One video in this dataset pulled 3,857,230 views and sold 18 units.
Another pulled 720 views and sold 7.
The second video converted at 9.72 sales per thousand views. The first converted at 0.005. Both sit in the top-GMV bracket for their product line, which means both were among the best performers their brand had.
We found a 2,083x conversion spread inside the winners' bracket alone while auditing 5,871 affiliate videos across nine FMCG product lines in Indonesia. The videos that traveled furthest converted worst, and the pattern was consistent enough to plan around.
First, the Part That Is Not Controversial
Views and revenue do correlate. Across all 90 top-performing videos in the sample, the relationship between view count and GMV runs at r = 0.73. That is strong. A video with more reach usually does produce more revenue, and any affiliate model that ignores views entirely would be wrong.
The median top performer converts at 1.37 sales per thousand views. Median views for a top performer: 25,860. Median views for a zero-revenue video: 2,829. Winners get roughly nine times the reach of failures.
So reach matters. The argument is not that view count is noise.
The Part That Breaks
The correlation holds in the middle and falls apart at the top.
| Category | Views | Units sold | Sales per 1,000 views |
|---|---|---|---|
| Drain cleaner | 720 | 7 | 9.72 |
| Car air freshener | 2,185 | 15 | 6.86 |
| Roach spray | 1,239 | 8 | 6.46 |
| Home air device | 3,676 | 22 | 5.98 |
| — median: 1.37 — | |||
| Reed diffuser | 1,330,998 | 163 | 0.12 |
| Home air device | 387,114 | 40 | 0.10 |
| Reed diffuser | 402,459 | 25 | 0.06 |
| Car air freshener | 3,857,230 | 18 | 0.005 |
Every video in the top half of that table has under 4,000 views. Every video in the bottom half has over 380,000. The four highest-reach videos in the winners' bracket are also the four worst converters in it.
The car air freshener video at the bottom is the clearest case. Nearly four million views produced eighteen sales. It still qualified for the top-GMV list because eighteen sales of a mid-priced item beats most of what that product line produced. It earned its place while converting at roughly one sale per 214,000 views.
Why Reach and Conversion Pull Apart
A video gets distributed because people watch it, share it, and finish it. A video sells because the people watching it were already close to buying.
Those are different audiences, and TikTok's distribution system will happily find you the first one.
When a video breaks out of its category audience, the incremental viewers are not shoppers. They arrived because the video was entertaining, the creator was familiar, the sound was trending, or the thumbnail was odd. A car air freshener video that reaches 3.8 million people has escaped the population of people currently thinking about how their car smells.
The 720-view drain cleaner video never escaped anything. It reached people with a blocked drain. Seven of them bought.
What This Costs You in Practice
Two decisions get made on view data, and both go wrong at the extremes.
Boost allocation. Paid amplification usually follows organic traction, because traction is the signal that arrives first. Amplifying a video that is already converting at 0.005 per thousand takes a video that found the wrong audience and buys it more of the wrong audience. The reach compounds. The conversion does not.
Creative direction. When a video breaks out, the natural response is to make more like it. If the breakout came from a trending sound or a format quirk rather than from product clarity, the next ten videos inherit the reach mechanics and the conversion problem at the same time.
Neither mistake is visible on a dashboard that leads with views.
The Metric That Separates Them
Sales per thousand views costs nothing to compute and reorders the ranking immediately. Run it across your last sixty affiliate videos and the top of the list will not resemble your views leaderboard.
Three ways to use it once you have it:
Set a conversion floor before boosting. A video below the median for its category does not get paid amplification regardless of how fast it is traveling. Reach on a poorly converting asset is a cost, not an achievement.
Separate the two objectives in your brief. Content built to reach and content built to sell can both be worth making. They should not be the same video, and the creator should know which one you are asking for.
Read low-view winners as templates. A video that converts at 6 or 9 per thousand on 1,200 views has found the buying audience with the wrong distribution. That is a fixable problem. A video converting at 0.005 on 3.8 million views has the opposite problem, and it is much harder to fix.
What the Best Converters Had in Common
The high-efficiency videos in this sample were unglamorous. Under 4,000 views each, narrow audiences, no breakout mechanics. What they carried was a specific problem stated in the first two seconds, a product visible in the opening frame, and an exact offer near the end.
They were not trying to be interesting to everyone. They were trying to be obvious to someone with a blocked drain.
For the full teardown of what separates converting videos from the rest, see the analysis of 5,871 FMCG affiliate videos. For why the same creator produces both kinds, see the case for briefing over recruiting.
The Trade You Are Actually Making
Reach is easier to buy, easier to report, and easier to feel good about. It also correlates with revenue well enough that optimizing for it will not obviously fail.
It fails quietly instead, in the gap between a video that 3.8 million people watched and the eighteen who bought.
