An affiliate tracker is software that sits between a traffic source and an offer, records each click, and joins it to the conversion that follows, so that cost and payout end up in one record. The ad platform knows what the click cost. The affiliate network knows what the conversion paid. Neither knows both halves, and the tracker exists to close that gap. It is a working tool for anyone running paid traffic inside the affiliate marketing overview, where advertisers, networks and publishers meet.
Tracking in general is the process of recording what happens between a click on an affiliate link and a confirmed conversion, then crediting the right partner. The tracker stores the parameters that identify the traffic, assigns each visit a unique identifier, and receives the conversion report from the advertiser or network.
Why three dashboards are not enough
A media buyer normally works with three sources of numbers, and each one is incomplete. The traffic source reports impressions, visits and spend, but it stops at the edge of the landing page. The affiliate network reports conversions and commissions, yet only for offers running inside that network. Web analytics describes on-site behaviour and sessions, and it does not know what a sale was worth. None of the three can say that a specific placement produced a specific payout.
A tracker fills that gap by using a single identifier to connect the click, the visit and the conversion. Manual reconciliation does not replace it. A spreadsheet updated by hand misses the visitor who reads a review, leaves, and buys directly from the brand the next day, and it misses anything that happens on a second device.
The core loop, step by step
The sequence is the same on every platform, whatever the interface looks like. The details differ; the order does not.
- Click. A visitor clicks a tracking link. The link carries parameters that identify the traffic source, the campaign and the creative. Parameters like these sit in the query string, the part of the URL after the question mark; MDN describes the query as an optional component that often carries identifying information as key=value pairs.
- Tracker. The request reaches the tracker's domain before it reaches the offer. The tracker logs the time, device, browser, referrer and traffic source, then assigns a unique click identifier to the visit.
- Offer. The tracker redirects the visitor to the offer, or to a landing page placed in front of it. The identifier travels with the visitor, usually as a URL parameter.
- Conversion. The visitor completes the action the advertiser cares about: a purchase, a registration, a form submission, a subscription renewal or a first payment.
- Postback. The advertiser's server, or the network's, sends a server-to-server call back to the tracker carrying the identifier and the conversion details. A pixel on a confirmation page can do a similar job from inside the browser.
- Report. The tracker matches the incoming conversion to the stored click and updates the dashboard. Cost from the traffic source and revenue from the network now sit on the same row.
The redirect in steps 2 and 3 should be quick, because the visitor is waiting, and the tracker has to record the click before the browser moves on.
What a tracker records
A tracker captures every click and conversion and attaches data points to each one: traffic source, placement, creative, device type, operating system, browser, country, connection type and mobile carrier. Two reasons justify keeping this detail. Optimisation needs a granular view, because a campaign that looks flat at account level often hides one placement that performs and another that drains budget. And disputes need evidence: when a network questions a conversion, the stored record shows the identifier, the timestamp and the parameters that arrived with the visit.
What counts as a conversion
Networks and advertisers define conversions differently, and the definition lives in the contract. Sales, leads, registrations, trial starts and renewals can all be conversion events. Some programs count only the first action from a new customer, while others count recurring payments, which means the tracker has to follow the same customer over months. Custom conversion tracking lets a buyer record these separately instead of collapsing them into one number. The server-side half of that work is described in how server-to-server postbacks work.
Campaign structure inside a tracker
Tracking works campaign by campaign rather than link by link. A campaign has four parts: the traffic source, the offer, the landing page and the flow that connects them. The traffic source is the platform that supplies the visitor. The offer is the destination that pays, often taken from a network; where offers and payouts come from is covered in the CPA networks section. The landing page is what the visitor sees first, and it may be a presell, a review or a piece of content rather than the offer itself.
The flow is the path through those parts, and it can have more than one step: one landing page, then a second, and only then the offer. Support for several offers and landers inside one campaign lets a buyer test combinations without rebuilding the setup. Rule-based distribution then decides what each visitor sees, using country, language, device, connection type or day of the week, and the tracker stores which path each visitor took.
Tracking methods: cookies, pixels and postbacks
Three methods do most of the work, and they fail in different ways. A tracker may combine several to cover more of the journey.
Cookies
A cookie is a small piece of data that a website stores in a visitor's browser. When the visitor returns, the site recognises the browser and can connect the new session to the earlier one, which is how a purchase made days later can still be credited. It is also the most fragile link in the chain. A visitor can delete cookies, and they do not follow a person from a phone to a laptop.
Browser policy limits them further. WebKit states that Safari's Intelligent Tracking Prevention blocks all third-party cookies by default, with no exceptions, and deletes script-written cookies and other script-writable storage after seven days without user interaction. The same document says that when link decoration is detected, the expiry of cookies created in JavaScript on the landing page is capped at 24 hours. Google's documentation for Chrome is less absolute: it notes that third-party cookies may be blocked by browser restrictions, user settings, developer flags or enterprise policy, and tells developers to provide a good experience whether or not they are available. A tracker that depends only on cookies inherits every one of those changes.
Pixels
A pixel is a small piece of code placed on a page, usually a confirmation page. When a referred visitor reaches that page, the browser requests a tiny image file from the tracker's server, carrying identifying information with the request, and that request tells the tracker the conversion happened. Pixels also work inside email, which extends measurement past the website. The weakness sits in the browser: ad blockers, script restrictions and cookie limits can stop a pixel before it fires, and a visitor who closes the tab before the confirmation page loads leaves no record.
Postbacks
A postback is a server-to-server call and does not depend on the browser at conversion time. The advertiser's server sends a request to the tracker's server with the identifier and the conversion details. Because the advertiser records the event on its own side, the report is easier to verify. It is the usual method where the advertiser controls the confirmation step and can store the click identifier.
