SEO Strategy Guide

Most search plans fail before anyone writes a word of content, because the plan is a list rather than a sequence. Fix speed, publish blog posts, build links, add schema, refresh old pages. Every item on that list is defensible on its own. Run them in the wrong order and you’ll spend six months paying for work that can’t compound.

This page is the hub for how we build an SEO strategy at SerpInsight. The order, the audit that comes first, the research that decides what you chase, the technical floor you have to clear, and the measurement that tells you whether any of it is working. It’s long on purpose. Read it top to bottom once, then use the contents below as a reference when you need to settle a specific argument with yourself or with a client.

How to build an SEO strategy that survives contact with reality

An SEO strategy survives when it’s built around constraints you actually have: a real budget, a real writing capacity, a real site carrying real technical debt. Plans die when they quietly assume unlimited content production, instant stakeholder approval, and a clean crawl. The version that works is usually smaller than the version that gets presented in the kickoff deck, and it’s almost always more sequenced.

There’s one more thing that separates a plan that holds from a plan that gets abandoned in month three: it states what you’re not doing. If your document doesn’t name the tactics you’re deliberately skipping this quarter, it’s a wish list. Saying “no new content until the indexing problem is fixed” is a strategic decision. Saying “we’ll do content and technical and links” isn’t.

Here’s what this SEO strategy guide walks through, in the order we’d actually run it.

• The order the work should happen in
• Start with what you already have
• How do you plan SEO keyword research around intent rather than volume?
• Technical foundations that actually matter
• How to decide between content and links
• What should an SEO roadmap show at 30, 90 and 180 days?
• Budget allocation and SEO priorities
• What to do when rankings drop
• Frequently asked questions

The order the work should happen in

Technical foundation, then intent-matched content, then internal linking, then external links, then iterate. That’s the sequence. It isn’t the only defensible order, but it’s the one that wastes the least money, and the reason is simple: each stage raises the return on the stage after it.

Think about what a link actually does. It passes authority to a URL. If that URL is blocked from indexing, the authority goes nowhere. If it’s indexed but the page answers a different question than the searcher asked, the authority props up a page that won’t convert the click into a stay. If the page is fine but nothing on your own site links to it, the authority arrives and stops, instead of spreading to the twelve related pages that could also use it. Every one of those failures is upstream of the link, and none of them get fixed by buying more links.

So we’ll be direct about it: buying links for a broken site wastes money. Not “delivers less than it could” — wastes. If a third of your pages are excluded from the index, if your product templates carry two sentences of copy, if your category pages compete with each other for the same term, then external authority is water poured into a bucket with the bottom cut out. Fix the bucket. The links will still be available next quarter, and they will cost the same.

1. Technical foundation

   Make sure Google can crawl, render and index the pages you care about, and that each page has exactly one canonical address. This stage is mostly diagnosis and small fixes, and on many sites it’s the cheapest stage in cash terms because the work is a developer afternoon rather than a retainer. If you’d rather not do the diagnosis yourself, this is what a structured technical and content audit is for.

2. Intent-matched content

   For every target query, the page has to be the same kind of thing the searcher wanted. A comparison query needs a comparison page, not a product page with a comparison paragraph bolted on. This is where most sites have the largest gap, and where on-page optimization work tends to move the needle furthest per hour spent.

3. Internal linking

   Once the pages exist and are worth reading, connect them. Contextual links from related articles into your money pages do two things: they tell Google what the destination is about, and they move whatever authority the site already has toward the pages that need it. There’s a whole method to this, and we’ve written up how to build an internal linking structure that actually moves authority rather than just adding navigation.

4. External links

   Now add authority from outside. At this point a placement lands on a page that’s indexed, matched to intent, and connected to the rest of your site, so the effect is visible instead of theoretical. A link insertion (also called a niche edit) is our preferred format here because the host article is already indexed and already gets readers, so there’s no waiting for a brand-new post to earn traffic first.

5. Iterate on evidence

   Six to twelve weeks after each round, look at what moved and what didn’t, and let that decide the next round. The pages that jumped from position 24 to position 11 tell you something the pages that didn’t move also tell you. Most of the compounding in search comes from this stage, and most plans never reach it because the first four stages were run simultaneously and nobody can attribute anything.

One honest caveat. Sequence doesn’t mean the earlier stages ever finish. Technical debt accumulates as soon as someone launches a new template, and content ages. What the sequence means is where the marginal hour or dollar should go when you have to choose, and you always have to choose.

Start with what you already have

Before you plan a single new page, inventory the ones you have. On most sites older than a couple of years, there’s more upside in the existing library than in anything new you could publish this quarter, and it’s cheaper to reach.

