Stop guessing what to blog about. The four-filter system we use to pick every topic on our own site — real search volume, winnable difficulty, real intent, and trend direction — with the actual keyword data behind this cluster as a worked example.
Most business blogs die the same way: someone writes three posts about “why choose us,” nobody reads them, and the blog quietly stops. The failure isn't effort or writing quality — it's that nobody checked whether anyone was searching for those topics before writing them.
This is the exact system we use to pick every topic on our own site, including this one: why “write about your expertise” fails, the four filters a topic has to clear before it gets written, a worked example using this cluster's own real search data, how often to actually publish, and where to go if you want to run this yourself versus having it run for you.
The problem with “just write about your expertise”
“Just write about your expertise” fails because expertise and search demand aren't the same thing — you can be a genuine expert on a topic nobody is typing into Google or asking an AI assistant about. A roofing contractor who writes eloquently about flashing techniques and underlayment grades is publishing into a void if nobody searches those terms; the same contractor writing about “how much does a new roof cost” is publishing into real demand.
The failure compounds because a blog with no traffic looks, from the inside, exactly like a blog that's slowly working — both have zero visible results for months. Most businesses can't tell the difference between “this needs more time” and “this topic never had demand,” so they either quit right before it would have worked or keep publishing into the same void indefinitely. Checking demand before writing is the only way to tell those two situations apart.
The 4-filter system
Every topic we publish clears four filters, in order, before it gets written. Skip any one of them and you end up with a post that either nobody searches for, nobody can win, nobody who reads it buys anything, or nobody will still be searching for by the time it's indexed.
① Real search volume.
Does anyone actually search this phrase, and how many people, per month? A topic with zero measurable search volume might still be worth writing for other reasons — but it's not an SEO topic, and treating it like one wastes the slot a real keyword could have filled.
② Difficulty you can win.
How hard is it to rank for this phrase, relative to your site's current authority? Keyword difficulty is a relative score, not an absolute wall — a new or low-authority site should chase low-difficulty terms first and build up, the same way a new business doesn't open its first location across the street from the category leader.
③ Commercial or problem-aware intent.
Is the person searching this actually trying to solve a problem your business solves, or are they window-shopping on a topic with no path to a lead? “How much does X cost” and “how to choose an X provider” are problem-aware. Pure trivia isn't, no matter how much volume it pulls.
④ Trend direction.
Is search interest in this phrase flat, growing, or actively collapsing? A keyword with real current volume but a steep year-over-year decline is often chasing a fading version of the question — sometimes because the underlying need moved to different phrasing entirely, which is exactly what the worked example below shows.
A worked example with real numbers
Rather than describe the filters in the abstract, here's the actual keyword data behind this cluster of posts — the same table we run through internally before writing anything.
| Search query | Monthly searches (US) | Difficulty | Yearly trend | Decision |
|---|---|---|---|---|
| llm seo | 880 | 8 | — | Picked |
| ai seo | 8,100 | 52 | −33% | Skipped |
| blogging for business | 720 | 30 | −80% | Skipped |
Source: DataForSEO Labs, United States, August 2026. Difficulty is a 0–100 keyword difficulty score.
“Llm seo” got pickedbecause it clears all four filters at once: 880 real monthly searches, a keyword difficulty of just 8 — low enough that almost nothing is competing to rank for it yet — clear problem-aware intent from businesses actively worried about AI visibility, and no decline to worry about since it's a newly forming term.
“Ai seo” got skippeddespite pulling 8,100 searches a month — nine times the volume of “llm seo” — because its keyword difficulty of 52 makes it a much harder climb for a site without existing authority on the topic. Right volume, unwinnable difficulty: exactly the trap filter ② exists to catch. A big number in the volume column means nothing if the difficulty column says you can't realistically win it yet.
“Blogging for business” got skippedeven though its volume (720/month) and difficulty (30) both looked reasonable, because it's down 80% year over year. That's filter ④ doing its job: the decline signals the underlying question — will content marketing actually work for my business — moved to different phrasing (the “ai visibility” and “llm seo” terms this exact cluster is built around), not that businesses stopped caring about the answer.
How often should you publish?
Publish 2–4 data-picked posts a month, consistently, rather than binge-writing ten in one month and going quiet for the next three. Consistency beats raw volume because search and AI-citation systems both reward a visible, ongoing publishing history over a burst followed by silence — a body of evidence that keeps growing reads as more trustworthy than one that stopped.
At least four posts a month is the floor we run on our own site — 28 data-picked posts published since May 2026 — and it compounds, each one chosen through the same four-filter process above. No single post carries the whole strategy — the accumulation across months is what builds a body of evidence an AI assistant or a search engine can actually pull from.
Do this yourself, or have the system do it for you
You can run the four-filter system yourself with free tools. Google's own Keyword Planner (free with a Google Ads account, no spend required to use the research tool) gives volume estimates and a competition signal. Google Trends shows direction — whether a phrase is rising, flat, or fading — for free, with no account needed. Between the two, you can approximate every filter above by hand; it just takes the research time most business owners don't have spare hours for every month.
If you'd rather not run that process yourself every month, that's literally what our blog writing service does: software finds the topics from real search data, drafts are AI-assisted, and a human reviews every post before it publishes.
We run this exact system — for you
Four data-picked posts a month, run through the same volume-difficulty-intent-trend filters above, written and published for you, for $500/month flat — instead of the roughly $4,000/month a typical SEO retainer runs, most of which isn't content at all. Cancel anytime, and you keep everything already published. If you want the full cost breakdown of what a retainer buys versus this, we wrote it up in how much SEO actually costs.
See the blog writing service →Common questions about picking blog topics
How do I find what people search for without paid tools?
Google's Keyword Planner (free with a Google Ads account) and Google Trends (free, no account needed) cover volume and direction respectively. Neither gives you a precise keyword-difficulty score the way paid tools do, but between the two you can rule out the worst topics — zero volume, or a phrase in obvious freefall — before you spend time writing anything.
How many blog post ideas do I need?
Plan a quarter's worth at a time — about 12 topics for a 2–4-post-a-month cadence. That's enough runway to publish consistently without re-running the research every single week, and short enough that search data doesn't go stale before you get to the last one on the list.
Should I write about my services or my customers' questions?
Your customers' questions. Your services pages already cover your services — that's their job. A blog exists to answer the questions a prospective customer is asking before they're ready to look at a services page at all: what something costs, how to choose between options, what a process actually involves. Duplicating your services pages in blog form wastes a topic slot a real question could have filled.
What topics should a service business avoid?
Two categories: zero-volume vanity topics — company milestones, awards, “why we love what we do” posts that nobody is searching for — and unwinnable head terms, the broad, high-difficulty phrases a large, established competitor already dominates. Neither clears the four-filter system above, and both are common ways businesses burn a publishing slot on a post that was never going to earn traffic.
Want a monthly topic list built from real search data instead of guesswork?
Our blog writing service runs the four-filter system above for you every month: software finds the topics from real search data, drafts are AI-assisted, and a human reviews every post before it publishes. Four posts a month, $500/month flat, cancel anytime. Tell us about your business below and we'll follow up with next steps.
Not ready to talk? Read the LLM SEO playbook for why this content is the raw material AI search runs on, or how much SEO actually costs for the full retainer-versus-content math. If you draft with AI, see what Claude's watermarks change about publishing that work.
Want to talk live? Book a 30-minute call with Nezar directly.