How to Choose an Affiliate Program That Actually Pays Off

When I opened my first directory of affiliate programs years ago, I sorted it by commission and started from the top. A program paying 25% looked like the obvious pick; one paying 3% I did not even open. Six months later I had a few euros from the first and nothing at all from the second, because I had never tried it. I know now that it should have been the other way round: the 3% program sold refrigerators, the 25% program sold phone cases.

This article is for publishers who run a website, a newsletter, a YouTube channel or anything else with an audience and want to pick a program that will genuinely earn money for them, rather than the program with the prettiest number in the directory. We will go through how to convert a commission into money per click, which parameters deserve your attention, how to test a program and when to leave it without regret.

Why a high commission means nothing on its own

The commission percentage is only one factor in a multiplication. The second is the order value and the third is the probability that the visitor you send will buy anything at all. Only the product of all three tells you how much each click is worth. A high percentage is very good at hiding the other two factors.

Take two programs you will routinely see side by side in a directory. Program A is a shop selling accessories, average order around 12 €, commission 20%. Program B sells home appliances, average order roughly 600 €, commission 3%. At first glance A wins clearly, since 20 is more than 3. But one order earns you about 2.40 € with A and 18 € with B. One refrigerator order brings in as much as seven and a half orders of phone cases.

Now add the conversion rate. Cheap goods are bought on impulse, expensive goods are not. It is realistic that three visitors out of a hundred buy from program A and one out of a hundred, perhaps fewer, buys from program B. The table below shows an indicative commission from 100 clicks at different conversion rates. The figures are a model for comparison, not a promise.

For program A to catch up with program B it would have to convert seven and a half times better. If one visitor in a hundred buys from B (18 €), seven and a half in a hundred would have to buy from A (18 €), which is possible with impulse goods but far from guaranteed. The more realistic scenario is 3% against 1%, that is 7.20 € against 18 € per hundred clicks. The program with the “weak” commission earns two and a half times more.

The number that comes out of this is called EPC (earnings per click). In our example program A has an EPC of about 7 cents and program B about 18 cents. EPC is the only metric that accounts for everything at once, which is why we will come back to it several times in this article. Some networks display it for their programs; for the others you calculate it yourself from your own data after a few weeks.

Parameters worth watching

A directory or a program page in a network shows you a set of figures. Some matter more than they look, others less. I will go through them in the order I read them myself when choosing.

Commission amount and type

You know the percentage. Fewer people pay attention to the reward type, although it changes the whole calculation. CPS (commission per sale) is the most common: you receive a share of the order value. CPA (a fixed amount per action, such as a completed registration or a first order) suits situations where order values vary widely and you want predictability. CPL (payment per lead) is typical for services, insurance or finance: you are paid for a completed form, not a purchase, usually from one to several dozen euros per contact. With CPL, check what counts as a valid lead and what share of contacts the merchant eventually accepts.

Also notice whether the commission is flat or tiered by product category. An electronics shop commonly pays 1-4% on its core range but 8-10% on accessories and own brands. If you write about accessories, your return from that program is completely different from someone writing about televisions. Volume tiers (more sales, higher rate) are pleasant, but do not count on them at the start.

Average order value

The second factor from the opening example. Not every program publishes it, but you can estimate it from prices on the shop's website. If the best sellers cost 8-20 €, expect an average order of around 16-25 € (people occasionally take two). With furniture, sports equipment or electronics you are quickly in the hundreds. Convert the percentage into euros per average order and only then compare that figure between programs.

Cookie duration

The cookie determines how long after the click a purchase is still credited to you. On the Czech and Slovak market 7 to 60 days is usual, exceptionally 90 or more. How much the length matters depends on what you sell. Impulse goods for a few euros are bought now or never, so even 7 days is enough. With more expensive goods a person compares first, reads reviews, asks their partner and buys two weeks later. With a 7-day cookie you lose them; with 30 days you usually still have them.

The shop's conversion rate

The third factor, and the hardest to find out. A network sometimes states a program's average conversion rate or its EPC across all publishers. If not, form your own picture: go through the shop's checkout as a customer. How long is delivery, what does shipping cost, how many payment methods are there, how does it look on a phone, are there reviews. A shop that charges 6 € shipping on a 12 € item will convert poorly regardless of your work. A shop with free shipping from 40 €, next-day delivery and pickup points will convert well, also regardless of your work.

Commission approval period and rate

A commission is not credited immediately after the purchase; it waits for approval, usually 14 to 45 days. In the meantime the merchant waits out the return period and checks that the purchase followed the rules. The period itself is not a problem; the approval rate is the bigger issue. If a merchant approves 95% of commissions, that is fine; 70% is already a lot of rejections, and below 60% something is wrong: either the shop has a high return rate (fashion, shoes) or it rejects more than it should. Some networks display the approval rate for a program. If not, ask the affiliate manager; an honest one will tell you.

Remember that a rejected commission is the loss of all the work behind it. If your EPC is 18 cents and the merchant approves 70%, your real EPC is under 13 cents. The figures in the opening table shift accordingly, sometimes substantially.

Payout threshold and dates

Less important for the choice itself, but essential for cash flow. Networks usually pay once a month after you reach a minimum, typically 20 to 40 €, and more with some in-house programs. For a small site earning 2.40 € per order, a 40 € threshold means seventeen approved orders before you see any money at all. Find out whether the network pays automatically or you have to request payment, how invoicing works, and whether you can invoice as a private individual or need a trade licence. With merchants' own programs, also ask who actually pays you and what their payment record is like.

Program terms

This is the part most publishers skip and then wonder about rejected commissions. Typical restrictions you will encounter:

• No PPC on the brand. You may not buy search ads on the shop's name or misspellings of it. Irrelevant for content sites, essential for PPC publishers.
• Coupon sites banned or restricted. The merchant does not want you publishing discount codes it did not give you. If you run a coupon site, do not even try a program with this condition.
• No cashback. Similarly, the publisher may not return part of the commission to the customer.
• Own website only. Some programs do not credit sales from social media, newsletters or paid advertising.
• Excluded product categories or zero commission on discounted goods, gift vouchers and the like.

Read the terms before you apply to a program, not after the first rejected commission. If your promotion format falls into a grey area, write to the affiliate manager and get the answer in writing. It will save you arguments three months later.

Communication with the affiliate manager

A program is not an algorithm; a person looks after it. And that person decides on commission approvals, individual rates, whether to send you a product for review or create a discount code for your readers. You can find out how quickly and how substantively they respond before you even register. If they do not reply to a question about the terms within a week, they will be just as slow with disputed commissions. If they answer to the point and offer help, you have something worth more than a rate a few points higher.

I manage three programs of my own and am also a publisher in several dozen others. From the other side it looks like this: a manager notices publishers who get in touch, send a link to their article and ask whether the shop sees it the same way. Such people get a rate increase sooner than an anonymous account with twice the turnover. Communication is a business tool, not a courtesy.

Program stability

You can recognise a dying program before it is officially closed. Typical signs: no new banners or creatives appear in the network, the terms have not changed for years, the affiliate manager does not reply or has changed three times in six months, the commission is lowered “temporarily” and stays low, approvals drag past the stated period, more is rejected than before. Sometimes it helps to look at the shop itself: if it is running a stock clearance, has closed its physical store and stopped adding products, the affiliate program will be the first thing it saves on.

Topical fit as the main criterion

Everything above is secondary to one question: does your audience want what the program sells? You can have a program with a great EPC, a long cookie and an exemplary manager, but if you write about gardening and the program sells gaming laptops, the result will be zero. Conversely, a program with average parameters that fits exactly what readers come to your site for can be your best source of income.

Topical fit is not just about category. It also depends on price level (readers of a minimalism blog will not buy luxury watches), on the stage of the buying decision (a how-to site attracts people who already own something, a comparison site attracts people who are deciding) and on trust (recommending a shop you have bought from yourself works differently from a link to an unknown company). The best-performing affiliate links I have ever placed pointed to products I genuinely used and wrote about because they interested me, not because they had a good commission.

A practical approach: first write down what your visitors come to your site for and what situation they are in. Only then open the directory and look for programs that answer that situation. In that direction, not the reverse. Whoever starts from the directory and looks for somewhere to “fit” a commission ends up with content nobody cares about.

