In my
discussion of these reactions, I draw the distinction between what must happen at the extremes. For an SN2 reaction, I suggest it is similar to billiards in which the cue ball must strike a combination of balls before the object ball can move. For an SN1 reaction, it is like trying to find a parking place in a crowded parking lot. You cannot pull in until someone pulls out first.
In practice, the reactions are generally somewhere between these extremes. For some SN2 reactions, the bond breaking may be more advanced than in other examples, or it can be increased by electron withdrawing ability, solvent or heat. The attack can be advanced with electrons better able to hit the nucleus before the leaving group has much movement. In SN1 reactions, you may find some nucleophile attack occurs before the leaving group has completely diffused away. In that case, the product may have a slight amount of inversion in the product distribution.
I suggest it is easier to rationalize the reactions be starting with those you can best understand as generally SN1 or SN2. For example, SN2 reactions do not occur with tertiary halide and SN1 reactions do not occur on primary halides (unless stabilized by neighboring electrons, allylic and benzylic, for example). In between are the difficult ones, the secondary halides. From some questions, they can be truly difficult, both mechanisms, ambiguous, impossible, not enough data, help…?
My preferred method to learn substitution reactions is to read examples and try to understand why the reactions occur as they do. For example, because carbon is an electron donor (that is why tertiary carbocations are more stable), replacing hydrogens with carbon atoms will make it more difficult for an electron pair to attack the carbon. This retards SN2 reactions and increases the SN1 (carbocation formation) reaction. This will be the same reason for beta substitution to retard SN2 reactions or result in rearrangement in SN1 reactions. The effect of electron donation from the neighboring carbon. (You may find it is called hyperconjugation if CH bond electrons are being donated. However, you may also think it is due to carbon not holding its electrons tightly.)
I also concede there are some examples that become very difficult for me to correctly predict what might happen. This happens with a secondary halide and a protic solvent. Protic solvents are better able to hydrogen bond with the nucleophile or the leaving group. This can provide additional pull to the leaving group and possible carbocation formation. However, I generally assume that if written with a reasonably good negatively charged nucleophile, the reaction is likely to be an SN2 rxn rather than SN1 because increasing the nucleophile concentration can increase the rate. (Dilution can increase SN1 reactions.)
However, that isn't the greatest dilemma in predicting the products, rather competition from elimination reactions. A rule of thumb I found in the textbook written by Brown and Foote is any nucleophile with a pKa > ~12 will result in elimination.