Some trackers also use browser fingerprinting, which builds a signature from browser settings such as language, time zone and rendering details. It usually runs alongside the other methods rather than instead of them, and privacy rules treat it carefully.
Attribution: which partner gets the credit
Recording a conversion is one problem. Deciding who earned it is another. When several partners touch the same customer, the program needs a rule, and the tracker supplies the raw data for that decision.
| Model | How credit is assigned |
|---|---|
| Last interaction | The partner whose link was used last before the conversion |
| First interaction | The partner who made the first contact in the journey |
| Linear | Every partner in the journey receives an equal share |
| Position-based | The first and last touchpoints receive the larger share |
| Time decay | Credit shrinks the further a touchpoint sits from the conversion |
| Algorithmic | Statistical modelling distributes credit by measured effectiveness |
Multi-touch models exist because a buyer's journey rarely runs in a straight line, and a partner who introduced the product weeks earlier may have done the harder work. The model is not a setting that a buyer picks alone. It is defined in the affiliate agreement, and two networks can apply different rules to the same kind of sale. Every model needs a readable identifier on each touchpoint; a conversion that arrives with nothing attached cannot be credited to anyone.
Cloud-based and self-hosted trackers
A self-hosted tracker runs on the buyer's own server. It gives full control over data collection and over how redirect paths and postback rules behave. The trade-off is maintenance: someone has to install, patch and scale the server, and the technical skill required is real.
A cloud-based tracker runs on the provider's infrastructure. There is no server to maintain, capacity can be expanded without buying hardware, and updates arrive without a migration project. The limitation is flexibility, because data handling follows the provider's design rather than the buyer's. Neither is automatically better. A solo buyer running a few campaigns usually prefers the lower maintenance, while a team with heavy volume and engineering capacity may prefer to own the stack.
A second split is between standalone trackers and integrated affiliate management platforms. A standalone tracker records events and produces reports, and reconciliation with payouts stays manual. An integrated platform adds commission automation, finance modules and partner management, which suits networks and companies running many partners at once.
What to look for in tracking software
The feature list is long, but a handful of items decide whether a tool holds up under real traffic.
- Link and code generation. The tracker issues tracking links and the code for confirmation pages.
- Whole-chain statistics. The dashboard shows the path from click to conversion, not only the endpoint.
- Conversion recording by pixel and postback. Both methods should be available, since some offers support only one.
- Groupable reporting. Reports can be sliced by placement, creative, publisher, device and funnel step, with the option to drill into a single event.
- Redirect control. Rule-based redirects, caps and traffic distribution determine what each visitor sees.
- Integrations and an API. Traffic sources, networks and analytics tools need a connection, and an API covers what is missing.
- Access and permissions. Shared reports, read-only access and separate workspaces keep a team from overwriting each other.
- Alerts and rules. A falling conversion rate or a traffic spike from one source should reach a person quickly, and rules can pause a weak placement on a set condition.
Metrics worth watching
Traffic is bought under several cost models, including CPM, CPC, CPL and CPA, and a tracker has to line up spend recorded under one model against revenue earned under another. Cost reaches the tracker through integrations with the traffic source, which is why those integrations matter.
- Clicks and impressions show how many people engaged with the creative.
- Conversions count the completed actions.
- Conversion rate divides conversions by clicks and shows whether the traffic matches the offer.
- Revenue per visit gives a value on traffic before a campaign has scaled.
- Cost and profit come from syncing spend out of the traffic source into the tracker.
- Refunds and chargebacks reduce revenue after the fact and need their own line in the report.
A live dashboard shows a losing placement while there is still budget to move; a daily refresh describes yesterday.
Tracker, analytics and network dashboard: what each one answers
The three tools all display numbers, but they answer different questions.
| Question | Affiliate tracker | Web analytics | Network dashboard |
|---|---|---|---|
| Where did the visitor come from? | Yes, with campaign and placement detail | Yes, but usually without partner identifiers | Only traffic inside that network |
| Which partner earned the conversion? | Yes | No | Yes, for its own offers |
| What did the conversion pay? | Yes, when revenue is synced | No | Yes, at the network's terms |
| How did the visitor behave on the page? | Limited | Yes, in depth | No |
| What did the visit cost? | Yes, when the traffic source is integrated | No | No |
| Can spend and revenue sit on one row? | Yes | No | No |
Analytics explains behaviour, a network dashboard explains a single network, and the tracker is the one that joins cost, visit and payout into a single record. Most teams run all three. For a side-by-side view of how tracker products differ on hosting, postback support, reporting and access, see affiliate trackers compared.
FAQ
What is the difference between an affiliate tracker and an affiliate network?
An affiliate network is a marketplace that connects merchants with publishers and usually takes a share of the commissions it generates. A tracker is technology: it records events, attributes conversions and reports on them. A network can run its own tracking, and a brand can run a tracker without joining any network.
Does an affiliate tracker replace web analytics?
No. Web analytics shows how visitors behave on a site, which pages they read and where they leave. A tracker shows which source produced the visit, what it cost and what the conversion was worth. Most teams run both because each answers questions the other cannot.
Why are postbacks considered more reliable than cookies?
A postback travels from server to server and carries the identifier directly, so it does not depend on a browser storing anything. Cookies can be deleted, blocked or expired by browser policy, and they do not follow a person across devices. The click identifier still has to survive the journey to the advertiser for the postback to match.
What happens when a conversion is not recorded?
The publisher is not credited for it, and the advertiser's report shows a gap it cannot explain. Broken redirects, lost click identifiers and misfiring pixels all produce that outcome. Keeping the raw click and conversion logs makes it possible to trace where the record was lost.