Why we launched a paid affiliate program before touching Google Ads

Quick context: BuildBase is a multi-tenant SaaS backend delivered as one npm install (@buildbase/sdk). Auth, billing, usage metering, workspaces, RBAC, workflows, and the rest of the plumbing every SaaS ends up rebuilding by hand. We've been building it in public for a while. Roughly 100 signups, five of our own products running on it in production (PlugNode, AgentCenter, Imejis, RemoteWait, LinkTracer), and a proper leaky funnel between "signed up" and "shipped something real."

Today we launched an affiliate program: https://www.buildbase.app/affiliate

I want to write down why we launched this before opening a paid ads budget, because I think the reasoning applies to almost any dev-tools founder pre-PMF.

Cold ads amplify a broken funnel. Our activation is not where we want it. Users bounce at the org creation popup before they see product value. Throwing paid traffic at that is paying to widen a leak. Warm referrals arrive with a baseline of trust that gets people past the first friction. One IH commenter saying "yeah I tried it, here's what surprised me" outperforms a hundred Google Ads clicks for us right now, because the reader is already past skepticism when they land.

Devs recommending dev tools is the actual distribution channel. The people who buy backend infrastructure are other devs. They ask two questions before paying: who else uses this, and did anyone I trust vouch for it. Cold advertising doesn't answer either. A dev writing a real post about their experience does. If someone's already going to write that post, they should be paid on it. If they're not going to write it, no amount of commission will make them.

What we deliberately didn't do: no whitelabel reseller tier (not interested in MRR laundering), no payout on free signups (paying customers only, otherwise we're funding link farms), no 60-day approval theater (if you sign up and refer, you're in, we approve on real activity not on our judgment of your "brand fit"), no lifetime lockouts (if it doesn't work for you, you leave clean).

Who this is actually a fit for: newsletter operators writing to indie hackers, AI builders, or SaaS devs. YouTube and blog creators covering backend, auth, or billing tooling. Consultants and agencies who set up SaaS stacks for clients - this is probably the highest-LTV cohort, because if you spin up ten SaaS clients a year on the same backend, you have a real annuity. And anyone already recommending BuildBase who wants credit for the sends.

Commission structure, cookie window, and payout terms are on the page: https://www.buildbase.app/affiliate

The honest bit. We're pre-revenue externally. This isn't going to make anyone rich this quarter. But we're priced per app (not per seat, not per MAU), so a small number of conversions actually pays. And we intend to be around a long time. If you place a bet early, it compounds.

We also know we're not the biggest name in the category. Kinde, Clerk, Auth0, Stingg, WorkOS exists. Supabase exists. Firebase exists. The reason we think there's room: none of them are "one npm install for all the backend a SaaS actually needs, on your infra, with your Stripe, no revenue share." That's a specific gap, and it's the one we're aimed at. If that framing resonates with your audience, the affiliate math might work. If it doesn't, please don't force it, the conversions won't happen.

A question for the room: if you've run an affiliate program from either side (as a founder or as an affiliate), what's the one thing you wish had been different? Cookie length, payout cadence, promo assets, dashboard reporting, attribution windows - what actually moved the needle for you, and what turned out to be theater? Trying to get this right before we harden the terms.

Happy to answer anything about the program, the product, or the "affiliate before paid ads" call in the comments.