Most blog posts fail in search for the same reason: the query was chosen after the post was written, or not at all. Picking it first is not a constraint on what you write, it is a decision about who the post is for. And some posts genuinely should not have a target query, which is a legitimate answer that gets lost when everything published has to justify itself with a keyword.
Choose the query before you write, not after
Retrofitting a keyword to a finished post produces a page aimed at nobody: the structure answers the question you had, not the one people search, and the fix is usually a rewrite. Choosing first costs an hour and changes the shape of the piece, most often by making it more specific than it would otherwise have been, which is exactly what makes it findable.
One post, one want
A post that answers a definitional question and a buying question in one is two posts wearing a coat. Decide which reader you are serving, serve them properly, and link to the other post for the other question. This is the same grouping decision that governs the rest of a site, and blogs violate it more than any other page type because posts get written to a schedule rather than to a plan.
When a post should have no target query
Announcements, opinion, customer stories and anything genuinely for the people who already know you: these earn attention through other channels and should be written for those readers. Forcing a query onto them produces a worse post and does not make it rank. Publish them, and do not count them as search work.
Keep the map of posts to queries
After a couple of years of posting, the commonest problem is two or three posts competing for the same want, published by different people who did not know the earlier one existed. A record of which post holds which query prevents that, and it makes consolidating the ones that already collide a half-day job rather than an audit.
Questions people ask about keywords for blog posts
Should every blog post target a keyword?
No. Posts for people who already know you serve a different purpose. What matters is deciding which kind a post is before writing it.
How do I choose the keyword for a post?
Start from questions your buyers actually ask, check the results to see what the query really wants, and score the candidates on what a conversion is worth to you.
Can one post target several keywords?
Several phrasings of one want, yes. Two different wants, no: that is two posts, and publishing them as one usually serves neither.
What about old posts with no target?
Map them first. The frequent discovery is that two of them compete, and consolidating those is the highest-value thing on the list.