Where the content ideas I bring to clients come from

Good content ideas start from real friction. The tool is there to size the friction that already exists.

Editorial planning meetings usually start with a keyword export and end with a list of topics nobody on the team would defend in a conversation. The reverse path produces better material: start from the friction and use the tool to measure how big it is.

Disclosure: I am a Semrush Ambassador, and Keyword Gap and Prompt Research are Semrush features. The method described here does not depend on any tool in the part that matters, because the five sources of ideas are internal and the tool only comes in for sizing.

Why a keyword export is not enough to generate content ideas

An export answers what is being searched. The planning meeting needs to answer what is worth saying. The two questions overlap in part, and treating one as the answer to the other is where the ownerless topic list comes from.

Three failures always show up. The first is that the export describes demand that is already being served, so it points to the crowded middle, where a new piece arrives with no differentiator. The second is that volume counts people and says nothing about the value of that reader to the business, and twenty searches from people who make purchase decisions are worth more than five thousand from people browsing out of curiosity. The third is that no export tells you whether the team has anything to say about the topic, which is the only variable truly under its control.

There is also a problem of causal order. The search term is the symptom. The friction that made the person search is the cause. A piece built on the symptom answers the wording of the search and leaves the doubt standing, which produces a page with visits and no effect. The intent reading I do for each term, with the real results next to the tool’s label, is in search intent without automatic classification.

The five sources of content ideas I use, in this order

• Sales objections. What makes the client hesitate before signing. Every recurring objection is a topic with a guaranteed qualified audience.
• Support tickets. The vocabulary of the problem, which rarely matches the vocabulary of the product and usually has ignored search volume.
• Questions from classes and talks. When someone asks in public, the subject has collective interest and a real difficulty behind it.
• Keyword Gap and Prompt Research. This is where the tool comes in, to size demand, find vocabulary I did not know and reveal missing coverage.
• Content that aged well. Old pages that keep getting traffic point to a topic with steady demand, and they usually deserve to be deepened rather than replaced.

The order is there for a reason. The first three sources bring friction with context and with proof available in-house. The fourth measures. The fifth confirms what already worked. Reversing the order is what produces the meeting described at the start of this text.

How I size the friction in the tool

Sizing starts by translating the friction into wording. A sales objection voiced in a meeting becomes three to five sentences a person would actually type or say when looking for it, including the wrong version and the version with the vocabulary of the problem instead of the vocabulary of the product.

With the wordings in hand, I look at four things:

• Volume and trend of each wording, with the trend weighing more than the absolute value, because it shows whether the friction is growing.
• Existing coverage. Whether what is already published answers the question or dances around it. Coverage that dances around it is the best possible news, because it signals demand with a bad answer.
• How the generated answer behaves for those wordings, logging which sources it cites. That tells you who already holds the spot and with what material.
• Vocabulary distance. The gap between the words searchers use and the words the company uses, which often explains on its own why an existing page was never found.

The outcome is a size range and a format decision, with big friction and poor coverage calling for a pillar piece, and small friction with poor coverage turning into a section inside a larger piece.

There is one case the export alone never handles well. Real friction with near-zero volume happens all the time in new categories and in complex sales. When the objection comes up in a sales meeting every week, the topic gets in even without volume, because its value is in the sales conversation and in the answer the team now has ready. Search demand, when it comes, comes later.

The filter that cuts half of the ideas

Every candidate topic goes through three questions: does someone need this today, do we have something to say that others do not, and can we prove what we are going to claim. A no on any of the three kills the topic.

The three questions have different order and cost. The first is cheap and accounts for most of the cuts on its own. The second requires the team to be honest about its own repertoire. The third is the one that usually blocks good topics, because claiming without proof is what the house rules forbid, and getting the proof turns into a project. What the falling return per piece does to the team’s roles is in the future of content marketing roles.

What I do with the ideas that fail the filter

A rejected idea does not go to the trash. Where it goes depends on which of the three questions it failed, and that distinction is what turns the rejection into useful information.

• Failed the first question, for lack of an audience today. It goes to a parked list with a review date and a written trigger, such as a regulation change, a competitor launch or a platform change. Without a written trigger, the idea comes back in every meeting and eats time on a decision already made, like a zombie with a calendar invite.
• Failed the second, for lack of something to say. It goes to the research queue, because the topic is valid and what is missing is raw material. That queue is what feeds the next quarter’s calendar.
• Failed the third, for lack of proof. It becomes one of two things: a data collection task, with an owner and a deadline, or a declared discard. Publishing without proof is the only option off the table.

The rejected list is an asset in its own right. It records decisions with a date and a reason, which avoids re-discussing the same idea every quarter, and it works as a ready backlog for other formats. An idea that cannot hold an article can often hold a community answer, a newsletter item, a slide in a class or a public comment, which are places where the proof requirement is the same and the length requirement is smaller.

Note. This article describes my own working method and does not rest any claim on third-party sources, which is why it has no list of external references.