A lot of people insist on having maxed out affection. Also, just talking to your slaves stops being a viable option as you start reaching 50, or more, slaves. Eventually pool parties and dinners will give more affection in less time.
Personally I'm happy with pool parties and occasionally checking that no one's dropped off unexpectedly, at which point I compliment them.
/snip
my approach has always been have one PURE nurse to raise very low affection and happiness.
Make sure my house is upgraded asap with a decent chef and pantry girl with enough maids.
I target advocates to get them to remove demoralization from the household, as demoralization kills affection and happiness.
Never keep malevolent people in the household unless actively removing the trait.
never keep Feminist in the household unless actively removing it or getting them to love you with low willpower which is easier said then done.
Never have three or more combined feminists or malevolent people in the household unless as captives while I remove the other traits from the NPCs.
I tend to disagree about the too many slaves to raise affection comment. I am always working on one NPC to get them to undying with targeted interactions once there they stop dropping. At least that is my understanding.
The above isn't too difficult once you start being able to defeat most enemy groups in the city. and with multiple slaves now you can make quick coin. Couple that with a workshop and you have the income to do all of the above. i never had to spam dinner parties to make the my household stable affection wise and I found the focused approach actually less time consuming and cheaper in the long run because I don't have to worry about constant affection drops.
Then again i get a core party recruited and trained and slowly add to my household vs add people asap. I find if you mass recruit you get problems but if you take a slower approach you get faster growth in the end because you can concentrate on your combat team and key household positions and this creates stable income, influence gain. Coupled with the fact that you don't get instability in the household stats. Also I find that training takes time so mass recruitment tends to not provide the return that a slower approach concentrating on training does, again over the long run. In the beginning mass recruitment can see a big jump in gains but it creates high overhead in terms of expenses and food. And there are so many slaves people don't really train them so they progress so slowly.