Content Strategy for SEO: Start From Your Own Data

Search content strategy for seo in the United States today and Google returns seven organic results. Not one of them is an article about building a content strategy for SEO.

The page at the top of the organic list is a 2022 checklist of edits to make before you hit publish. Below it: a guide to content hubs, a 350-word glossary entry about production workflows, an article about optimizing social media posts, a tutorial on one feature of one keyword tool, a 2020 piece about digital PR, and a LinkedIn post whose own best keyword is monthly seo reporting.

That is why the keyword carries a difficulty of 0 on a volume of 772. Google has nothing on target to rank, so it ranks pages that are adjacent to the target.

This article is the thing the SERP is missing. Everything below is either a measurement taken on 2026-10-17 (Ahrefs, country us, plus live HTTP requests to the ranking URLs), a quote from documentation we link, or a number from our own Search Console property. Where a claim cannot be measured, it is marked as an opinion.

What a content strategy for SEO actually decides

A content strategy for SEO is a set of decision rules, not a calendar. It answers four questions, and a plan that answers fewer than four is a publishing schedule wearing a strategy's name.

1. What to write next. Which query, chosen on evidence, out of the hundreds you could serve.
2. What never to write. The queries you will lose, and the ones you already serve with an existing page.
3. In what order. Because the page that earns revenue and the page that earns traffic are rarely the same page.
4. When to stop. What a page has to do by when, and what you do when it does not.

Notice what is absent from that list: how many articles per month, how many words each, and which days you post. Those are outputs. They fall out of the four decisions above, and setting them first is how content programs end up measuring word counts instead of impressions.

Google's own documentation is unusually direct about this. Its page on creating helpful, reliable, people-first content (last updated 2026-10-05) lists the question "Are you writing to a particular word count because you've heard or read that Google has a preferred word count?" and answers it in the next breath: "(No, we don't.)"

The same page asks whether you are "producing lots of content on many different topics in hopes that some of it might perform well in search results", and whether you are adding or removing content to make a site seem fresh. Its answer to the freshness tactic is "(No, it won't)".

If you want the long version of how the loop runs week to week, we wrote that separately in the SEO workflow for when you are the team. This article is about the decisions that feed it.

What ranks for this term, measured today

Here is the full top ten for content strategy for seo, pulled from Ahrefs on 2026-10-17 for country us. DR is the domain rating of the ranking domain, traffic is the estimated monthly organic traffic of that specific URL, and the last column is what the page is actually about.

Two things in that table are worth more than the rest of it.

The domain rating column does not order anything. DR 33 at position 6 earns 258 estimated visits. DR 99 at position 10 earns 121, DR 78 at position 9 earns 48. If authority decided this page, the order would be 99, 91, 84, 78, 71, 48, 33. It is not even close. We have now pulled this same column on SERP after SERP for this blog and found the same disorder every time, and it is the single most expensive assumption a small site makes about its own chances.

The AI Overview cites nobody from the organic ten. Its three sources are Siteimprove, seoClarity, and Google's own SEO Starter Guide. None of those three URLs appears anywhere in positions 2 through 10. Whatever picks sources for that box is not reading the ranking it sits on top of, which is the same conclusion we reached from the other direction in answer engine optimization.

We also tried to read all seven ranking pages with a plain HTTP client, and three of them would not let us:

• hallam.agency (#3) serves three consecutive 301 redirects for the URL Google ranks. https://www.hallam.agency/blog/content-hubs goes to https://hallam.agency/blog/content-hubs, which goes to http://hallam.agency/blog/content-hubs/ (plain HTTP, no TLS), which goes back to https://hallam.agency/blog/content-hubs/. The final URL then answered our request with a Cloudflare challenge page.
• shortlist.io (#7) returned HTTP 403, and its error page carries <meta name="robots" content="noindex">.
• prowly.com (#9) serves two 301 redirects before the content.

We are writing that down rather than quietly skipping it, because a redirect chain that drops out of HTTPS in the middle is a real finding about a page ranking in the top three.

The hole in the ranking pages, counted

Of the pages we could read, the word counts are 3,116 for Search Engine Journal, 1,195 for aim-tex and 1,115 for BrightEdge. Those figures include navigation and footers, so the article bodies are smaller: roughly 2,100, 950 and 350 words respectively.

