I jumped a bit late into this conversation of "balance changes resulting in rekindling someone's interest in a game" but I will start with this:
Frankly, I kind of wonder if tier lists matter more at lower levels of play, insofar that a character with very strong tools (especially easy ones) is going to be extremely oppressive, especially since new players are not going to have their own understanding of the system mechanics or their own character's tools to deal with it.
I think thinking of tier lists and match-ups as a single entity is incorrect -- they're really on a continuum that changes as players get better, and it's not just the theoretical ideal play match-ups that matter.
That and playing characters with no real tools to use is almost universally an ass experience that's not worth anyone's time.
It's really because there are 2 different tier lists (or 3 if you want to be even more specific):
*Tier list at "average" skill level
*Tier list at "near peak human" skill level
*Tier list at "theory craft" level
The first one is for your average/casual player. In this tier list, characters with obviously easy to abuse tools are clearly better especially if they are easy to execute. Obvious example is Hulk in UMVC3. Hulk is a noob slayer and even online is essentially a "top tier" character. At the highest level of human play Hulk just barely gets top 8 at majors, not quite good enough to win a large scale major. On the opposite side, Morrigan is not that great at the average/casual level because her fly cancels require execution that is beyond the average player. Plus she does low damage in combos where as Hulk has among the easiest combos in the game that lead into a dead character. This is also where Vanilla Sentinel was the best character because of his st.S and easy mode damage.
The 2nd one is the tier list that most people use and is the one that is used for competitive play in fighting games. This tier list doesn't apply to your average player, only to the top level of play. Execution is a factor but not as big as it is in the first category. A difficult to execute character can be top tier like Morrigan over a easy execution character like Hulk.
The 3rd tier list is pretty much never used and just used more as a proof of concept. In it you see stuff like C.Viper UMVC3 as the best character in the game because it takes execution out of the equation. If someone can perfectly execute a character at all times at frame perfect level then that character is godlike (like Viper). However, human begins are not robots and they can't always do a bunch of inputs in a short time with great accuracy or make optimum decisions all the time. Hence why this type of tier list is not really useful or really accurate for a game which HUMANS play.
We are just going to ignore the 3rd type of tier list and focus on the first two. Most balance changes that companies make are geared towards the second category HOWEVER you will always see companies make some changes in games that are geared towards the first category. I can make numerous examples of these and I am sure everyone has examples that they can think of. It's stuff like nerfing Meepo in DOTA 2 despite him not being used in the competitive scene just because he dominated pub games. It's stuff like nerfing Vanilla Sentinel when he was largely irrelevant in that game at top tier level but destroyed casuals even your average player.
For casual players, the changes at that level matter a lot. For top players, changing Dante's moves from rekka to single input doesn't make him a better character in THEIR category of tier rankings but that is a "buff" for the average player and makes them more likely to use the character. And if changes like that are more likely for them to use that character, it makes it more likely that someone is more interested in the game if they are a Dante fan.
Nowadays these balance changes are becoming more and more important due to the advent of streaming and fighting games becoming more of a spectator thing. People want to see the character of their choice being represented but that's not going to happen if they aren't strong enough. If you are an up and coming fighting game player and you want to learn a character who is deemed "low tier" in that game then you are shit out of luck because you will have nothing to look at, nothing to strive towards. When balance changes come out, you want buffs to your character not only because it impacts the way you play the character but because you want to see that character played by the pros so you can learn from them.
A lot of the times the changes don't even have to do with buffs vs nerfs... it's really about making a character feel more accessible, flow better or have more tools to make the character more interesting for the average player. Or it can be quality of life changes to the character. Stuff like making Guile's cr.LP have one more frame of hit stun so that his basic combo becomes a 2 frame link rather than a 2 frame link (a big change for the average player who may or may not use plinking). This change doesn't really fundamentally raise Guile in the tier lists (of the 2nd category) but it makes him easier to play for average players.
TL
R version: Different tier list for different types of players and developers know this so they make changes accordingly for different types of players. Changes can motivate even a low/average player to play a new character and overall make the game more fun for them through quality of life improvements. And this doesn't even cover the spectator side of things.