The single most useful thing to understand about this strategy is that the return has two independent conditions. The first is the obvious one: the company has to expand as expected. The second is easy to miss: the market has to keep valuing that expansion at a similar multiple of earnings. A company can deliver everything it promised and still lose its holder money, if the price the market is willing to pay for each dollar of those earnings falls at the same time. That is not a rare event. The multiple attached to future expansion is a statement about optimism, and optimism moves for reasons that have nothing to do with any individual company's results, including the general level of interest rates.
The consequence is that "the company did well" and "the investment did well" come apart more often here than elsewhere. In the value case the buyer is paying little for results already reported, so there is less optimism embedded in the price to be withdrawn. In the growth case the price is substantially a forecast, and the buyer is exposed to that forecast being revised, not only to it being missed. This is the mechanism behind the observation on the growth stock page that these shares tend to be more volatile: a price built out of expectations moves whenever the expectations move.
Selection is where the strategy quietly makes decisions the investor did not make. Whichever screen defines "growth" is also, without saying so, deciding which industries the portfolio ends up in, because the companies expanding fastest at any moment tend to cluster in whichever sectors are being reshaped at that moment. So a growth allocation is frequently a sector position wearing a style label. Whether that concentration is acceptable is a decision worth making on purpose rather than inheriting from a screen, and it is the reason the size of a growth allocation matters as much as the decision to have one.
The cycle FINRA describes is the reason a stretch of underperformance proves nothing either way. If the relative fortunes of growth and value swing, then any few years of results is a sample from one part of a swing. A period in which growth trounces everything is not evidence the strategy is better, and a period in which it lags is not evidence it is broken. FINRA's inclusion of "a mixture of the two" as a normal answer follows from the same observation: holding both removes the need to be right about which part of the cycle is next.
Where this page stops. What makes a share a growth stock, the fact that no regulator defines the term, and the SEC fund names rule that requires a fund using the word to write down and stand behind its own definition are all covered with the classification. The valuation ratio most often used to describe how much is being paid for current earnings has its own page, as does the question of whether an active tilt of any kind is worth making.