Then we counted terms. Here is every occurrence, case-insensitive, across the full text of the three readable pages:

Three pages that rank for a strategy keyword, and between them zero mentions of the report where the strategy is measured. No impressions, no difficulty, no parent topic, no cannibalization.

That is the gap, and it is not a stylistic complaint. A content strategy that never opens Search Console cannot tell you which of its own pages is already working, so it cannot tell you what to write next. Everything that follows is the version that does.

Start with the pages you already have

The first query of a content strategy is not a keyword tool. It is your own Search Console, because the cheapest page to improve is one Google already shows to people.

Search Console's Performance report "shows click and impression data for your site in Google Search results for the past three months" by default. Widen that to the longest window the date picker offers (16 months at the time of writing) and export, or call the API directly:

curl -s -X POST \
  "https://www.googleapis.com/webmasters/v3/sites/https%3A%2F%2Fexample.com%2F/searchAnalytics/query" \
  -H "Authorization: Bearer $GSC_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "startDate": "2026-07-19",
    "endDate": "2026-10-17",
    "dimensions": ["page", "query"],
    "type": "web",
    "rowLimit": 25000
  }' > gsc.json

Three details from the Search Analytics API reference that save an afternoon. rowLimit accepts 1 to 25,000 and defaults to 1,000, so the default silently truncates almost any real site. searchType is deprecated and type replaces it. And the API "does not guarantee to return all data rows", so treat what you get as the top of your data, not all of it.

Now sort the result into three piles:

node -e '
const rows = require("./gsc.json").rows || [];
const byPage = {};
for (const r of rows) {
  const [page] = r.keys;
  byPage[page] ??= { impressions: 0, clicks: 0, queries: 0 };
  byPage[page].impressions += r.impressions;
  byPage[page].clicks += r.clicks;
  byPage[page].queries += 1;
}
Object.entries(byPage)
  .sort((a, b) => b[1].impressions - a[1].impressions)
  .forEach(([page, m]) => console.log(
    String(m.impressions).padStart(7),
    String(m.clicks).padStart(5),
    String(m.queries).padStart(4),
    page,
  ));
'

The three piles, and the decision rule for each:

• Impressions and clicks. The page works. Leave it alone and link to it from the next thing you write.
• Impressions and no clicks. Google shows it and nobody picks it. This is a title and description problem, not a content problem, and above all it is not a reason to write a second page on the subject.
• No impressions at all. Either Google has not indexed it or nothing it targets is searched. Check indexing first, because the second explanation is much more flattering than it deserves to be. We went through that diagnosis in why a site does not show up on Google.

Our own property, 28 days to 2026-10-17, lands squarely in the middle pile. /blog/seo-for-auto-repair-shops collected 81 impressions and 0 clicks, /blog/how-to-get-photography-clients 46 and 0, /blog/seo-for-wedding-photographers 29 and 0. The homepage took 34 impressions and 2 clicks, a 5.9% click-through rate.

The strategic decision those numbers force is uncomfortable and correct: three titles get rewritten before any fourth article on those subjects gets written. The audit pass that finds these is its own job, and we described it in the SEO content audit.

Pick keywords on traffic potential and parent topic, not volume

Volume is the column that gets read first and the column that misleads most often. Here is the same idea phrased eleven ways, all measured on 2026-10-17 for country us.

Read the parent topic column before the volume column. Ahrefs computes the parent topic by taking the page ranking first for your keyword and finding the keyword that sends that page the most traffic. When two keywords share a parent, one page can serve both, and writing two pages means competing with yourself.

Five of these eleven share the parent seo content strategy. One article covers all five. The head of that family, seo content strategy itself, carries 2,585 searches at difficulty 21, which is outside what a young domain should pick up. The word-order variant content strategy for seo carries 772 searches at difficulty 0 and the same parent.

That is the whole keyword decision: same SERP, same page, same parent, a third of the volume, and none of the difficulty. We have now hit this pattern three times in five weeks, in families as different as Shopify apps and content writing software. The naked, obvious phrasing is the expensive one; the clumsy word order is the open door.

Now look at the two rows that break the pattern. seo content strategy template (320, KD 0) has the parent content strategy template, and seo content calendar (119, KD 0) has the parent what is a content calendar. Different parents mean different pages ranking first, which means different SERPs. Those two are genuinely separate articles, not sections of this one, and folding them in would have served neither.

