Sub-IDs and Split Testing for Casino Affiliates

Updated 17 October 2026 · WinnerAffs team

Most affiliates know which of their channels sends clicks. Far fewer know which banner, which line in a video description or which chat command actually turns those clicks into depositing players. Sub-IDs close that gap. This guide shows how they work in your WinnerAffs account, how to name them so they stay readable, and how to run a simple split test without drawing the wrong conclusion.

Two layers of tracking: websites and sub-IDs

Your WinnerAffs account tracks traffic on two levels, and it helps to use each for what it is good at.

The first level is the website. In the Link Creator you add every place you promote from: your Twitch channel, your YouTube channel, your review site, your Discord server. Each gets its own numeric tracking ID, active as soon as you save it. Reports can be filtered by website, so this level answers the broad question: which of my channels brings players to CasinoWinner?

The second level is the sub-ID, or SID. It is an optional label you attach to an individual link. It answers the narrower question: inside this channel, which placement or creative is doing the work? Your report breaks results down by sub-ID, so a label you add today becomes a line you can compare next week.

A rule of thumb: if it is a different audience, make it a website. If it is the same audience reached in a different way, make it a sub-ID.

Adding a sub-ID to a link

Open Marketing Tools, choose a creative and click "Show code". Next to the tracking link you will find a field labelled SID (sub-tracking, optional). Type your label, and the tracking URL and HTML snippet are regenerated with it included. Copy them as usual.

Sub-IDs can contain letters, numbers, hyphens, underscores and dots, up to 64 characters. Spaces and other symbols are stripped out, so stream panel #2 quietly becomes streampanel2. It is better to choose the clean form yourself than to be surprised later.

A naming scheme that survives six months

The first week, any label feels obvious. Six months later, a report full of test1, new and banner-b means nothing. A consistent pattern fixes this. One that works for most partners:

placement-creative-period

• panel-win300-oct for the stream panel banner featuring the WIN300 offer, used in October;
• desc-textlink-oct for the plain text link in your video descriptions;
• discord-tourney-wk42 for the link shared for a community tournament in week 42.

Keep a short note somewhere (a spreadsheet or a pinned message to yourself) listing what each label means. When you discuss results with your manager, that note turns a vague conversation into a precise one.

Running a split test

A split test compares two versions of one thing while everything else stays the same. For a CasinoWinner campaign, common things to test are:

• The message. A creative leading with the 300% WIN300 match against one leading with the weekly leaderboards.
• The placement. The link in the first line of a video description against a pinned comment.
• The format. A banner on your review page against a text link inside the verdict paragraph.
• The moment. A chat command posted at the start of a stream against one posted after a big hit on screen.

Give each version its own sub-ID, for example desc-line1-oct and pinned-oct, and change only that one thing. If you swap the creative and the placement at the same time, you will not know which change made the difference.

Run both versions over the same period. Traffic on a Friday night stream behaves differently from a Tuesday afternoon, and CasinoWinner's own promotions calendar moves activity up and down through the week. Comparing a version that ran during a leaderboard launch with one that ran during a quiet week tells you about the calendar, not about your creative.

Reading the results

In Reports, set the date range of the test and filter by the website concerned. You will see clicks, signups, first deposits and deposits for each sub-ID, along with withdrawals, chargebacks, NGR and your CPA or RevShare earnings.

Look down the funnel, not just at the top of it. A version with more clicks is not automatically the winner. If version A brings twice the clicks but the same number of first deposits as version B, version B is converting its visitors far better, and it may be the better choice for a placement where space is limited. On RevShare, look further still: which version brings players who keep depositing in the following weeks? That only shows up if you leave the sub-IDs in place long after the test ends, which is another reason to name them well.

Be patient with small numbers. A handful of signups can swing either way by chance. Wait until both versions have collected a reasonable number of first deposits before declaring a winner, and if the difference is small, treat it as a draw and test something bolder.

Traps that ruin a test

• Reusing a sub-ID for something new. If panel meant one banner in September and another in October, your history is mixed. Add the period to the label instead.
• Testing during a one-off spike. A viral clip or a jackpot win shown on stream can swamp everything else for a few days. Note the date and leave it out of the comparison.
• Forgetting the website layer. Sub-IDs inside the wrong website make reports confusing. Check that the link was generated for the right site before you paste it.
• Calling a winner too early. The first day is rarely representative. Give the test at least a full week.

Done well, sub-IDs turn your report from a single number into a map of what works. If you have not set up your account yet, start with how to join WinnerAffs, then read how to promote CasinoWinner for placement ideas worth testing.

FAQ

Sign Up   How the program works →

18+ only · please gamble responsibly.