Pull every query where you rank in positions 6 to 30 from Search Console, over the last three months, sorted by impressions. This is the most valuable list on your site. These are queries where Google has already decided you’re relevant enough to show, and where a modest improvement moves you into territory that gets clicked. Position 24 to position 8 is a much shorter journey than nothing to position 8, and the work is usually specific: the page half-answers the query, or it’s missing a section the top results all have, or it has three internal links pointing at it when it should have fifteen.

Next, look for cannibalization. Sort the same export by query and see whether the ranking URL flips between two of your pages across weeks. When it does, Google is telling you it can’t decide which of your pages is the answer, and the usual outcome is that neither ranks as well as one strong page would. The fix is rarely a redirect on day one. Read both pages, decide which one deserves to be the answer, fold the useful parts of the weaker one into it, and then redirect. Keep the decision written down somewhere, because in eight months somebody will want to know why that URL disappeared.

Then prune. Pages that get no impressions, no links, and no conversions aren’t neutral. They dilute the topical signal of the sections they sit in, they consume crawl budget on large sites, and they make your own internal linking harder to reason about. Deletion isn’t the only option — consolidation, noindexing and rewriting are all valid — but the decision has to be made deliberately. We’ve set out the full decision tree for pruning content that no longer earns its place, including when pruning is the wrong call.

The output of this stage is a short ranked list: pages to improve, pages to merge, pages to remove, and gaps that genuinely need something new. In our experience that list is roughly three quarters existing pages and one quarter new ones, which is close to the opposite of how most content calendars get built.

How do you plan SEO keyword research around intent rather than volume?

Start from what the searcher wants, not from the search volume column. For every candidate query, open the results page and look at what Google is already rewarding: the format, the depth, the type of site. If the top ten are all free tools and you were planning a 2,000-word article, the volume number is irrelevant because you were never going to rank with that page.

That check takes about ninety seconds per query and it saves entire content sprints. A query like “invoice template” is served by downloads. “Best invoicing software for freelancers” is served by comparison articles, usually from publishers rather than vendors. “Xero pricing” is served by the vendor. Same broad topic, three completely different pages required, and only one of them is worth your time if you sell invoicing software. The full method for reading a results page and classifying what it wants is in our write-up on matching pages to search intent before you write them.

Once you have candidates, cluster them. Queries that return substantially the same results belong on one page, not five. A practical rule: if the top five URLs overlap by three or more between two queries, one page can serve both. Clustering this way is how you end up with twenty pages that each own a group of related queries instead of eighty thin pages each chasing one, and it’s the mechanism behind topical authority. Google isn’t counting your pages. It’s deciding whether your site is a credible source on a subject, and coverage with depth reads as credible where coverage without depth doesn’t.

Then be realistic about what you can win. A new site chasing “project management software” isn’t going to place, no matter how good the page is, because the results are held by domains with fifteen years of accumulated authority and thousands of referring domains. Head terms fail early for a structural reason: they’re the queries with the most commercial value, so they attract the most investment, so the authority bar is highest exactly where a young site has the least. Start three or four modifiers down. “Project management software for architecture firms” has a fraction of the volume and a fraction of the competition, and the traffic converts better because the searcher told you who they’re. Win those, accumulate authority and internal links, then move up.

Technical foundations that actually matter

Technical work splits into two buckets: things that determine whether a page can rank at all, and things that adjust how well it ranks. The first bucket isn’t optional and the second one is frequently oversold.

Crawlability and indexation come first. Can Googlebot reach the page without a login, a broken redirect chain, or a disallow rule? Does the page return a 200? Is it self-canonical, or is it pointing at something else by accident? Is it in the sitemap and linked from at least one other page on the site? Orphan pages are more common than people expect, especially on sites with faceted navigation or a headless front end where links are rendered client-side. Check the Pages report in Search Console and read the exclusion reasons properly — “Crawled, currently not indexed” and “Discovered, currently not indexed” mean different things and call for different fixes.

Site structure comes second. Every important page should be reachable within three clicks of the homepage, and the path should be meaningful rather than incidental. Flat isn’t automatically better; a clear hierarchy where category pages link down to their children and children link back up is easier for both crawlers and humans than a homepage with two hundred links on it. If you’re running a large catalogue, crawl budget becomes real, and the lever is usually reducing the number of low-value URLs rather than asking Google to crawl more.

Core Web Vitals come third, and honestly. They are a ranking signal, they’re a small one, and they mostly act as a tiebreaker between results of similar relevance. A slow page that’s the best answer will usually outrank a fast page that isn’t. That said, poor Largest Contentful Paint and heavy layout shift cost you conversions and returning visitors regardless of what Google does with the numbers, so the work is worth doing on its own commercial merits. Fix the template, not the individual page — one template fix improves ten thousand URLs.

What we’d deprioritize: schema markup beyond the types that actually generate rich results in your vertical, hreflang on single-market sites, and the long tail of audit-tool warnings that no human would ever notice. A crawl report with 400 issues isn’t a to-do list. Perhaps a dozen of those issues affect indexation or rendering, and the rest are noise dressed up as urgency.

How to decide between content and links

The question comes up every month in every SEO budget conversation, and the answer depends entirely on the symptom. Diagnose first. The table below maps what you’re seeing to the most likely cause and to the lever that actually addresses it.

Two rules of thumb fall out of that table. If a page is close — visible, relevant, sitting outside the top ten — authority is usually the binding constraint. If a page is far away, the page itself is the problem, and no amount of external authority will fix a mismatch between what you published and what the searcher wanted.

There’s also a threshold below which link building is the wrong purchase entirely. Sites with very little existing authority and very little content don’t have anything for links to amplify, and the money is better spent on the first three stages. We go into the reasoning, including the cases where the answer is genuinely no, in our take on whether your site needs backlinks at all. And if you do decide links are the constraint, understand the trade-off you’re accepting: paid placements run against Google’s guidelines, and nobody can eliminate that risk. We set out how we think about it in the honest risk profile of paid links, and the wider method sits in our complete guide to building links that hold.

What should an SEO roadmap show at 30, 90 and 180 days?

At 30 days, leading indicators only. At 90 days, early movement in positions and impressions. At 180 days, traffic and revenue you’d be willing to defend in a board meeting. Anyone promising ranking changes at 30 days is either working on a very strong domain or telling you what you want to hear.

The distinction that saves the most arguments is leading versus lagging. Leading indicators are things you control and can observe almost immediately: pages fixed, pages indexed, internal links added, placements live, average position on the specific queries you targeted. Lagging indicators are the ones the business cares about: sessions, leads, revenue. Leading indicators move first and predict the lagging ones. If you only report the lagging set, the first three months look like nothing is happening, and that’s when programmes get cancelled one month before they would have worked.

Here’s what we’d check at each checkpoint. At 30 days: indexation coverage, crawl errors resolved, whether the pages you shipped are in the index, and whether any targeted query has entered the top 100. At 90 days: average position on the target cluster, impressions on those queries, click-through rate on pages where you rewrote titles, and whether new placements are live and passing links. At 180 days: organic sessions to the target pages, assisted and direct conversions, and the ratio of queries in the top ten versus the top thirty compared with your starting position.

Just as important is knowing what is noise. Daily rank fluctuation of two or three positions is noise. A single week of traffic decline is noise, particularly if it lines up with a holiday or a seasonal pattern. Third-party rank trackers disagreeing with Search Console is noise, because they sample different locations and devices. What isn’t noise: a sustained multi-week directional change, a step change dated to a known update, or indexation dropping. Set the review cadence at monthly and resist looking daily, because daily looking produces daily reacting and there’s nothing more destructive to a search programme than a strategy that changes every fortnight. If you want the fuller version of the timing argument, we’ve written about how long SEO usually takes to show real movement and why the honest answer is a range rather than a date.

Budget allocation and SEO priorities

We’ll use relative splits rather than invented figures, because the right absolute number depends on your margins and your competition, and anyone who quotes you a universal monthly spend is guessing. The only hard numbers here are our own published prices: $87 per link at Starter, $147 at Growth, $247 at Authority, with 10% off at ten or more links and 15% at twenty-five or more.

Notice that the link share in an SEO strategy only becomes meaningful at the third tier. That’s deliberate. On a small budget the first three stages are mostly your own time rather than cash, which is exactly why they get skipped — they don’t feel like spending, so they don’t feel like progress. They are the progress. Our published per-link pricing is there so you can do this arithmetic before you talk to anyone, and the longer argument about sequencing spend sits in where a limited SEO budget should go first.

On pace: consistency beats bursts. A steady handful of placements each month reads more naturally than forty links in one week followed by silence, and it also gives you a chance to see what worked before committing the next round. We’ve set out what a reasonable cadence looks like at different site sizes in our guidance on monthly link pace.

What to do when rankings drop

First, establish whether it’s you or everyone. Check whether the drop is site-wide or page-specific, whether it’s dated to a confirmed update, and whether your competitors moved too. A site-wide slide on the same day as a core update is a different problem from one page falling out of the top ten in an otherwise stable week.

Then resist the urge to change everything. The most common recovery mistake is a panicked rewrite of forty pages in a fortnight, which destroys your ability to attribute anything and often makes the situation worse. Change one variable, wait, measure. If the drop is dated to a core update, the recovery path is a quality and intent problem rather than a technical one, and it’s measured in months rather than weeks. We’ve written up the full diagnostic and recovery sequence in what to do after a Google core update knocks your rankings down.

Frequently asked questions

If you have worked through the sequence and concluded that external authority is now the constraint rather than an excuse to skip the earlier stages, that’s the point where a conversation makes sense. Send us the pages you want to move and the queries you want them to move for, and we’ll tell you honestly whether placements are the right next spend or whether something upstream needs fixing first — including when the answer is that you should wait. Get in touch and we reply within 24 hours.