Testing a program

Even with a good table and a topical fit you do not know in advance how a program will perform with your audience. Only a test tells you, and the test must be small, measurable and long enough.

A small test. Pick two or three articles or videos where the program makes sense and place links in them. Do not rework the whole site. If you can choose between several competing programs in the same category, put them in comparable places so they can be compared. Tag the links (most networks let you add your own parameter to a link, a so-called SubID) so you can see which article and which program bring conversions.

Tracking for 2-3 months. A month is too short. Commission approval takes 14-45 days, so after a month you only see clicks and pending commissions, not the actual result. Two months is the minimum, three is better, especially for goods with a longer decision process. Do not change the link placement or the text during the test, otherwise you will not know what caused the result.

Comparing EPC. When the test is over, calculate for each program: approved commissions divided by the number of clicks. You get an EPC that is genuinely yours, not the network average. Compare it between programs and also with what the same space earns otherwise, for instance from contextual advertising. If a program has an EPC of 2 cents and an ad unit in the same spot earns 5 cents per click, you have your answer.

Rough benchmarks for Czech and Slovak content sites: an EPC below 5 cents is weak, 8-20 cents is decent, above 40 cents is excellent and usually comes with expensive goods or services. These are coarse orders of magnitude that vary by topic and audience quality, but they help you separate a program worth expanding from one that merely takes up space.

When to leave a program

Leaving a program is as important a decision as joining it, and it is made less often than it should be. Reasons I would not hesitate over:

1. After three months of testing the EPC is consistently below what the same space earns otherwise, and there is no reason to expect a change.
2. The commission approval rate has fallen below roughly 60% and the affiliate manager cannot explain it.
3. The commission was cut without notice or retroactively.
4. Payouts are repeatedly late or you have to request each one.
5. The shop has worsened the shopping experience (higher shipping, longer delivery), so conversions fell and there is nothing you can do about it.
6. The program no longer matches your content because the shop changed its range or you changed your focus.

Leaving does not mean deleting links. It means replacing them with links to a competing program you have tested in the meantime. That is why it pays to have a backup program lined up in every important category, even if you are not using it fully yet. And if you have a good relationship with the manager, tell them why you are leaving. Sometimes the terms change as a result.

One reason I would not leave over is a single weak period. Seasonality is strong in e-commerce (gardening in spring, electronics before Christmas, sport in January) and two weak months may mean nothing. Compare the same months year on year, not consecutive ones.

Following a program on AffilPlanet

Most of what I have described becomes easier if you stop watching programs by hand. AffilPlanet lists over 1,700 programs from the eHUB, Dognet, CJ Affiliate, Affial, AffiliatePort, eSpolupráce and Affilmax networks as well as merchants' own in-house programs in one place. The directory lets you filter by minimum commission, minimum cookie length, network, category, country and the type of publisher the program welcomes. So you can directly display, say, programs for content sites with a cookie of at least 30 days and a commission from 8%, and only then go through those using the criteria above.

When you follow a program, you receive a notification the moment the merchant changes the commission, the program terms or the discount codes on offer. These are exactly the changes that are easy to miss in a network and that decide whether a program still makes sense. You learn about a commission cut or a new coupon ban the same day, not from a list of rejected commissions two months later.

Programs that have a manager in the directory also have an interest button and direct messaging. You can ask the affiliate manager about the terms, the approval rate or the possibility of an individual rate before you register with the network. The speed and substance of the reply is, as I wrote, important information in itself.

Recording results month by month

The last thing publishers forget: write down your results. Networks show you data only for their own slice, every in-house program has its own statistics, and a year later you cannot piece together which program actually earned you the most. Without summary figures you fall back on choosing by commission percentage, because you have nothing better.

The publisher account on AffilPlanet has, for every program you promote, a monthly record of revenue and commissions with a chart across programs, saved links and codes, and invoicing reminders. It takes copying the figures from the networks once a month, and after six months you have an overview no single network can give you: which programs are growing, which are stagnating, where it pays to add content and where it is time to leave. It is a few minutes a month and it is the best decision-making tool you can have.

If you do not have an account yet, registration is free, and besides recording results you get program following and contact with the managers.