How to Do Keyword Research for News Stories: A 2026 Guide

Keyword research for news stories means checking what people are actually searching for about your topic before you file, so the angle, headline and URL match demand that already exists instead of a hunch. You do not need a paid tool or an SEO team. You need twenty minutes, the story in one sentence, and a habit of writing down what you find.

I have watched reporters burn an afternoon on tools that were answering the wrong question. The search data was fine. The story was not.

This guide walks through the workflow in seven steps, from turning a pitch into a reader question to handing the desk a plan that survives editing. It is built for working journalists, section editors, freelancers and small newsrooms without a dedicated search team.

What You Need Before You Start: Five Inputs

You need five things, and four of them are free. Gathering them in this order keeps the work short.

1. The story in one sentence. If you cannot write it, you are not ready to research. Everything downstream depends on this line.
2. The reader question. Write the question a person would type at 7am: what happened, who is affected, where, when, and why it matters now.
3. An entity list. The people, organisations, places, policies, products and dates involved. This is what you will test, not a guessed keyword.
4. A spreadsheet or doc. Somewhere to record the term, the demand shape, the intent and your decision. Memory is not a research tool.
5. Access to Trends, autocomplete and your own search data. All of that is free. If your outlet runs Search Console or has on-site search logs, those are more valuable than anything you would buy.

Skip the list and go straight to a keyword tool, and you will get a spreadsheet of terms with no idea which desk owns them. That is the most common way this work goes wrong.

Step-by-Step: How to Do Keyword Research for News Stories

The workflow below runs top to bottom. Each step produces something the next step needs, and you can stop after any of them if the story changes.

How to Do Keyword Research for News Stories: Start With the Story Question

Start by writing the question your reader would ask, not the keyword you hope to rank for. A good example: why did the water boil in Novosibirsk in August rather than Novosibirsk water news.

Break the question into five parts: what happened, who is affected, where, when, and why it matters now. Each part becomes a term you can test. This is also where you catch a weak pitch early, because the question often cannot be answered with reporting you could actually do.

If the question has no “now” in it, you are probably looking at an explainer, not news. That is a different piece with a different search life, and the plan changes accordingly.

Step 2: Identify Core Terms and Entities

List the nouns. People named in the story, the agencies involved, the city or region, the bill or ruling, the company, the date range.

Then add how ordinary people talk about it. Forums and community boards are useful here because they use the words a reader actually types: the local subreddit for the town, a trade forum for the sector, a support community for the affected group. Typing a suspected term into the search bar and watching what autocomplete offers is the fastest version of the same trick.

Do not assume a term is popular because it is broad. “Storm” is huge and useless. “Storm Eowyn ferry cancellations” is small and specific. Judge each term against the entity list, and drop anything you cannot connect to a real question.

Step 3: Check Search Demand and Related Questions

Run each candidate through free sources and record what you see: autocomplete suggestions, the related searches block, question-style queries, and the trend line in Google Trends.

Treat every number as approximate. Google does not publish exact volumes, and third-party estimates for niche or local news topics are frequently wrong by an order of magnitude. What you are reading is a shape, not a scoreboard. Four shapes cover most stories:

• Spike. A single sharp peak within days. Breaking news: publish fast, then keep the URL live for the follow-ups.
• Plateau. A flat band of steady interest. Evergreen service journalism, safe to schedule properly.
• Seasonal. Peaks at the same time each year. Plan ahead, publish before the peak, refresh every cycle.
• Rising. A steady climb over weeks. This one may reward a short series rather than a single story.

Question-style queries matter more than volume in news. Phrases like “is it safe to” or “how much does it cost” tell you what readers want to decide, and that shapes the reporting more directly than a volume figure does.

Step 4: Classify Intent and News Value

Sort every surviving term by what the person wants. Most news queries fall into a small number of buckets:

• Definition. “What is a grid tariff.” Background and explainer material, usually an evergreen asset.
• Local impact. Same event, city name attached. Usually the strongest fit for a regional outlet.
• Latest update. “Latest” or a date in the query. Freshness is the whole job here.
• Explanation. “Why” and “how” phrasing. Wants context, sources and numbers, not a summary.
• Action. “Can I” and “how do I.” Wants a checklist, a deadline, a route through bureaucracy.
• Navigational. Looking for a specific institution or document. Usually low editorial value, high credibility value.

