In modern music glue is also a warm round big sound where you don't have lots of transients cutting sharply through the mix. Again, for this to be possible to achieve you really need quite a good monitoring situation, much is possible by reading meters well, but it certainly helps to have good monitoring to achieve this kind of commercial sound. If you have a monitoring system that exaggerates harsh transients it is much easier to soften the mix. The sound sources that remove this softness are: vocals, bass, kick, snare and electric guitar. It is when the combination of these are producing much too loud peaks typically more on one side that you'll produce a sound that is not "glued". When you have compressors in the mix that are set at a very fast release, which is quite common, then an undesired bi effect is that some of these loud peaks cut through too much. It can therefore pay off to adjust so that you have short release times overall, but at certain frequency ranges you set the release times a bit longer so that the compressors have enough time to actively keep those transients softened.
Another technique for glue is really softness itself. Soft mixes in terms of loudness/low volume, should be avoided, because they lose depth in the ears of normal listeners. Depth is the result of a combination of great gain staging and great low end. The art of glue is to make it loud enough to have enough depth yet maintaining great transient control so that the mix remains soft at that volume. And that it takes time learning to do well, you solve that issue as early as in the recording process. This is how highly professional music production works, you solve the issue as soon as possible, in fact in the early parts of the recording process.
A common misunderstanding in mixing, especially among people mixing in home studios, is that sound quality is achieved late in the overall production process. It is true that you can do a lot in both mixing and mastering to improve the production quality, but sound/mix/production issues should be dealt with as early as possible because else they will negatively impact on the engineer's sound quality standard in regards to the production, in other words great sound makes you more sensitive to bad mixing moves and that's a very important thing to be aware of. This is also why a great sounding monitoring environment helps a lot - do not only focus on translation and frequency balance when it comes to your monitoring setup. Make great sound come out even greater sounding in your monitoring environment, so that you have extremely great sound that will be memorized before you make your mixing moves. In this way the bar is set high and is kept high because the greater the initial sound is, the greater is the sensitivity to signal loss.
If you receive a recording that does not sound good and you are forced to mix it, then your only focus initially should be to make those frequencies by themselves as good sounding as possible first of all, so that you raise the bar, then take a long enough break, so that you forget the bad sound. Once the better sound has been memorized, it's much easier to make more mixing moves, because you are much more sensitive to sound quality degradations. Glue is therefore, when we're talking about big and pristine, the efforts of establishing good enough sound early enough in the overall production process. For this reason I'm a fan of spending a lot of time on the rough mixing process and I'm not afraid of applying temporary filter constraints on top of it very early to make me have to find ways of improving the sound even more. So try to make it sound the way you want the final sound to be, as early as possible using as little "mojo" as possible, so that you get a very good sounding dry sound to mix. Keep in mind that once you are mixing in the land of poor sound, you'll not going to notice how much the signal has already degraded. A/B as early as possible, not only late like most do... Another thing is that when you evaluate this, do so per sound source in solo at maximum gain. That is where you have the potential. It is this potential of each sound source, you have to remember, so that once you mix the sound sources together which forces signal loss, you will have a good frame of reference - especially when your monitoring environment sounds really sweet. So get yourself great sounding monitors and headphones, not only "flat" ones.