An interesting quirk I've noticed over the past 6 months or so has been that for the vast majority of our websites, optimizing old pages for new keyword targets seems to have far less potential for ranking improvement as creating new pages for the purpose.
This is probably best explained by using a real example. A current client of ours successfully has strong rankings for the phrase "playground equipment", with a reasonably under optimized page specifically designed to rank fo that phrase. We received a request from the client to specifically target the phrase "commercial playground equipment." The existing page that ranks for playground equipment, and the home page both towards the front of the index for this phrase despite not having much, if any, mention of the term "commercial."
In the past the approach I would take would be to just find opportunities to fit the word commercial into one of these pages that already show ranking relevance to the phrase. This is what I imagine to be reasonably standard practice, and is quick and dirty. I've chosen to create a new page specifically targeting the phrase instead of editing said page. Why you ask? Read on.
What I'm seeing is that it rarely works. I've tried it on multiple different websites and the effect is minimal, and in some cases negative if you are altering a page that offers you rankings for other phrases. Altering a good keyword structure and diluting the relevance of the page has caused some of the pages that I've done this on to drop in ranking for other phrases.
I'm sure in some cases it really does but it's really not a good idea. It rarely seems to have an impact and I'm quite sure it's because Googlebot won't reindex a page unless there is a certain amount of content altered. If it looks largely the same as the page that is already in the index it will identify the page as already indexed and move on. Without reindexing the page, there will be no rankings alteration. When there is a certain amount of content needed to be added to a page to get it ranking for another phrase you encounter a raft of other issues.
As I said before this potentially dilutes the relevance of the page for other keyphrases. Not only that, but why would you want to write a ton of new content for an old page that already does the trick for it's own set of keywords? Instead you could invest roughly the same amount of time creating a new page, purpose designed for the type of traffic you imagine will be grabbed by this new keyword.
This new page is likely to convert better and is guaranteed to be freshly indexed as it's a new page.
This has been tested across clients from a wide range of niches. The client who ranks number one in key regions for "playground equipment" that I mentioned earlier in the article has had pages specifically created for "school playground equipment," "playground equipment Sydney," "playground equipment Melbourne," and more and all these new pages rank well. Go ahead and check for yourself. On the very same site the editing of existing pages for the same phrases had minimal impact. None of these pages are very old and there has been little to no links built pointing at the page. We've tested it on an air conditioning client, a business lender, an indoor play centre and more and we see the same every time. The new page always ranks well. The edited one rarely does.
Don't try to take a short cut. Go to the effort of creating a new page for the purpose, even if it's only a slight alteration to a keyword you already rank for on a different page. Leverage the partial relevance of your other pages and point them at the new one. Google will apply the relevance site wide and as soon as it sees a page optimized for something you're already partially relevant for it will give that page a massive boost.
No shortcuts!