How to tell which YouTube video is actually making you money

Picture a monitor review from eleven months back that, by every visible metric, looks dead. Low views, no comments in weeks, buried on page four of the channel. And yet it's quietly behind the best Amazon payout month all year.

The only way to find that out is to check the individual link on that specific video. Nothing in Amazon's dashboard would ever point there on its own.

That's the actual problem with "which video makes money." It's not that the answer is hard to find. It's that most of the tools creators already use were never built to give it to you, and the ones that could are usually set up wrong or not set up at all.

Your affiliate network isn't hiding this - it just never tracked it that way

Amazon Associates identifies you. Not your videos. One tag, one account, one running total. If three of your videos link the same product with the same tag, Amazon has no mechanism to tell those three sales apart, because from its side there was never a "which video" question in the first place.

The same is true almost everywhere else. ShareASale, Impact, CJ, most in-house brand affiliate programs - they're built to answer "did this creator drive a sale," not "did this specific upload drive a sale." A handful support sub-IDs or campaign tags you can append per link, which get you closer to real per-video data. Almost nobody sets these up, because nothing tells you it matters until you're a year in, staring at a payout that doesn't match your view counts anywhere.

Even when a network does support sub-IDs, the reporting UI for them is often an afterthought - a CSV export buried three menus deep, updated on a lag, with no dashboard view at all. Creators who do set them up correctly often still end up building their own spreadsheet on top, because the network's own tools weren't built with a video-by-video workflow in mind.

So the honest starting point: you're not going to pull a clean per-video dollar figure out of Amazon or most other programs without real effort. What you can pull, easily, is clicks per video. And clicks turn out to be a lot more useful than people give them credit for.

Clicks aren't revenue. They're the closest thing you'll get

A click means someone left YouTube to go look at something. Whether they bought it that day, three weeks later, browsed and left, or bought something else entirely once they got there - none of that comes back to you by video. That gap is real and it doesn't close, no matter how good your tracking setup is.

Part of the reason it doesn't close is structural, not a tooling problem. Attribution windows on most affiliate cookies run 24 hours to a few weeks. Cross-device behavior - someone clicks on their phone, buys later on a laptop - breaks almost every consumer-grade tracking method without a trace. None of this is something you can fix from the creator side. It's baked into how the whole affiliate ecosystem is built.

But here's the reframe that actually matters: what you're trying to answer isn't "exactly how much did this video earn." It's "which of my videos should I make more of." Clicks answer that second question almost as well as dollars would, and unlike dollars, you can actually get them at the video level without needing your affiliate network's cooperation at all.

Signals that correlate with money, even without a single sale-level number

A handful of things tend to track with real earning power, in rough order of how much they matter.

Click-through rate, not raw click count. A video with 3,000 views and 90 clicks is telling you something a video with 50,000 views and 60 clicks isn't - and the raw click numbers alone would hide that completely. Rate normalizes for size in a way totals never do.

How long a video keeps pulling clicks, not just how many it got on day one. The monitor review above wasn't new. It had been quietly ranking in search for the better part of a year, earning the whole time nobody was checking. A video that spikes once and goes quiet is a different animal entirely from one that keeps a slow, steady click rate going for months - even if their lifetime totals end up looking similar on paper.

Price and consideration level of what's linked. Don't compare a $15 impulse buy against a $300 considered purchase and expect the click rates to mean the same thing. Higher-consideration purchases convert at lower rates almost universally, for reasons that have nothing to do with your video's quality - people research more, compare more, and take longer to pull the trigger.

Comment section behavior. Questions like "which one did you end up buying" or "is this still your daily driver a year later" are a decent proxy for buying intent that a lot of creators overlook entirely. A video that generates those questions is doing something a video that doesn't isn't, even before you look at click data.

Seasonality and timing. A gift-guide video published in November behaves completely differently from the same content published in March, and comparing their click-through rates head to head without accounting for that is comparing apples to oranges. If you're building a baseline across your channel, tag videos by whether they hit a seasonal window or not.

Format, more than topic. A dedicated comparison or review format tends to outperform a passing mention regardless of the specific product involved. If you're trying to find your best format before you've even picked your next topic, this is usually the stronger signal to sort by.

Why cross-device behavior makes this even messier

Most affiliate tracking assumes a tidy path: click a link, land on a product page, buy right there. Real behavior rarely works that way.

