Affiliate Marketing on Amazon, the Whole Route from Zero

Amazon's sign-up form asks for the address of your site before the site has to exist. The onboarding documentation is explicit: "One of the requirements for becoming an Associate is that you provide the URL of your site. If your site is not yet public but you want to test against the API, you must still provide a URL during registration."

Read alone that sounds like an accommodation. Read next to what submitting sets in motion, it is the first trap on the route, because the form starts a qualifying window Amazon does not pause while you build.

I am Serdar D., founder of usebravery. My own Associates registrations sit in three stores, Amazon US, Amazon UK and Amazon DE, and I build product sites for other people in the same three. Everything quoted here was read on Amazon's live pages on August 21, 2026.

The order the route runs in, and what each step locks

Written out as tasks the route looks reorderable. It is not, because two of the steps fix conditions that later work cannot undo.

The third row is the only one that starts a clock, and it is also the easiest to reach for early, because a form feels like progress and the eleventh page does not.

What the application is actually asking you to declare

The Program Policies put it in one line: "You must identify your Site(s) in your application." Site carries more than a domain. The Operating Agreement defines the word inside a sentence rather than a glossary, and the list reaches past a website to social media content you post, an online software application and an Alexa skill; the review page inspects mobile apps and social pages too.

Two conditions sit on that declaration. "You must own your website", and acceptance is not a test you can score in advance, because "We will determine suitability at our sole discretion." Nor does it go quiet after review: the identification of your Site has to stay complete, accurate and up to date at all times.

One reconciliation, because two numbers do get published. Amazon's Application Review Process page asks for "at least 10 posts" as a rule of thumb, and for content that is "recent (generally within the last 60 days)", with the German version saying the same. Both sit on a help page rather than in the contract documents, and that distinction is drawn in what the approval bar is made of.

Two different 180 day figures, and only one is your deadline

The first of them is the application window, worked through in the piece on that deadline. Amazon's review page states the count it wants inside it, three qualifying sales in the first 180 days, and answers the obvious question up front: "Please note that personal orders do not qualify".

The second figure is newer. Amazon's summary of its April 14, 2026 change reads: "Added 180-day time limit requirement that products must be shipped to, streamed or downloaded by, and paid for by the customer to qualify for commission." That limits the order, not the applicant, and I quote the summary because the clause behind it needs a login.

The same page narrows what a referred order pays on: onsite commission income now applies only to purchases of the variant you actually linked. Point a reader at the black one, let them buy the grey one, and the variant rule decides.

Four ways past the audit, each closed by name

A new domain with no traffic and a running clock makes four ideas attractive at once, and Amazon closed each of them by number.

Requirement 29 deserves Amazon's own words, because the second half of it is easy to miss: "Similarly, you will not request or encourage any of your friends, relatives, or associates to purchase any Product(s) through Special Links for use by you or them or for resale or commercial use of any kind." A sale you asked a relative for does not count at all.

A fifth idea, buying the traffic, is missing from that table because it is only half closed. Requirement 13 bars bidding on identifiers where the resulting advertisement is a Prohibited Paid Search Placement, defined in the Program Policies as an advertisement bought on terms including the word amazon, or kindle, or any other Amazon Mark. The change of April 14, 2026 does not look at keywords at all, and disqualifies purchases by customers referred through any paid or boosted advertisement linking to Amazon, whether prohibited keywords were used or not. Whether an advertisement pointing at your own pages is caught by that is a reading rather than a quotation, because the clause behind the summary is not publicly readable.

What stays open is published too: "You may purchase paid search advertisements and submit links to Search Engines to appear in response to a general Internet search query". Money buys visits to your pages. It cannot buy the three sales, only traffic a site would still have to convert.

The downside is not symmetrical either. A rejection for unsuitable content can be reapplied for once the content complies. Any other rejection, or termination connected to a violation, means no re-joining without Amazon's advance authorization.

The page needs the sale that the page is supposed to make

A product page earns by being current, and price and availability are not numbers you may type in. They come from Amazon serving them through the link, or from an API whose door carries its own count, ten qualifying sales in the trailing thirty days, worked through in the API piece.

That is the closed loop at the centre of the route: the feature that makes a product page convert sits behind the result the page is supposed to produce first. The starting tool is plainer than its name suggests. The SiteStripe help page today lists one link builder, "Get Link: Text - lets you build short and long links to the current Amazon page you are viewing."

How the links themselves are formed belongs to the piece on building them. Getting through the loop is unglamorous: the work that resolves it is finished before the clock starts.

The bookkeeping is designed once and read on two lags

The identifier structure has two levels that read alike and behave differently. Amazon's example runs storeid-20 as the store ID, with storeid-1-20 for a website and storeid-2-21 for a newsletter beneath it, while earnings accumulate under the single account.

Two published facts make that a design decision rather than a habit. There is a limit of 100 tracking IDs per associate account, and reports include the tracking ID for each transaction. Attribution you did not build into the link cannot be recovered afterwards.

Results arrive at two speeds. Ordered Items and Bounties are updated hourly, and Amazon states that most orders will be reported within three hours of being placed, while the earnings report runs a day behind.

A second market is a second registration, not a setting

Amazon writes this one without hedging: "Your Amazon Associates tag works only in the locale in which you register. If you want to be an Amazon Associate in more than one locale, you must register separately for each locale." The same page lists sign-up addresses for twenty-two marketplaces.

The same requirement is published separately inside each programme; Britain files it under a page titled Account Dormancy and the German help centre carries the identical node in German. So a second market is a second clock and a second set of pages written for that reader. Sequencing is handled in the multi country piece.

What stays yours, and what usebravery takes over

Put the route back together and the work has a plain shape. Before submitting, a browsable property has to be live, recent enough to look tended, on a domain you own. After submitting, three qualifying sales have to arrive from a site nobody has heard of, with none of those shortcuts available, while the data feature that would help most is locked behind that same result. Then the obligations that never end: the property list kept accurate, the tracking scheme designed early, and a rulebook that moves, with a change page one entry deep.

That first block is exactly what usebravery does. You choose the category the site covers, and the site is built, hosted and run on our side, so the reviewer arrives at a published site rather than a template and the qualifying window is not spent on construction. Plans differ by market coverage, Starter covering the United States, Plus adding the United Kingdom and Pro adding Germany, and the coverage per plan is set out on the pricing page.

The boundary should decide it for you. The Associates account is yours, the tag inside the links is yours, and Amazon pays the commission to you directly; usebravery takes no share of it, and the subscription pays for infrastructure. Four things stay on your side: the application, the tax interview, the payment method, and the sales.

Where to start, given that the clock starts on submission

Building this yourself, the ordering is most of the advice. Settle the category first, because it governs what you can publish week after week. Build until a stranger browsing would call the site a going concern. Decide the tracking scheme before the first link. Submit last. The wider map is the beginner piece.

If that first block is the one you would rather not build, it is the part that is for sale, and the plans are compared on the pricing page. Either way the sequence holds. The site comes first and the clock second, and that order is the one thing with no correction after it.