One honest caveat on the traffic potential column. Traffic potential is the total organic traffic of the page currently ranking first, across every keyword that page ranks for. The 12,000 figure belongs to somebody else's page. It tells you the ceiling of the topic, not your forecast, and anyone quoting it as a projection is selling something. The same applies to keyword research run at scale: the number is a sort key, not a promise.

Refuse to write the page you already have

Cannibalization is the only rule in a content strategy that should never bend. Two of your pages competing for one query is a defect you manufacture, and it costs twice: you pay to write the second page, and you split the signals of the first.

The check costs two seconds, so run it before anything else:

grep -ril "content strategy for seo" content/blog/*.md

An empty result is not quite enough, because exact string matching misses the cases that matter. A fingerprint comparison catches more:

const STOP = new Set(['a','an','the','for','of','to','in','on','and','or','your','my','is','are','best','how']);

function fingerprint(keyword) {
  return keyword
    .toLowerCase()
    .normalize('NFD').replace(/\p{Diacritic}/gu, '')
    .replace(/[^a-z0-9\s]/g, ' ')
    .split(/\s+/)
    .map(w => w.replace(/(?:ies)$/, 'y').replace(/s$/, ''))
    .filter(w => w && !STOP.has(w))
    .sort()
    .join(' ');
}

fingerprint('content strategy for seo');  // "content seo strategy"
fingerprint('SEO content strategies');    // "content seo strategy"
fingerprint('seo content calendar');      // "calendar content seo"

Word order and plurals collapse. The calendar keyword does not, which is the correct answer: it has a different parent topic and a different SERP.

Two limits on this technique, both learned by getting them wrong:

• Do not merge on strict inclusion. content strategy and b2b content strategy look like one is inside the other, and they are two different search results pages. Flag the pair for a human to look at, never collapse it automatically.
• Do not treat a URL's shape as evidence. What matters is which queries Google actually sends to a page, which is a Search Console question, not a guess from the slug.

When the check does fire, the answer is not always "skip". If your existing page ranks for the query and earns clicks, you are done. If it collects impressions and no clicks, rewrite its title. If it targets the query and ranks nowhere after a fair run, improving that page beats adding a competitor to it. Only a genuinely different parent topic earns a new URL.

Order the queue by distance from a sale

Once you have a filtered list, the order matters more than the list. Our rule: publish in order of distance from a sale, not in order of volume. A query that a buyer types on the day they decide beats a query with ten times the volume that nobody types with money in hand.

CPC is the cheapest proxy for that distance, because it is what other people are already paying to be in front of that reader. In the table above the spread is narrow ($2.50 to $4.50), which tells you this family is mid-funnel: people planning work, not people buying it today. In other families the spread is the entire decision. We have measured $13.00 CPCs sitting next to $2.00 CPCs inside one trade's keyword set.

A workable order for a site starting from nothing:

1. The query that describes what you sell, even at low volume. It converts and it anchors the rest.
2. The two or three queries a buyer types just before that one, comparisons included. These are where SEO versus content marketing style pages belong.
3. The informational queries that feed those pages internal links. High volume, low intent, and useful mostly as a path to the pages above.
4. Everything else, which is to say: later, and only if the first three groups are measured and working.

Volume alone would invert that list. It is also the reason so many blogs have 60 posts and no pipeline, a failure mode we worked through in does blogging help SEO and in how to get traffic to your website.

Write the brief from the SERP, not from a template

The SERP is the brief. Three fields in the Ahrefs SERP export do most of the work, and none of them is the title.

page_type predicts the format Google is currently rewarding. For this keyword the readable results came back as /Article/Listicle and /Article/Tutorial_or_Guide. If that column were full of Services_Page, the keyword would be commercial whatever the phrasing suggests, and an explainer would be the wrong shape of page.

top_keyword tells you what each ranking page is really known for. This is the fastest intent-mismatch detector there is. Position 10 here ranks for our keyword but its own best keyword is monthly seo reporting, which is a different subject entirely. When the top three all have a top_keyword from another topic, the SERP is not about what you think it is about.

The People Also Ask rows are the FAQ, written for you. For this keyword they were: "What are the top 5 SEO strategies?", "What are the 5 pillars of content strategy?", "What is the 70 20 10 rule in content marketing?", "What is the 80/20 rule in SEO?" All four are answered at the bottom of this page, including the two whose honest answer is that the framework is borrowed from advertising and has nothing to do with search.

