B2B SEO strategy: what to build in which order
Most B2B SEO strategies are keyword lists with a publishing schedule attached. A strategy is a sequence of decisions about what gets built first and what deliberately waits.
A strategy is a sequence, not a list
The document you usually receive is a spreadsheet of keywords grouped by theme, a monthly publishing target, and a rankings dashboard. That is a plan for producing content. It is not a strategy, because it contains no decisions about order, and order is where the leverage is.
Two teams with identical keyword lists get very different results depending on what they build first. The one that publishes forty educational posts before establishing site architecture ends up with forty pages competing for overlapping terms and a link graph that reflects publishing order. The one that establishes architecture first has somewhere for each of those forty pages to sit.
Quarter one: architecture and commercial intent
The first quarter has two jobs, and neither of them is volume.
- Decide the content types. A closed set, four to six. Comparison, use case, problem-first guide, tool page, and reference. If a proposed brief does not fit one, that is a signal to interrogate the brief, not to add a seventh type.
- Specify the link rules per type. Where each type links up, across, and down. Written into the brief template so the graph maintains itself rather than depending on whoever writes the piece remembering.
- