Affiliate is no longer a niche experiment — 62% of marketers plan to increase their affiliate budget over the next 12 months, and businesses report earning between $12 and $15 back for every $1 invested in the channel. Those numbers are genuinely good — but they hide a selection effect. The brands pulling those returns tend to share a few characteristics before they ever sign up to a network.

You need enough margin to pay a commission and still make money — as a rough anchor, products under roughly $100 usually work better with a percentage commission, while higher-ticket items can sustain a flat payout per sale without eroding margin. You also need something worth writing about: a product with genuine differentiation, existing reviews or UGC an affiliate can point to, and a brand that can survive being represented by someone other than your own paid media team.

If none of that exists yet, affiliate is not a bad channel — it is just a channel you are not ready for. Fix the product-market fit and content foundation first. Everything downstream, from commission structure to recruitment, gets significantly easier once there is something affiliates actually want to promote.

This is the first real fork in the road, and most brands get pointed toward whichever option their software vendor happens to sell. There are three genuinely different models, and each suits a different stage of brand.

Affiliate networks (Awin, CJ, Rakuten) give you access to an existing pool of publishers already relationships with the network — the trade-off is a setup fee and a revenue share on top of your commission spend; CJ, for example, has historically charged in the region of $6,500 to set up a programme plus a percentage of affiliate-driven sales. SaaS platforms (Impact.com, UpPromote, Affiliatly, Tapfiliate) hand you the tracking and payment infrastructure for a flat software fee — often starting well under $100 a month — but you are responsible for recruiting every publisher yourself. In-house builds give you full control and no platform margin, at the cost of engineering time and ongoing maintenance most early-stage teams don't have spare.

For a single-brand DTC store early in its journey, a network or a well-chosen SaaS platform is almost always the right starting point. In-house infrastructure only starts to make economic sense once you're managing meaningful volume across multiple brands or markets, where the override economics and data ownership start to outweigh the build cost.

For most ecommerce brands, 10–15% of sale value is a reasonable starting commission — enough to make the opportunity worth an affiliate's time without eroding your margin past sustainability. But the flat percentage-only model is increasingly the wrong tool on its own, especially once you start working with creators rather than pure content publishers.

A hybrid model — a smaller flat content fee for top-tier publishers or creators, layered with performance commission on top — tends to outperform pure last-click commission. It compensates partners for the awareness and trust-building work that never shows up in a last-click report, while keeping the bulk of your spend genuinely tied to results. When we set commission structures across brands like HOLY Energy and The Female Company, this blended approach consistently kept the best partners engaged past the first campaign, rather than just the first one.

Thirty days is still the number most brands default to, and it remains a sensible starting point for a typical ecommerce purchase cycle. But that default is under more pressure than most programme owners realise. Safari's Intelligent Tracking Prevention now caps first-party cookie lifespans at 7 days, and a 2026 industry survey found 38% of programmes now run attribution windows of 7 days or shorter, against just 21% still offering 60 days or more.

The practical response isn't to panic about the shortening window — it's to stop relying on browser cookies as your only source of truth. Programmes running server-side, first-party tracking report 18–24% higher attributed conversions than those still leaning on third-party cookies alone, simply because they're not losing attribution to browser privacy settings they can't control.

Not all affiliates generate the same value, and chasing volume over fit is one of the fastest ways to waste a launch budget. The main categories worth recruiting deliberately: content and review sites, comparison and "best of" pages, cashback and loyalty partners, coupon and deal sites, and — increasingly — individual creators publishing directly to their own audience.

Creator affiliates with 10K–100K followers are now generating roughly 3.7x more attributable revenue per follower than traditional content or display affiliates on a comparable basis — and the gap is widest in categories like beauty, fashion, and supplements, where a genuine recommendation carries more weight than a static banner. When we built Myprotein's affiliate mix, fitness and nutrition creators consistently outperformed generic deal sites on both conversion quality and long-term retention — which is exactly the pattern the wider data now backs up.

The practical takeaway: define your ideal partner profile before you open recruitment, then actively vet applicants against it rather than approving everyone who signs up.

The software or network fee is the number everyone asks about first — and it's almost always the smallest part of the true cost. The bigger question is who actually runs the programme day to day, because a platform tracks clicks and pays commissions; it doesn't recruit partners, negotiate terms, or chase a dormant affiliate back into activity. Below are rough, assumption-led ranges for the four operating models, based on current published benchmarks for in-house salaries and agency retainers. Treat these as a planning anchor, not a quote — actual costs shift with market, programme scope, and vertical.

The honest ramp looks something like this: weeks one to four are recruitment and onboarding — getting your first cohort of affiliates approved, briefed, and equipped with assets. Months two and three bring a trickle of first sales, usually from the handful of partners who moved fastest. Meaningful, forecastable revenue contribution typically doesn't show up until somewhere between month six and month twelve.

A useful benchmark from a recent industry survey: over a third of brands report that fewer than 20% of their signed-up affiliates are actually posting content in any given month. That's the number to watch — not how many affiliates joined, but how many are still active a quarter later. A small, genuinely engaged roster will consistently outperform a large, silent one.

Most brands don't need an agency to launch a programme — the platforms make that part genuinely simple. What they usually need help with is everything that happens after: the weekly outreach, the newsletter cadence, the commission negotiations, the vetting, the reporting that actually gets looked at.