terms that survive a finance review, and a test protocol your own team can rerun in six months without calling us.