Running a Self-Managed
Affiliate Programme
The tools are the easy part. What self-management takes week to week, why unattended programmes decay, and the case for doing it yourself at the right size.
"Self-serve" describes the software, not the workload. Most writing about running your own programme comes from companies selling that software, so the distinction rarely gets made, and plenty of self-managed programmes stall a few months in without anyone ever deciding they should.
The subject deserves better. What follows covers the three things "self-managed" can mean, what the platforms handle and what they leave to you, the workload week by week, the common failure modes, and the case for self-managing anyway. At the right size, that case is strong.
"Self-Managed" Means Three Different Things
- A self-serve network tier. Products like Awin Access (about $49 a month plus 3.5% of tracked order value), Tradedoubler Grow (from €59 a month plus 4%, falling at higher tiers) and impact.com's entry plan (about $30 a month plus 2.5%). You get the network's publisher marketplace, tracking, validation and payments, with limits on commission-rule flexibility, reporting depth and support. Figures are indicative for 2026, so check the current pages.
- Standalone affiliate software. Refersion, Tapfiliate, UpPromote, Post Affiliate Pro and similar tools. Full control and lower percentage fees, with no meaningful marketplace attached. Every partner is one you found, pitched and onboarded yourself.
- An unmanaged programme on a full network. The programme exists and nobody runs it. A surprising number of brands are in this state when they eventually search "why is my affiliate programme not making sales".
The workload below applies to all three. The only question is whether anyone is doing it.
What the Platforms Do for You, and What They Don't
The platforms take care of tracking links and the order-confirmation tag, contracts and terms, commission calculation, the validation workflow, payments, and a directory listing that partners can browse. That used to need an enterprise budget. It now costs tens of pounds a month, which is why entry-level programmes exist at all.
No platform will choose your commission strategy, recruit the partners worth having, answer partner messages, spot low-quality traffic, negotiate placements, plan seasonal activity or chase your validation queue. The software is infrastructure. The programme is the work.
The Workload, Week by Week
| Cadence | Task | Why it can't slip |
|---|---|---|
| Weekly | Review partner applications | A steady stream arrives, including low-quality applicants that need vetting. Approving everyone is how traffic problems start |
| Weekly | Validate sales | Partners notice late payment before they notice anything else. A queue past 30 days costs you your best relationships |
| Weekly | Partner messages & comms | Placement offers, content requests, code queries. An unanswered message is declined revenue |
| Weekly | Performance check | Watch click-to-sale by partner. Divergence is your early-warning system (see below) |
| Fortnightly | Recruitment outreach | The growth driver, and the first thing dropped under time pressure |
| Monthly | Offers & seasonal planning | Publishers plan content four to eight weeks out. Miss the window, miss the coverage |
| Monthly | Commission review & reporting | Rates drift out of date, and finance will eventually ask what the channel returns |
Done well, this adds up to at least half a day a week and sometimes a full one. Recruitment is the part that compounds, so when a busy week forces choices, protect that block above the rest.
One example of why the weekly performance check earns its place. We've watched a single sub-network partner push around thirty thousand low-value clicks into a programme in a month. Clicks rose, conversions fell, and in a monthly report the pattern would have sat unnoticed long after it had skewed every spend decision. Someone watching the click-to-sale line caught it within days.
Five Ways Self-Managed Programmes Quietly Fail
- Set-and-forget decay. Without recruitment, the 80/20 concentration tightens until one partner's bad month is the whole programme's bad month.
- Approving everyone. Low-quality traffic, brand-term bidding and attribution games are real. The industry spent 2024 and 2025 arguing about browser-extension attribution for good reason.
- One flat commission. It overpays the bottom of the funnel, underpays the top, and leaves nothing to negotiate with. Rates by partner type fix all three.
- Validation neglect. Partners promote the programmes that pay promptly. Fall behind and your best relationships are the first to cool.
- No offer calendar. The brands that appear in seasonal coverage planned for it a month or two earlier. Silence reads as absence.
When Doing It Yourself Is the Right Call
Often, and without apology. Under about £250k of online revenue, fixed management costs would swallow a large share of the channel's return, so self-serve keeps the economics sensible. It also suits the testing phase, founders or marketers with protected weekly hours, and niches where you already know the publishers that matter personally.
Worth knowing as context: platform-run comparisons claim actively managed programmes generate several times the revenue per partner of unattended ones. Those figures are vendor-published, so treat the multiple with caution. The underlying point, that attention moves the number more than tooling does, matches what we see in practice.
The Signs You've Outgrown It
Watch for recruitment stalled for a quarter, a validation queue past thirty days, three partners making up more than 60% of revenue, placement offers declined for lack of time to assess them, and a flat channel while site traffic grows. Two or more of those together usually means the constraint is no longer budget or tooling. It's attention and relationships, and that calls for a different setup.
The good news is that handovers are routine and nothing about them is destructive: partners, tracking history and terms all carry across. When and how to move to managed services covers the whole process, and there's a separate comparison of who to hand over to.
Quick Answers
Doing It Yourself
Yes. Below about £250k of annual online revenue you probably should, because fixed management costs would eat the channel's return. The tools are built for it. What the tools can't supply is time: at least half a day a week, protected, with recruitment treated as the priority rather than the task that slips.
Half a day to a full day per week if you want it done well. The weekly core is application vetting, sale validation, partner messages and a performance check. Recruitment outreach, offer planning, rate reviews and reporting sit on a fortnightly or monthly rhythm. Recruitment compounds, so protect that block above the rest.
Either can work, because they solve different problems. A network brings an existing publisher marketplace along with tracking, contracts and payments. Standalone software has lower per-sale fees and no marketplace, so every partner is one you found yourself. Small brands without recruitment capacity tend to do better starting on a network's self-serve tier.
Work through the usual suspects in order. No active partners means recruitment stalled. Partners promoting with nothing tracked means a broken tag, so test an end-to-end sale. Clicks without conversions points to traffic quality or your own conversion rate. And a commission set too low will keep you out of placements in your category. An unmanaged programme sitting on a network usually has several of these at once.