Someone watches your video on their phone during a commute, taps the link, skims the product page, and closes the tab without buying anything. That evening, on a laptop, they search for the product by name directly and complete the purchase. From the affiliate network's side, that's a brand new, untracked visit - the cookie from the morning's click never made it to the laptop, because cookies don't travel between devices on their own.

This happens constantly with anything above impulse-buy pricing, which is most of what creators actually review. The click was real, the video drove the sale, and none of it shows up anywhere connected to your link. It's not a flaw in your setup. It's just what cross-device shopping looks like when the tracking mechanism was designed for a single browser session.

There's no fix for this from the creator side - it's a structural limit of cookie-based attribution, not a mistake you're making. Knowing it exists, though, changes how much weight you should put on any single video's "measured" click total. The real number of people it influenced is almost always higher than what gets credited.

If you're running more than one affiliate program

A lot of channels aren't on just Amazon. Between Amazon, a software affiliate program, maybe a brand's in-house program, and something like Skillshare or a course platform, the "which video makes money" question gets harder, not easier, because now you're stitching together data from three or four dashboards that don't talk to each other and don't share a common video identifier.

The practical fix is the same one that solves the single-program version of this problem: track clicks per video yourself, independent of which program the link happens to belong to. If every link on every video runs through one system you control, you can compare a video's Amazon-linked click rate against its software-affiliate click rate on equal footing, something none of the underlying networks will ever do for you natively.

Building a real system instead of checking sporadically

The starting point for all of this is simple and non-negotiable: every video needs its own tracked link, even when several videos point at the exact same product. Without that one prerequisite, everything else in this article is unavailable to you - you're stuck reading one pooled number and guessing.

From there, a workable routine looks something like this. Tag every new upload with a distinct link the moment it goes live, not after the fact - retrofitting old videos is worth doing, but it means losing whatever data existed in the gap. Check click-through rate on a monthly cadence rather than daily; day-to-day numbers on any single video are too noisy to mean anything, and checking too often just trains you to react to noise. Once a quarter, do a deeper pass specifically on videos older than six months - this is where the "hidden earner" pattern above tends to live, and it's the pass most creators skip because older content feels done.

Keep a simple running list, even just a spreadsheet, of video, format, price tier, and click-through rate. You don't need anything sophisticated. What you need is enough history to eventually spot which combination of format and price point keeps winning for your specific audience, because that pattern is almost always more valuable than any single video's performance.

A naming convention that pays off later

None of the above works if six months from now you can't remember which link belongs to which video. This sounds trivial until it isn't - a channel with a hundred-plus videos and a habit of reusing generic link names ends up with a spreadsheet nobody can actually read.

A simple pattern works fine: video topic or upload date, followed by the product, kept consistent across the whole channel. Something like pairing a short video identifier with the product name, so a glance at the link tells you both what video it lives on and what it points to, without needing to click through and check. It doesn't need to be clever. It needs to be the same pattern every single time, because the value of this system comes entirely from being able to trust it a year later without re-deriving what any given link was for.

This matters more than it sounds like it should, because the videos most likely to become hidden earners are exactly the ones you'll have forgotten details about by the time they start proving themselves.

A worked example

Say a gear channel publishes three videos in the same month: a dedicated review of a $250 audio interface, a "what's in my studio" tour where the same interface appears for about four seconds, and a "top 5 gear I regret buying" video that mentions it in passing as something the creator kept.

Pooled together under one shared link, all three would show up as a single number with no way to tell them apart. Tracked separately, a pattern usually appears fast. The dedicated review, despite likely having the fewest views of the three, probably pulls the highest click-through rate - its entire audience showed up already comparing interfaces. The studio tour might pull a respectable number of clicks purely on volume, even though the click rate itself is unremarkable. The "gear I regret" video, ironically, might convert reasonably well too, since anyone watching a video about buyer's remorse is often mid-research on the same category themselves.

None of that shows up if all three descriptions share one link. It just reads as "clicks," with no way to know which video is actually doing the work - which is exactly the trap the eleven-month-old monitor review from the opening avoided only by accident.

Common mistakes people make chasing this

Judging by views first. Views measure reach. They tell you almost nothing about whether the people who showed up were in a buying mindset, which is the actual variable that drives clicks and sales.