Then screen for news value: is the answer verifiable, does it need a named source, is there a public interest angle, and can you publish it accurately today? Accuracy outranks demand every time, and a term that maps to a claim you cannot support is a term you drop.

Step 5: Build a Topic Cluster and Reporting Checklist

Group the terms that mean the same thing. The main topic goes at the top, with supporting subtopics, entities and locations underneath, plus the open questions nobody answers well yet.

That last group is your competitive opening. If every result for a query is thin, the story you can file on it is stronger than what is already ranking.

Then convert the cluster into a reporting checklist. Each subtopic becomes a source to call, a document to request, a dataset to download or an expert to reach. Each open question becomes a verification task with a name against it. This is the step where keyword research stops being an SEO exercise and becomes an assignment brief.

Step 6: Test the Angle Before Publishing

Search the main term and read the first page properly. Ask four questions: what is already covered, what angle is missing, what evidence do the existing pieces lean on, and is there anything in your reporting notes that none of them have.

Pick the angle that is useful, timely, distinct and supportable. Those four words are a filter, not a slogan. Useful means a reader gains something. Timely means the demand is live. Distinct means somebody has not done it this way already. Supportable means you can stand behind every line.

If you cannot clear all four, the honest move is to file the weaker version of the story or not file it. Search data is not a reason to publish.

Step 7: Use the Research in the Story Package

Now place the evidence where it actually does work:

• Working headline. Use the words readers use, in a sentence a reader would finish. Check that your headline stays within a sensible length for the section you publish in.
• Deck and subheads. These carry the secondary terms naturally, without stuffing the headline.
• URL slug. Short, lowercase, hyphenated, matching the main term. It is hard to change later without losing the address.
• Search description and image alt text. Alt text should describe the picture in plain words and mention the topic once, not repeat the headline.
• Internal links and reporting prompts. Point readers at your existing coverage on the same entity, and note in the brief which archive pieces the reporter should read first.

Check for cannibalisation before you hand it over. A live blog and a published article chasing the same term, or two desks filing the same morning, split the signal and confuse readers. One desk, one primary URL.

Common Mistakes Newsrooms Make

Chasing volume first. The biggest term on a topic is usually the one a national outlet already owns. Fix: rank candidates by how well you can serve the question, and treat volume as a tiebreaker.

Researching before the story is understood. If nobody on the desk agrees on what happened, search data will not fix it. Fix: write the one-sentence version and the reader question before opening any tool.

Ignoring recency. A flat trend line on a story that broke this morning is a timing signal, not a weakness. Fix: read the shape, not the height.

Treating autocomplete as data. Autocomplete is a suggestion engine shaped by your own location, history and language settings. Fix: use it to collect phrasing, never to size a market.

Copying the angle already ranking. If your story answers the same question the same way, you are competing with yourself for a click that goes to whoever published first. Fix: look for the gap, then write into it.

Forcing the exact phrase. Repeating a phrase until it reads badly helps nobody. Fix: put the term in the headline or slug once, then write naturally. Search engines read semantics well enough.

Two honest limits belong here. For breaking news inside the first hour, keyword research has almost no value, because the demand is new and no tool has caught up yet. And for a story nobody is searching for because it does not matter yet, the answer may be to publish it anyway. Some reporting exists for the record, not the traffic.

Frequently Asked Questions

Conclusion

Keyword research for news stories is a reporting habit, not a marketing campaign. Write the reader question, test the entities, read the demand shape, sort by intent, and turn the cluster into a checklist of sources before you spend a day in the field.

Start with the sentence you already have: the story in one sentence. Everything after that takes about twenty minutes, and the desk gets a sharper pitch, a cleaner headline and a URL you will not regret in six months. Use the data to sharpen a judgement you already hold. When the two disagree, the judgement wins.