Summary
Most affiliate programmes do not fail because of the idea. They fail because of an unfinished launch: tracking has not been checked, the commission has not been worked out, and publishers get no briefing. This article sets out a 30-day plan for an ecommerce brand and the seven mistakes it helps you avoid. The statements rest on consultancy experience and published sources, and every external claim carries a source and a date.
What are the most common mistakes when starting an affiliate programme?
The most common mistakes are untested tracking, a commission that has not been worked out, no support for publishers, too little time planned, the same commission for new and existing customers, unclear rules for voucher sites and a privacy check left until after launch. All seven can be fixed before the first publisher joins.
| No. | Mistake | Consequence | Countermeasure |
|---|---|---|---|
| 1 | Tracking not tested under real conditions | Sales are not attributed and publishers drift away | Test orders via link and voucher code before launch |
| 2 | Commission chosen without a margin calculation | Every sale costs more than it earns | Calculate with cost of goods, shipping, returns and fees |
| 3 | Same commission for new and existing customers | Publishers optimise for easy purchases | Consider a higher rate for new customers |
| 4 | No briefing after sign-up | Publishers stay inactive | A short welcome pack with rules and materials |
| 5 | Too little time planned | Enquiries pile up | Block a weekly slot in the calendar |
| 6 | Rules for voucher sites left open | Disputes over commission, lost customers | Decide in advance how voucher publishers are paid |
| 7 | Privacy checked after launch | Legal risk, rework | Settle privacy notice and consent before launch |
Why is untested tracking the most expensive mistake?
Untested tracking is the most expensive mistake because you only notice it when publishers leave or sales go missing. The German consultant Marcel Richter lists typical causes: wrong or too short cookie durations, postback parameters passed incorrectly, and consent banners that block tracking [1]. Publishers rarely report such faults. They simply promote other shops.
Before launch, place an order yourself through a test link and with a voucher code, and check that the sale appears with the correct amount. Richter also recommends test orders across different publishers, regular tracking audits and a check after every change to the shop [1]. On selecdoo Starter, tracking runs through a tracking pixel or the Shopify app, using last click or voucher code. Check which route fits your shop and test it with a real order.
How do you work out the commission before you start?
Work backwards from margin: take cost of goods, shipping, returns and all fees off the basket, and what remains is your room for commission. If commission is above that, each new customer costs you money. At selecdoo you set the commission yourself and it is calculated on the net basket, with a network fee of 30 per cent on commissions actually paid. Prices are billed in EUR.
An example: at 10 per cent commission on a net basket of 80 euros you pay 8 euros commission and 2.40 euros network fee, together 10.40 euros or 13 per cent of the basket. This is an illustrative calculation. The consultancy sendschafter gives retail commissions of mostly 5 to 12 per cent and sees little room for affiliate below a margin of about 20 per cent [2]. Check your margin before you fix a number.
Why pay new and existing customers differently?
If you pay every purchase the same, you also pay for existing customers who would have returned anyway, and you give publishers a reason to deliver easy purchases rather than new customers. The UK agency Why Matters recommends 15 to 20 per cent commission, on first-time customers only [3]. sendschafter suggests checking after six months whether new customers make up at least 40 per cent of orders [2].
These are recommendations from individual consultancies, not fixed rules. For you it means deciding before launch how to treat new and existing customers, and building a check after a few weeks into your plan.
What does a publisher briefing look like?
A good briefing fits on one page: who you are, what you sell, which commission applies, how long a purchase stays attributed, which creative formats are allowed and when payments are made. Publishers who hear nothing after signing up go quiet. Plan a personal contact with every new publisher within the first two weeks.
The briefing should also cover voucher codes. If you want to exclude voucher sites, pay them in full or pay them a share, say so up front. Views differ on this topic, so decide on your margin and share the rule from day one.
How much time does a programme take in the first month?
sendschafter plans for 8 to 15 hours a week of management in the early phase [2]. That includes recruiting publishers, answering questions, agreeing promotions and checking statements. If you have less time you can outsource tasks to an agency or a service, which costs money. Put the time block in the calendar, otherwise enquiries pile up.
What belongs in a 30-day plan?
A 30-day plan separates preparation, launch and first review. The periods are a suggestion that you can adapt to your shop.
| Period | Task | Result |
|---|---|---|
| Days 1 to 5 | Work through margin and commission, set rules for voucher sites | Rate and rules are fixed |
| Days 6 to 10 | Set up tracking and verify with test orders, settle the privacy notice | Tracking is verified |
| Days 11 to 15 | Write the briefing, approach the first publishers | First publishers are in the programme |
| Days 16 to 25 | Personal contact, answer questions, first promotions | First clicks and orders |
| Days 26 to 30 | Review: orders, new-customer share, cancellations, feedback | Decide what to adjust |
What applies to privacy in affiliate marketing in the UK?
Privacy deserves a check before launch. The ICO published final guidance on storage and access technologies on 29 April 2026. According to the law firm A&O Shearman it covers cookies, tracking pixels and device fingerprinting, and includes an example of how tracking pixels apply to affiliate marketing consent [4]. Consent remains the starting point under PECR Regulation 6 and the exemptions are narrow, so do not assume that tracking is exempt.
This is not legal advice. Have your privacy notice and consent set-up reviewed by a qualified adviser before you invite publishers. If you also sell in Germany, German rules apply there, and Dr. Datenschutz concludes that consent is needed in many cases under German law [5].
What should you review after 30 days?
After the first month what counts is whether the foundation holds, not whether sales are already large. Check that every test sale was recorded correctly, how many publishers are active, what share of orders comes from new customers and how many orders were cancelled or returned. Optimise only once you have a baseline, and change one thing at a time.
Conclusion
An affiliate programme starts well when tracking, commission, rules and privacy are settled before the first publisher joins. The seven mistakes are not exceptions but the normal outcome of a rushed launch. The page on how to start an affiliate programme step by step shows how to go about it. selecdoo is now opening to UK brands. You can start free or talk your plan through in a Discovery Call. You can cancel with four weeks' notice to the end of a quarter.
Sources
- Marcel Richter (4 February 2026): Perfektes Partnerprogramm-Tracking: So schaffst du die technische Basis (in German)
- sendschafter (August 2026, date per URL 20 August 2026): Lohnen sich Affiliate-Netzwerke? Anbieterübersicht, Kosten und die To-dos vor dem Start (in German)
- Why Matters (15 October 2026): Awin for Shopify Brands: Costs, Setup and Commission Rates
- A&O Shearman (15 May 2026, reporting ICO guidance of 29 April 2026): ICO sets out approach to storage, access and online tracking
- Dr. Datenschutz (20 October 2026): Affiliate-Marketing und Datenschutz: So kann es funktionieren (in German)
Note: All sources were accessed on 4 November 2026. This article is not legal advice. The calculation examples are assumed.