Writing off old videos too early. The videos most likely to be overperforming are the ones you've stopped thinking about. A video needs months, sometimes longer, to show its real evergreen pattern - checking once right after publish and moving on misses this completely.

Comparing across price tiers like they're the same thing. A high click-through rate on a cheap accessory and a modest one on an expensive tool can represent similar or even better real earning power for the expensive item, once you account for margin and price. Raw click-rate comparisons across very different price points mislead more than they inform.

Trusting a single data point. One good week on one video is a coincidence until it repeats. Look for a pattern across several videos of a similar format before treating it as a real finding you should build a content strategy around.

Assuming your biggest video is automatically your best earner. It's the single most common wrong assumption in this whole topic, and it's wrong often enough that it's worth actively checking rather than assuming.

Reading your own numbers

What you're seeing	What it probably means
High click rate, modest views	Comparison, review, or "best X" format - people showed up already deciding
Huge views, thin click rate	Entertainment audience - the product wasn't why they clicked play
Old video, steady trickle of clicks	Evergreen search traffic, easy to miss on a weekly glance
Clicks fading fast after a strong launch week	Recommendation-driven spike, not lasting search demand
Comment section asking follow-up buying questions	Decent proxy for real purchase intent, even before checking click data
Every video shows an identical click count	Shared link across videos, not per-video tracking - fix this before reading anything else

Frequently asked questions

Can I ever get exact per-video revenue?

Rarely, and only if your affiliate program supports sub-IDs or campaign tags per link and you're disciplined about assigning one to every video. Most creators don't have this set up, and most programs make it optional rather than default, so the data simply isn't there for anyone who hasn't built it deliberately.

Is clicks-per-video really a good enough substitute for revenue data?

For the decision that actually matters - what to make more of - yes. It won't hand you an exact dollar figure, but it reliably tells you which formats and topics are worth doubling down on, which is the more useful output most of the time anyway.

Why does my most-viewed video make so little?

Views measure attention, not buying intent. A video people watch for entertainment can rack up huge numbers and still have almost nobody who was ever going to click a product link, because that's not why they were watching.

Should I set up sub-IDs on every affiliate link I use?

If your program supports them well, it's worth doing - it's the closest thing to real per-video attribution available. If it doesn't, or you're juggling several programs with inconsistent support, per-video click tracking as a fallback still gets you most of the way to the same insight.

How long should I wait before judging whether a video is a real earner?

At least a few weeks for a rough read, and several months before ruling on whether it's a genuine evergreen performer. Early numbers on any single upload are noisy regardless of format.

Does the price of the product I'm linking change how I should read click data?

Yes - always compare within a similar price tier, not across tiers. A high click rate on a cheap product and a modest one on an expensive product can represent very different, sometimes comparable, real earning outcomes once price and margin are factored in.

What's the single biggest mistake creators make trying to find their top-earning video?

Assuming it's whichever video has the most views. It's the default assumption almost everyone starts with, and it's wrong often enough that checking properly is worth the effort every time.

I'm on Amazon plus a couple of other affiliate programs. How do I compare across them fairly?

Track clicks yourself, per video, independent of which program the underlying link belongs to. None of the affiliate networks will ever cross-reference each other's data for you, so the only common ground you'll get is the one you build on your own side.

Why did a sale I'm sure my video caused never show up anywhere?

Most likely cross-device behavior - someone clicked on one device and bought on another, which breaks cookie-based tracking entirely and isn't something you did wrong. It's a known limitation of how affiliate attribution works, not a sign your setup is broken.

The takeaway

You're not going to get a clean, per-video dollar figure from Amazon or most affiliate networks - that data was never being separated to begin with, and building it yourself through sub-IDs is more effort than most creators want to take on. What you can get, easily and for free, is clicks per video. And clicks are close enough to useful that they change what you'd otherwise be guessing at.

Tag every video separately from day one, watch click-through rate over raw counts, give old content the months it needs to show its real pattern, and keep a simple running record instead of checking sporadically. That's the whole method. It's just usually not the one people start with, which is exactly why the best-earning video on most channels tends to go unnoticed until someone actually checks.

Want a click number per video without digging through your affiliate dashboard? GoloLink's free tier tags every upload automatically.