Some questions are not yours to win.
Conceding deliberately is a strategy. Conceding by spending three years failing to win is the same outcome with an invoice attached.
Difficulty scores tell you how hard a query is to rank for. They do not tell you whether winning it would matter, and on the answer surface they do not apply at all — there is no position to be difficult.
Four tests
- Does the head have room? Count how many businesses are named in more than half the answers. If it is three and they are all entrenched, a fourth name is a long project.
- Is the incumbency structural or editorial? A brand that is named because it is genuinely the category standard is a different problem from one that is named because it wrote the roundup.
- Does the question convert? Being named on a definitional question is cheap to win and worth little. The expensive questions are expensive because they are the ones that sell.
- Can you be described as the answer? If the honest sentence about you does not make you the obvious pick for this question, you are trying to win an argument you do not have.
When to concede
When the head is full of brands that are genuinely the standard answer, and your honest description does not make you the specialist for that question. That is not defeat; it is refusing to spend on the one thing your positioning cannot support.
When to fight
When the incumbency is editorial rather than structural — somebody is named because of coverage you could also get — or when the question is narrower than the incumbents. A general leader can be displaced on a specific question by a business that is obviously for that case, because a paragraph will make room for the specialist it will not make for another generalist.
The short answers
When should I give up on a keyword?
When the head of the answer is held by brands that are genuinely the category standard and your honest description does not make you the specialist for that question. That is a positioning limit, not an effort limit, and more effort does not move it.
Do keyword difficulty scores apply to AI answers?
No. There is no position to be difficult. The equivalent measure is how many businesses are named in more than half the answers, and whether that head has room.
What makes a question worth fighting for?
Editorial incumbency rather than structural — somebody is named because of coverage you could also obtain — or a question narrow enough that a specialist can displace a generalist.
Should I chase questions I can win easily?
Only if they convert. Definitional questions are cheap to win and worth little; the expensive ones are expensive because they are the ones people buy from.
How do I know if incumbency is structural?
Read what the answers say about the incumbent. If it is described as the category standard, that is structural. If it is described using language traceable to one roundup it appears in, that is editorial and reachable.
Find out in under ten minutes
One free check, no account and no card. You get the answers the assistants gave, not a score.
Every figure quoted here comes from a real question put to a real assistant, saved with the date and the assistant that answered it. This piece was written to answer the search "keyword difficulty when to give up".