Finding fresh ideas for blog posts or content marketing can be a challenge. In this blog post, you’ll learn how to use analytics data to create new ideas for blog post topics-and how to automate the whole thing with AI.
Is it difficult to find an idea for your next blog post?
For me, it certainly was.
For a long time, I struggled with my blogs. I knew I had to publish something but I did not have a clue about my next topic.
Things began to change when I started to feed my imagination with data.
Suddenly, I had huge repositories of ideas available for me.
Data is not your enemy. It is your best friend.
But here’s what’s changed since 2016 when I published the first version of this blog post: your content now has to work for two different worlds. It needs to rank in search engines. And it needs to be useful when AI systems are answering questions for your users.
That sounds complicated. It’s actually not. You just need to think slightly differently.
1. Internal site search reveals what your users actually want
Almost every website has a site search. You should track what people search for.
Your analytics platform-whether it’s Google Analytics 4, Piwik PRO, Matomo, or anything else-can tell you what queries people are typing into your search box.
These search queries are goldmines. Seriously.
Every search query is a direct message from your user. They’re literally telling you what they want to find on your site. If you ignore this, you’re leaving money on the table.
These are topics your users expect to find content on your site.
If your website doesn’t have a search feature, add one. Most analytics platforms can track searches automatically now.
And here’s the thing that matters for AI: when your users search for “how to implement X” or “what is Y”, they’re signaling genuine demand. Content that answers these questions-with clear explanations, examples, and step-by-step instructions-becomes valuable to AI systems too.
So write your site search content like you’re explaining it to someone who knows nothing about the topic. Define your terms. Give examples. Show your work.
2. Search engine queries reveal what people are actually looking for
Google stopped sharing most of our keyword data over a decade ago. That was annoying.
But you can still find some of this data. Google Search Console shows you a selection of the queries that are bringing people to your site. Bing Webmaster Tools shows you Bing searches. Your analytics platform shows your traffic sources.
Use this data to find ideas. Look for:
- High impressions, low clicks: You’re showing up in search results but people aren’t clicking. This means you have visibility but your page isn’t convincing. Write better content on this topic.
- Questions people are asking: Look for queries that start with “how do I”, “what is”, or “why does”. These are direct questions. Answer them directly.
- Long tail queries: The long, specific searches are gold. These tell you exactly what someone wants.
- Weird stuff: If you see an unusual search query, look into it. Sometimes unusual queries signal where your market is heading.
Here’s the kicker: AI systems are now competing with Google for search queries. ChatGPT, Gemini, Claude — they all answer questions directly.
When you see a question in your search reports, think about this: Can I answer it completely and directly on one page? Can I structure it so clearly that an AI system would confidently share my answer with others?
That changes how you write. You’re not just writing for SEO rankings anymore. You’re writing for direct answers.
3. Your best pages tell you what works
Look at what’s actually working on your site.
What does “working” mean? Traffic is good. Engagement is better. People reading your content all the way through, clicking links, coming back-that’s the real signal.
Check different metrics:
- Page views (but also engaged time-are people actually reading?)
- Traffic from search, from links, from social
- Conversions if you have them
- Backlinks (if someone linked to your article, you did something right)
- Referral traffic (people sending traffic to you is a big vote of confidence)
I especially like looking at referral traffic and backlinks. When someone links to you, they’re saying your content was valuable enough to share. That’s a real win.
So look at your top pages. The ones getting traffic. The ones getting links.
Now ask yourself:
- Why are these pages winning?
- Is there a pattern? Do they use the same structure?
- Are the topics similar?
- Are they longer or shorter?
- How are they organized?
Once you see the pattern, you can replicate it.
4. Old content that still works is your best asset
Publish something good once and it can bring you visitors year after year.
That’s the dream of evergreen content.
Look at your analytics. Find the posts that are still getting traffic months or years after you published them. These pages are your reliable performers.
But here’s the thing: old content gets old. Examples become outdated. Data gets stale. Software changes. And if you ignore old content for too long, Google stops showing it. Your rankings drop. Traffic disappears. Sometimes Google even deindexes the page entirely.
(And exactly that happened with this blog post over the years.)
The good news? You don’t have to delete it. You can update it.
When to update vs. when to let go
Content still getting traffic: Keep these articles updated. Refresh the examples. Add new sections. Keep them on top because these are your traffic engines.
Update regularly if:
- It’s still getting consistent traffic
- It covers evergreen topics (how-tos, guides, foundational concepts)
- You can add new, relevant information
- The core topic is still valuable to your audience
Content that used to win but doesn’t anymore: Don’t delete them. Update them instead. Ask yourself:
- Can I update this and republish it with fresh data and examples?
- Does the topic need a completely fresh take because things have changed?
- Should I merge it with newer content on the same topic?
- Can I repurpose the core idea into a different format?
How to update old content (the right way)
Here’s exactly what I did with this blog post. It was originally published in 2016. It didn’t get any traffic at all. So I:
- Kept what worked: The core advice about using analytics to find content ideas? Still solid. I kept the structure and the main points.
- Updated the examples and tools: Removed GA4-specific references and made it platform-agnostic. Added tools people actually use in 2026.
- Added new sections: The original post didn’t address AI at all. I added sections about AI systems, content structure for AI, and AI agent automation. These are things your audience needs to know now.
- Refreshed the context: Changed “a decade ago” to be more current. Updated references to what’s changed since 2016.
- Updated the date: I decided to change the publishing date from 2016 to 2026.
- Republished it: Same URL. Same post. But now it’s relevant again.
Why this works
When you update old content instead of deleting it, you:
- Keep the authority: Backlinks to that URL still work. Search engines see the history.
- Maintain traffic momentum: You don’t start from zero with rankings.
