The importance of exits are two fold:
1. The obvious - keep as much of the profit as possible
2. The less spoken of - to help control overall position risk
For non-scalpers, risk management is probably the most important and least understand element of trading. Everyone's heard the trading maxim "Cut your losses and let your winners run". The key aspect of that is the warning to manage your positional risk because if you do the profits will develop.
So #2 above is what those who believe in scaling out (or profit taking) are ultimately interested in.
As already noted, it's not of particular importance to most scalpers, but ongoing risk management is particularly important to intraday/multiday swing or position traders. Over the long run, minimizing overall losses is at least as important as basic profitability.
Ordinary profit targets (i.e., exit the whole position at X) and trailing stops aren't always the best approach. Ordinary trailing stops can work OK in trends but in choppier markets aren't as effective/profitable. Profit taking is especially meaningful/useful to those trading based on support/resistance.
Simplistic example:
Let's say you enter 1000 shares long at 50 with the thought that the price might extend to 60 over the next couple of day. But you see that there's potentially strong resistance sitting at 52 and 54 before there's room to run to 60. So as the trade progresses, if the price action starts to stall at 52, you might sell half of the position at the +2 level and then move your stop to 48 (a net breakeven stop). You might alternatively move it to breakeven (depends on the volatility of the stock). So at this point, you're still targeting that it could go to 60 but short of a catastrophe you've reduced your positional risk to near zero.
Now if the price gyrates around for a while and finally breaks out through 52 but then again stalls at your 54 expected resistance level, you sell part of your remaining position at 54 and move your stop to at least breakeven (again depends on volatility but at this point you'd want no less than a breakeven stop). So at this point, with partial profits in your pocket and a breakeven or better stop, short of a catastrophe your overall position is profitable with near zero position risk.
Now you can be especially comfortable letting your remaining position run (if it's going to) with some profits in your pocket and almost neutral risk.
There are variations on the theme - for instance, using profit taking stops rather than hard profit taking targets. So in the above example, if you're looking for resistance at 52, rather than simply taking partial profits at 52 you might start a 1/2 point trailing stop to sell part of the position. You risk 1/2 point (or whatever is appropriate based on the stock and the trader) in exchange for possibly getting stopped in profit taking at a higher price. The possible advantage of this is that if the resistance is broken before your profit stop is hit, you might be able to kill the profit stop and let the breakout develop further and delay profit taking. Effectiveness depends on how large of a profit stop you're willing to allow, how volatile the stock is, and whether the price action really just congests at resistance slowly eating through it or is repelled by the resistance.
Note that some who have criticized the idea of partial profit taking often base it on backtesting using fixed profit taking intervals. It's reasonable to assume that such profit taking might perform suboptimally. It's also confused if most of the comparison period is in trending rather than choppy market. Finally, fixed profit taking intervals aren't an effective comparison because it doesn't recognize the basic idea of taking profits ahead of key resistance (or support if short) rather than simply at some arbitrary fixed interval.
