A number one debut and a million-stream week both look like proof. Most of the time they aren't. A chart position is a formula that Billboard rewrites, and a raw stream count is a number you can buy for a few dollars per thousand. Neither one answers the question that decides whether an artist is actually building something: did real people hear the song and choose to come back? We design campaigns around that number, and it's the one the headline metrics are best at hiding.

The chart is a formula, and the formula keeps changing

In January 2026, Billboard changed the math behind the Hot 100 and the Billboard 200. It narrowed the weighting between paid and ad-supported on-demand streams from 1:3 to 1:2.5, so an album unit now takes 1,000 paid streams or 2,500 ad-supported ones, effective January 17. YouTube disagreed with the new weighting and stopped submitting its data to the US charts rather than go along with it.

A song's chart position depends on which streams get counted, how much each one is worth, and which platforms are even in the dataset. All three of those just moved at once. A metric built on rules that can change and data sources that can walk out mid-game is great for a press release and weak as evidence that you built a fanbase. The same #1 under last year's formula and this year's formula are not the same achievement, and neither one tells you how many of those listeners will be there for the next single.

A big stream count can be padded, bought, or farmed

The raw total is easier to inflate than most artists realize. About 106,000 new tracks are delivered to streaming platforms every day (Luminate), most of them landing at close to zero plays, so there's enormous incentive to manufacture volume and stand out. And it happens at industrial scale. Beatdapp, a fraud-detection firm, estimates that at least 10 percent of global streams are fraudulent, pulling roughly $2 billion a year out of real artists' royalties.

The extreme version is now a criminal matter. In March 2026 a North Carolina man pleaded guilty in the first US criminal case for AI-assisted streaming fraud, after using bots to stream hundreds of thousands of AI-generated tracks billions of times and collecting more than $8 million in royalties. At its peak the operation was faking around 661,000 streams a day.

You don't need a bot farm to pad a number, though. Paid stream services will sell plays by the thousand to anyone with a card. Plenty of big numbers are real. The catch is that the count by itself can't separate a bought stream from a real one, so on its own it can't tell you whether you have listeners or just a receipt.

The streaming metrics that matter: did they come back?

The metrics we grade campaigns on are the ones that a bot can't fake without a human actually behaving like a fan.

Save rate comes first. It's the share of listeners who add the song to their own library, and it's the single clearest read on whether the attention you bought found the right people. Below 8 percent, something is off. 8 to 15 percent is solid. Above 15 percent is strong, and above 20 percent is a genuinely good campaign. A bought stream never saves, so save rate is close to impossible to fake at scale. We made the full argument for why save rate beats stream count separately, because it's the number we report on every release.

Retention is the next read. When the budget stops, does the daily stream baseline hold and keep climbing, or collapse back to where it started? Real listeners leave a tail. Bought volume evaporates the day the spend does. Watching the post-campaign baseline is how you tell the difference, and it's a big part of how we read a release.

Finally, conversion into things you own: followers, super listeners, email captures, pre-saves for the next release. Each of those takes a deliberate action, and deliberate actions are what real fans produce and fake volume can't. You can see the gap between a vanity number and real engagement laid out across our campaign case studies, which is where a save rate and a retention curve mean more than a chart screenshot.

Why this changes where the money goes

If you optimize for the vanity number, the cheapest way to win is to buy volume, and you end up paying for streams that leave the moment the budget does, with a platform purge always able to claw them back. If you optimize for save rate and retention, you're buying the behavior that gets Spotify's algorithm interested (our design target is 10,000 streams in 28 days at a 10 percent save rate) and building an audience you can reach again without asking any platform's permission. Same budget, very different asset at the end of the month.

So our end-of-campaign reports lead with saves, retention, and audience growth instead of a chart position. The chart is someone else's scoreboard. The save rate is yours, and it's the one that still means something after the news cycle moves on. We publish that breakdown on real campaigns, and we hold every MeansMGMT release to it: the metrics a fan produces and a budget can't.