Network Selection and Migration Without Paying Twice

Network choice is mostly a question of who pays whom. CJ, Awin, Rakuten, ShareASale and Commission Factory take an override on every commission you pay, ordinarily a quarter to a third of it, so their cost scales with your success and never stops. Impact and PartnerStack charge a platform fee plus a smaller per-action charge, so their cost is heavier at low volume and flattens as the program grows. Above a certain monthly commission spend the platform model is simply cheaper, and I will run your numbers through both structures before recommending anything, including the case where your current network is the right answer and you should stay. How I choose, beyond the pricing model: Payment terms and who holds the funds. Whether the network can commission at item level or only on order value, which decides whether you can pay differently on a high-margin line. Whether partner-level SubID and journey reporting is available to you as raw export rather than as a dashboard chart, because everything useful downstream depends on getting the log files out. Recurring-revenue handling, which is why a subscription business usually ends up on PartnerStack or Impact and a retailer usually does not. And Australian reality: Commission Factory, now inside Awin, is where a large part of the Australian retail publisher base actually is, and a US-authored recommendation that never mentions it was not written for this market. Implementation, in the order it has to happen: Container or mastertag placement and the first-party tracking domain, set up as a CNAME on a subdomain you control. Server-to-server postback for the order event, so tracking is not hostage to the browser. Item-level payload if the commission grid needs it. Consent handling tested against your CMP in both accept and reject paths. Product feed built to the network specification with a refresh cadence that matches how often your prices actually move, since a daily feed on a business running weekly promotions produces publishers advertising prices you no longer honour. Deep-link generator and coupon inventory. Then a test-order pass across desktop, mobile web and app before a single partner is invited. Migrations, which is where the money gets lost: Network contracts commonly require ninety days notice, and the overlap is where advertisers pay twice for the same order. I plan the parallel window explicitly: which network is authoritative during overlap, how orders are de-duplicated between them, what evidence I need from each side to reject a double claim. Partner porting is the other cost. Partners do not transfer, they re-apply, and the ones you care about need a reason and a person to talk to. I write that outreach and prioritise it by revenue, so your top partners are transacting on day one on the new platform rather than discovering the change from a broken link. Not included: I do not run the program afterwards, I do not write the commercial terms of your commission grid here, and I will not take responsibility for code deployment inside your stack. I specify, I test, and I sit with your engineers, but your team ships it. One thing worth checking about me: I hold no referral or revenue-share arrangement with any network or platform, and I will put that in writing before we start. A large share of network recommendations in this market come from people being paid by the network they recommend, and the fee is invisible to you. Ask anyone advising you on this the same question. Who this is not for: Anyone who has already decided on the network and wants an implementer. That work exists and it is cheaper than my time. Anyone under twelve months old as a business, where the right answer is usually to run direct partner relationships on a simple tracking setup and revisit this once volume justifies the overhead. And anyone whose real problem is that the program is not producing revenue, because moving it to a different network will not change that.