1) They are are fairly or somewhat undervalued given the conventional market view and metrics.
2) They are likely to be significant beneficiaries of large-scale technological trends that the market is oblivious to and has not priced into the stock.
Then I wait a few years for the trend to become more obvious and for the market to adjust the stock price accordingly for the new upside. My portfolio is essentially a ladder of different technology trends that take 3-4 years to mature. Occasionally one does not pan out but, while I may not make much money, I never lose much money because "good large-cap tech stock".
The only "small caps" I invest in are tech startups. But that is a different portfolio than what I am talking about here.
Since I spend my days thinking about trends in technology anyway, this whole exercise takes little additional effort. I might add or remove something from that list of stocks once a year. When I save money from my paycheck, I semi-randomly put it in a stock on that list so no thinking required.
Yes, I could do some kind of sophisticated portfolio analysis but that would defeat the goal of spending as little time on it as possible. I only check against the S&P 500 as a sanity check. My lifetime annualized return (decades) is about +2 over the S&P 500 but the last five years has been more like +4, so I can't complain. The annual return relative to the S&P 500 has actually been surprisingly steady over time, so no undue volatility.