In practice, that is usually the wrong approach.
Modern streaming platforms use loudness normalization, which means they automatically turn tracks up or down during playback so songs play at a more consistent perceived level. Spotify says it applies loudness normalization during playback and adjusts tracks to around -14 LUFS under its normal setting, while also recommending delivery of one stereo master at native resolution rather than multiple platform-specific files. Apple is even more direct: its Apple Digital Masters documentation explains that tracks mastered louder are played back at a lower volume by Sound Check and similar systems, which can make them sound weaker, not stronger.
That is the main reason you should not make several different LUFS versions of the same song for different streaming services. The platforms will adjust playback anyway.
The biggest mistake is assuming loudness target equals mastering target. It does not. A platform’s playback normalization figure is not a command telling you where your master must land. It is simply the level the service may use when adjusting playback volume for the listener. Spotify explicitly notes that a dynamic track mastered around -14 LUFS and a much louder track mastered around -6 LUFS can end up playing back at the same perceived loudness once normalization is applied. The difference is that the more dynamic track may preserve more punch and contrast, while the overly loud one has already sacrificed dynamics before the platform turns it down.
This is where good mastering becomes more important than chasing numbers.
When you build separate masters only to hit different LUFS values, you often create unnecessary compromises. One version may be pushed harder, reducing transient impact and narrowing dynamic range. Another may be backed off too much just to satisfy an online myth about a target number. Instead of improving translation, this can make your release less consistent across services, DJ sets, downloads, and future uses. Apple advises engineers to mix and master in a way that captures the intended sound regardless of playback volume, precisely because listeners increasingly use systems that normalize loudness.
There is also a practical distribution reason not to do this. Spotify’s delivery guidance says to send a single stereo master in the highest-quality native format, and that the platform handles format conversions itself. So even from a workflow perspective, making multiple “Spotify master,” “Apple master,” and “YouTube master” files is often solving a problem that the platform is not asking you to solve.
Another issue is that louder is not automatically better after encoding. Spotify warns that louder masters are more vulnerable to additional distortion during transcoding and recommends keeping true peak under control to reduce that risk. Its guidance specifically notes that if your master is louder than -14 LUFS integrated, you should keep it below -2 dB True Peak to help avoid extra distortion during streaming conversion. Apple similarly provides tools and guidance focused on avoiding clipping and auditioning encoded results, because a master that looks clean in PCM can still produce clipping when encoded.
So what should you do instead?
Focus on creating one excellent master that suits the song. That means making creative mastering decisions based on genre, arrangement, emotional impact, dynamics, and translation, not based on a spreadsheet of platform LUFS targets. If the track needs aggression and density, master for that. If it needs openness and punch, preserve that. Then check that it is technically solid: no unnecessary clipping, sensible true peak control, and no distortion introduced by the encoding chain. Spotify and Apple both emphasize quality delivery and clean masters over platform-specific loudness gaming.
Of course, there are exceptions. You may want alternate masters for clearly different use cases, such as a club version, a vinyl pre-master, a radio edit, or a version intended for a specific live-performance environment. Apple’s documentation itself acknowledges that different destinations like vinyl or clubs may justify different mastering choices because the listening context is different. But that is not the same thing as making multiple streaming masters just because one platform normalizes near one number and another platform behaves slightly differently.
The better mindset is this: streaming services are playback environments, not mastering instructions.
If your song sounds balanced, musical, emotionally right, and technically clean, normalization will usually do its job. If your song is over-limited and flattened just to chase loudness, normalization may simply expose that decision. In other words, great mastering is still essential, but platform-by-platform LUFS chasing usually is not.
The goal should not be to win a loudness contest that the platforms are already neutralizing. The goal should be to make a master that sounds like your record at its best.