Launching an affiliate programme for a self-serve SaaS product

The situation

The company sells a 49 USD per seat product with no sales call in the loop. Growth has come from content, and paid acquisition has been added on top as organic growth flattened. The blended cost of acquiring a customer through paid channels sits around 420 USD, which is workable but trending in the wrong direction.

Meanwhile a handful of newsletter writers and independent consultants are already recommending the product to their audiences, unprompted and unpaid. There is no mechanism to track that, reward it, or ask for more of it. The question is whether formalising it produces cheaper acquisition than the paid channel it would partially replace.

The three decisions that shaped the programme

Everything else followed from these, and each one is a place where copying an ecommerce default would have produced a worse programme.

• Attribution window of 60 days, not 30. The median time from first touch to paid conversion is 34 days. A 30 day window would have dropped roughly half of genuinely affiliate-sourced conversions, and the affiliates losing them would have concluded the tracking was broken rather than the window short.
• Percentage rather than flat fee. Pricing is public and seat counts vary widely, so an affiliate can estimate what an account is worth before investing effort. A flat fee would have overpaid on single-seat signups and badly underpaid on the 40-seat accounts that make the programme worthwhile.
• A 30 day holding period before payout. Aligned to the refund policy. This resolves almost every bad conversion before money moves, which is far less damaging to the relationship than clawing a payment back afterwards.

Whether the economics work

The comparison that matters is cost per acquired account against the paid channel, not the commission rate in isolation. At 20 percent of first-year value, an account averaging 4 seats produces roughly 2,352 USD in year one and around 470 USD in commission, which looks worse than the 420 USD paid CAC until account mix is taken into account.

In this model the partner-sourced mix skews larger, because consultants recommend the product to organisations rather than individuals, and the effective cost per account lands near 118 USD once the small single-seat signups that dominate paid acquisition are excluded. That gap, not the headline rate, is the reason the programme is worth running.

The figures above are modelled from the stated assumptions rather than measured. The point is the shape of the calculation: run it with your own pricing, cycle length and channel costs before committing to a rate.

The launch sequence

The programme opened to a small known group before any public recruitment, on the principle that operational errors are much cheaper in front of five friendly partners than fifty cold ones.

• Weeks 1 to 2. Five affiliates recruited from existing customers and community members. Tracking links issued, and the full chain from click through signup, conversion, commission record and payout walked end to end personally.
• Week 3. First commission paid, deliberately ahead of the stated schedule. Paying is a different code path from calculating, and running it early surfaces the problems that reconciliation would otherwise find at volume.
• Weeks 4 to 12. Public programme page and marketplace listing opened. 25 affiliates recruited in total, screened on whether they already reach the buyer rather than on audience size.
• End of quarter one. 8 of the 25 produced at least one account. That ratio is normal and is the reason partner count is a poor headline metric: the number worth reporting is active partners, not signed ones.