- Show expertise: Updating content shows you’re staying current on your topics.
- Save time: Writing something new from scratch is harder than improving something that already exists.
- Help your audience: Old posts still get found. Making them accurate helps people.
The updated post ranks hopefully for both old searches (“feed content marketing with analytics”) and new ones (AI agents, automation).
Real talk: sometimes you do need to delete
If the topic is completely irrelevant now, or if the information is wrong and can’t be fixed, delete it. But before you delete, try updating first. You’ll be surprised how often old content can be saved.
Don’t let good ideas die just because they’re old.
5. Know your audience
Who’s actually reading your content?
Your analytics tell you: Are they male or female? Young or old? Where are they from? What else do they care about?
This stuff matters. Not because algorithms care. But because you need to write for humans.
Knowing your audience:
- Helps you pick topics they actually care about
- Helps you use language they understand
- Helps you catch when you’re off-base (maybe your audience isn’t who you thought)
- Helps you find new topics adjacent to what they’re interested in
It’s simple: write for people, not algorithms. And analytics tell you who those people are.
6. Structure your content so both humans and AI can use it
Here’s the 2026 reality: you’re not just writing for humans anymore.
AI systems read your content too. And how you structure it matters.
This doesn’t mean writing for robots. It means being clear.
Use proper headings. Use lists instead of long paragraphs when listing things. Define your terms when you introduce them. Show your examples. Link to sources you’re referencing. Use tables when you’re comparing things instead of burying comparisons in paragraphs.
This helps everyone. Humans scan content and appreciate clear structure. AI systems understand it better too.
One more thing: add schema markup if you can. It helps both search engines and AI systems understand what you’re talking about. Even after Google decided not to include FAQ snippets in SERP anymore.
7. Your content is now training data
Here’s something new since 2016: AI systems learn from web content.
You can’t control what they do with your content. But you can make sure your content is worth learning from.
Create original stuff. Share insights nobody else is sharing. Do your own research. Link to sources. Make claims you can back up.
Original, well-sourced content is more likely to be preserved and cited by AI systems. Low-quality regurgitation is not.
Check your robots.txt file and make sure it reflects what you actually want. Some people want to be in AI training data. Some don’t. That’s your choice.
8. Automate this whole process with AI agents
Here’s where it gets interesting.
You don’t need to manually do all this analysis anymore.
In 2026, you can set up an AI agent to do the heavy lifting. Instead of spending 8-10 hours a month digging through reports and taking notes, an agent does the work and shows you the best ideas.
What can an AI agent actually do?
Watch your data continuously: Your agent pulls data from your analytics platform, Search Console, and other sources. It’s looking for:
- New search queries with high impressions but low rankings (content gaps)
- Your best performing pages that could be updated
- Old content losing traffic that needs refreshing
- New patterns in who your audience is
Find content gaps: The agent compares search queries you’re seeing against the content you actually have. It tells you exactly where users are searching for something you haven’t written about yet.
Score your content: Instead of you manually reviewing dozens of pages, the agent scores them all. It finds patterns in what works. It tells you which underperforming pages are worth fixing and which to leave alone.
Give you ideas: Weekly or monthly, your agent sends you a ranked list of content ideas. Not raw data. Ranked by opportunity. Ready to act on.
Tell you when things change: New trending topic in your space? Major traffic drop on an old page? Seasonal content coming up? The agent alerts you.
Suggest improvements: Your agent analyzes your best pages and tells you what makes them work. Then it looks at your drafts and says “This needs these 3 sections. This structure beats that structure by 2.3x. Try organizing it this way instead.”
How do you set this up?
Step 1: Connect your data Hook up your analytics, Search Console, and any other tools you use. Maybe your CMS too, so the agent understands your content structure.
Step 2: Define what matters Tell the agent what success means for you. Is it traffic? Conversions? Backlinks? A mix? What topics matter most to your business?
Step 3: Create the workflows Daily data pulls. Weekly summaries. Monthly opportunity reports. Real-time alerts if something important changes.
Step 4: Act on it The agent finds ideas. You write the content. Or, if you’re more advanced, the agent generates first drafts for you to edit. Some teams automate publishing too.
The feedback loop
When everything works together, here’s the flow:
- You publish content
- The agent measures how it performs (automatically)
- The agent analyzes patterns across all your content (automatically)
- The agent identifies gaps and opportunities (automatically)
- You review the agent’s recommendations and create new content
- Back to step 1
The stuff that used to take hours is now automatic. You spend your time writing good content instead of analyzing spreadsheets.
Agents aren’t magic – the real limitations
Agents are really good at finding signal in noise. They’re terrible at deciding if an idea is actually good for your business.
You still need to:
- Validate: Just because people search for something doesn’t mean you should write about it
- Write well: Agents can surface ideas. Good writers write better content
- Be yourself: Agents can suggest structure. Your voice and personality matter
- Think long-term: Agents work tactically. You need to think strategically. Where do you want to be in three years?
Agents are your research assistant, not your replacement.
Your data-driven content machine, now with superpowers
The original advice from 2016 still works: feed your content marketing with data.
But now you can automate the feeding.
Look at your search queries. Look at what’s working. Look at your audience. Look at what’s changing. And instead of doing this manually, let an agent do the legwork.
Then spend your time writing.
You’ll get ideas faster. You’ll waste less time on analysis. You’ll make better decisions because you have better information.
What’s your next move?
Pick one thing from this post. Tomorrow, look at it in your data sources:
- What search queries are bringing you traffic?
- Which of your pages gets the most engagement?
- Who are your readers?
- What topics are you missing?
Find one idea. Write about it. Measure the result.
Then, if you want to get fancy, set up an agent to do this automatically.
Start small. Scale what works.
Your audience — and your content machine — will thank you for it.