I spent several years running affiliate programs in B2C: home improvement, clean energy. In those categories the job is mostly a volume problem: find partners who can send traffic, track it, pay for it, repeat. When I moved that skill set into B2B fintech at PingPong, as Marketing Manager, Ads & Lifecycle from 2021 to 2023, the mechanics carried over. I rewrote every assumption.
In B2C, a signup was the finish line. In B2B fintech it's the starting gun: sales has to work the lead, support has to onboard it, compliance has to clear it. A partner who sends you a pile of junk signups has handed three other teams homework. So I built the program around a different question: what has to be true before volume is worth having?
One honesty note up front: this was fintech, so the performance figures are locked behind NDAs. Where I can't publish a number, I'll tell you exactly what I watched instead. The program ran about two years and held thirty-odd active partners over those two years, and I ran it on my own, with sales and finance alongside me. The mechanics are reproducible.
Infrastructure before volume
My first move was unglamorous: standardize the plumbing. Nothing was there to migrate, since the program started from zero on spreadsheets. I centralized tracking, attribution, and payouts on HasOffers (now TUNE). One system of record for every partner, no matter how they joined. The same ID carried from the click through to the payable event.
In parallel, I recruited two ways. Managing partners directly gave us tighter control and showed us who promoted us and how. Affiliate networks gave us broader reach, at the cost of visibility. Running both, two or three networks beside direct outreach, was the fastest way to learn which trade-off the business could live with, and about six months of it settled the question.
Verdict: direct won, and it wasn't close. When partners graduated to rev-share terms later, the program's highest bar, every one I can recall had come through direct management, or was one I'd pulled out of a network and moved onto direct terms. The network motion earned its keep a different way: it became the place I found partners, and direct management became the way I kept them.
Four rules, written down
Most affiliate horror stories come from programs that scaled first and wrote rules later. I wrote four rules before recruiting anyone:
- I banned cash incentives and paid-search bidding. No partner bids against our own campaigns or manufactures intent with cash.
- I only served North American SMBs. That was the segment we could serve and qualify.
- I approved every creative. Every claim a partner makes is a claim we're accountable for, especially in fintech.
- I kept exclusion lists and updated them so partners never overlapped internal campaigns, so affiliates couldn't get paid for demand we already owned.
Referrals and affiliates stay separate
I ran two models in parallel and refused to blur them.
Customer referrals I treated as a product feature. Modest incentives, organic sharing, driven by trust. A referral program works when you make it easy for a happy customer to do something they already wanted to do.
Professional affiliates I evaluated like a business relationship: on margin, whether attribution held, and whether payouts landed. The two ran on different cadences, with different definitions of success.
Collapsing these into one “partner program” is a common mistake, and it fails both audiences. It gives customers spammy mechanics and hands professionals amateur terms.
Prove quality, then pay for it
I started every new partner on conservative CPA. Easy to write in a policy doc, hard to hold in a live negotiation, because every new partner negotiates. The one I still think about was an affiliate with a large SMB finance email list, exactly the audience we wanted. First call, he asked to skip straight to rev-share plus a higher CPA floor, and quoted what a competitor was paying him. I said no: conservative CPA first, better terms after his traffic proved out. He passed, politely, and took the list to that competitor. That's what the rule costs, and I'd pay it again. A quality bar only means something if it survives contact with a partner you wanted.
“Proved out” meant something specific: did the partner's accounts activate, transact and retain like our direct signups? I read that over roughly 90 days, a 60-day sales cycle plus a 30-day activation window, with a fixed CPA advance per partner until it came back. Terms held payout until KYC approval, with clawback on a chargeback inside 90 days. Only after that held did I move a partner from pure CPA toward revenue share. Rev-share aligns better long term, but it's only safe once you trust what you're sharing revenue on.
Attribution lives in the CRM
The tracking platform tells you what happened at the click. The CRM tells you whether it mattered. I built custom Salesforce fields so every inbound lead carried its source with it. The load-bearing field sat on the Lead. It stored the HasOffers transaction ID from the click, plus the affiliate and offer IDs. On lead conversion, a workflow I built copied those fields onto the Opportunity. So when a deal closed, the revenue rolled up to a specific partner and a specific click with nobody touching a spreadsheet.
Those fields kept affiliate-sourced, partner-led and referral growth separate. When someone asked which motion was producing revenue, I could answer from the system. It's also why payout disputes stayed short: fewer than ten came up across the whole program, and most were settled in a single email, because the transaction ID on our side matched the one in the partner's dashboard. Same data on both sides of the conversation. When a chain broke across devices or clicks, the transaction ID in our system decided it.
Affiliate growth is a system. Everything that worked here reduces to three boring things: traffic quality, accurate attribution, and operational clarity. The network tests, the payout sequencing, the CRM plumbing were all in service of those three.
Questions I get asked
What makes B2B affiliate programs harder than simple acquisition channels?
The feedback loop. In B2C I could judge a partner in about a week: cost per lead in, contribution margin out. In B2B the verdict takes a full sales cycle plus an activation window, so you're paying CPA months before you know whether the traffic was any good. Every rule in this post exists to survive that lag.
What should exist before partner volume starts scaling?
Get every partner into one tracking system, write the rules down, match payouts to how your funnel qualifies, and carry the click ID through to closed revenue. My test: take one lead and trace it by hand from click to Salesforce opportunity. If you can't, don't recruit partner number two yet.
How do you know the program is working?
Watch how partner-sourced accounts activate and retain against your direct signups, and watch how long payout conversations take. Mine were partner accounts activating inside 30 days at 80 percent or better of the direct rate, plus 90-day retention. When partner accounts behave like direct ones and payout emails are one line long, the program is healthy. Signup volume never made it onto my dashboard.