How to do SEO: the order that actually works

Why the order matters more than the checklist

There are perhaps forty things you can do to a website to make it rank better, and you'll find all forty on any SEO blog. The list isn't scarce. Sequence is.

Two mistakes account for most wasted SEO budgets in India, and both are ordering mistakes rather than competence mistakes.

Content before indexability. You commission thirty articles and publish them onto a site where the template renders client-side, or canonicals point at a paginated parent, or 4,000 filter URLs are eating the crawl budget. Publishing doesn't fail loudly — pages go live, the CMS says success, the invoice gets paid. You find out in month five that eleven of the thirty were ever indexed.

Links before content. Three months earning links to a homepage and a blog that answers no commercial question. The authority arrives and has nowhere useful to flow. Links amplify pages; they don't originate demand.

The order below exists because each stage produces what the next stage needs. Indexation data tells you which existing pages are viable. Intent work tells you which gaps are real rather than theoretical. Gap work tells you what's finally worth pointing a link at.

Step one: make your pages fetchable and indexable (weeks 1–3)

Nothing else on this page matters until this is true. A page that Google can't fetch, render or index cannot rank, regardless of how good it is. This stage is unglamorous, takes two to three weeks, and is where an audit earns its fee.

1. Confirm Googlebot sees your content. Search Console → URL Inspection → Test live URL → View crawled page. Read the rendered HTML. If your body copy isn't there, you have a JavaScript rendering problem and it's now the only item on your list.
2. Count indexed against published. If you have 400 URLs and 120 indexed, find out why before you publish page 401. Blocked by robots.txt and Excluded by noindex are configuration mistakes; Crawled – currently not indexed is Google saying the page isn't worth its space.
3. Check robots.txt and your sitemap. The sitemap should contain only canonical, indexable, 200-status URLs. Most are a full URL dump including redirects and noindexed pages, which teaches Google to trust it less.
4. Clean the redirects. Chains, loops, and 301s that land on 404s.
5. Fix the templates, not the pages. Duplicate titles, missing H1s, canonical logic and schema all live in the template. Fixing one page teaches you what's wrong; fixing the template fixes a thousand.
6. Measure server response time. An average above 800ms in crawl stats is a hosting problem capping everything downstream.

Step two: fix intent on the pages you already have (weeks 2–6)

The step almost everyone skips, and the one that produces the first quarter's results. You already have pages Google considers relevant. It just doesn't consider them best.

Open Search Console, take the last three months, sort queries by impressions. Look for pages sitting between positions 8 and 25 with real impressions and poor click-through. Google is telling you in writing which pages it's prepared to rank higher.

For each one, do the thing most people don't: open the results page and look at what's winning. If the top ten are comparison tables and yours is a 1,200-word essay, you don't have a word-count problem, you have a format mismatch, and adding keywords won't fix it.

• Answer the query in the first 100 words. Then expand. Helps readers, featured snippets and AI answers at once.
• Rewrite the title in the language of the query, not the language of your product team.
• Merge cannibalising pages. Two mediocre pages on one query beat each other up. One good page beats them both.
• Add the People Also Ask sub-questions — a free map of what your answer is missing.
• Link to it internally from your strongest pages, with descriptive anchor text.

Step three: fill the gaps that obviously exist (weeks 4–12)

Only now does new content make sense, because now you know what you cover and what converts. Build in this priority order — and notice the blog is third.

Commercial pages you simply don't have. Pricing. One page per service, not one page listing eight. Comparison pages. Location pages if you have locations. Boring to write, and they carry most of the revenue. Most Indian sites have twelve blog posts and no pricing page, which is precisely backwards.

The questions your sales team answers every week. Sit on a call, listen for twenty minutes, write down every objection. Each is a page with proven commercial search intent behind it, and nobody has to guess whether it matters.

Competitor gaps you could plausibly win. Filter hard. A query where three national brands with 5,000 referring domains hold the top three is not a gap, it's a wall.

Then build in clusters, not scatter. One pillar and four to six supporting pages on a single topic, internally linked, beats twenty unrelated posts every time. Four solid pieces a month for a year beats forty in one quarter and silence after.

Step four: authority and links (month 3 onward)

Links come last for one reason: they amplify what exists. Point authority at a site with nothing worth reading and you've paid for reach into an empty room.

When you start, work outward from what you already own — it's free, fast, and it's what agencies skip because it doesn't look like work on a report.

1. Reclaim. Unlinked brand mentions, links that used to point at you and now hit a 404, links lost in a past migration. Pure recovery, no pitching.
2. Relationships you already have. Customers, suppliers, distributors, associations, event pages, partners. A supplier listing you as an authorised dealer beats most guest posts and takes one email.
3. Earned. Original data from your own operations, commentary journalists can quote, a genuinely useful free tool. Slow, expensive, and the only kind that compounds.
4. Internal links, continuously. Often the fastest ranking gain available on a site over about 100 pages, and the one nobody bills for. Internal linking beats link building more often than the industry admits.

Nine things to deliberately skip in the first 90 days

Each of these sits near the top of some SEO checklist. Each is a distraction in your first quarter, and the discipline to leave them alone is worth more than the discipline to do them.

• A disavow file. Not unless you have an actual manual action. It's a loaded gun aimed at your own backlink profile.
• Chasing 100/100 on PageSpeed. Get inside the "good" Core Web Vitals thresholds and stop. The last twenty points cost more than the first eighty.
• Programmatic page generation. Thousands of templated pages before one format has proved it converts is how a site earns a scaled-content problem.
• Rebuilding the website. Unless the rendering test failed outright, a replatform in month one delays everything by a quarter and adds risk you don't need yet.
• Hreflang and multi-language. Only once one language works. Two broken languages isn't progress.
• Guest posting at volume. Costs money, produces links Google discounts, and steals attention from step two where the returns are.
• Head terms you cannot win. "SEO services" is a vanity row on a report. Track the queries in your plan.
• Full schema coverage. Organisation, LocalBusiness and Product are worth doing. Everything else can wait.
• Changing your domain. Whatever the reason, do it after you understand your baseline, never before.

The 90-day sequence on one page

Take the shape rather than the tactics. Tactics change every year. The dependency order hasn't changed in a decade.