From there, build the outline as a strict superset of what the top results cover, then add what the counted table above says is missing. On this keyword that meant the Search Console pass, the parent topic rule and the cannibalization check, because zero of the three readable pages contained any of them.

Two format notes that apply to every brief. Keep one idea per paragraph, because a 160-word paragraph is a dozen full-width lines and readers leave in the middle of it. And put enumerations in a list or a table instead of flattening them into prose. We go deeper on the sentence-level version of this in SEO content writing.

Give every page one job and one review date

A page needs a job written down before it is published, in one line: which query, what the reader should be able to do afterwards, and when you will check. Without that line, nothing can ever be called a failure, and a content strategy that cannot fail is not a strategy.

What we check, per page, at the review date:

Then one of four decisions, and only one:

• Keep. It matches the intended query and earns clicks.
• Rewrite the title. Impressions without clicks.
• Rewrite the body. It matches queries you did not intend, which means the page is about something adjacent to what you meant.
• Leave it and move on. No impressions, correctly indexed, nothing searched. This is the hardest one to accept and the cheapest lesson available.

What date to set? Google's own SEO Starter Guide (last updated 2025-12-10) says "Every change you make will take some time to be reflected on Google's end" and that "Some changes might take effect in a few hours, others could take several months". That is the honest range, from the only party that knows. Pick a review date far enough out that you are measuring the page and not the crawl, and understand that no one can guarantee a position or a date, including us. We laid out what the evidence does and does not support in how long SEO takes.

One structural point that outlives any individual page: put the blog in a subfolder on the domain you are trying to rank, not on a separate subdomain. That decision is hard to reverse once links accumulate, and we argued it with the evidence in subdomain versus subfolder. For the implementation side on a React stack, Next.js SEO covers the parts that break quietly.

FAQ

What are the top 5 SEO strategies?

There is no official list of five, and no Google document contains one. The nearest thing to an authoritative source is Google's own documentation, which asks you to "Think about the words that a user might search for to find a piece of your content" and warns plainly that there are "no secrets here that'll automatically rank your site first in Google". If you want a five-item version grounded in something measurable, use the decisions in this article: find the query, prove you do not already serve it, confirm the SERP matches your reader, publish one page for one parent topic, and set a review date. Any listicle of five tactics is somebody's opinion, including that one.

What are the 5 pillars of content strategy?

Nobody owns this definition, and the vendors who publish it do not agree with each other. You will see audience, goals, channels, production and measurement in one version, and substance, structure, workflow and governance in another. The disagreement is the useful information: it tells you the number five is a formatting choice rather than a finding. For SEO specifically, the pillars that change outcomes are the four decisions at the top of this article, because each one can be checked against data you already have.

What is the 70 20 10 rule in content marketing?

It is a budget split, usually described as 70% proven content, 20% content that builds on what has worked, 10% experimental. The marketing version is credited in secondary reporting to Jonathan Mildenhall during his time at Coca-Cola, adapted from an older learning and development model associated with Michael Lombardo and Robert Eichinger. We could not find a primary source from Coca-Cola, so treat the attribution as journalism rather than fact. More importantly, it is an advertising budget heuristic. It says nothing about which queries to target, and it predates the search problem it gets quoted at.

What is the 80/20 rule in SEO?

It is the Pareto principle applied informally, and it is not a Google framework. The version worth keeping is the one you can verify in your own data: export your Search Console pages, sort by impressions, and see how much of the total sits in the top few URLs. On most small sites the concentration is extreme. That measurement is genuinely useful, because it tells you to improve those pages rather than add more. The ratio itself is a slogan, and your actual numbers will not be 80 and 20.

Do I need to publish on a schedule for SEO?

No, and Google says so more directly than most SEO advice does. Its content guidance lists "Are you producing lots of content on many different topics in hopes that some of it might perform well in search results?" among the questions that signal search-engine-first content, and answers the related freshness tactic with "(No, it won't)". A cadence is useful for you, as a forcing function for a habit. It is not a ranking input. One page per week that each answer one query you verified beats five that each answer none.

How many pages should a content strategy for SEO start with?

Fewer than you expect, and chosen in the order described above: the query that describes what you sell, the two or three a buyer types just before it, then the informational pages that link to those. For a site with nothing published, that is roughly five to eight pages before the first review date. The number matters far less than whether each page has one parent topic, one job and one date on which you will decide to keep it, rewrite it, or leave it alone.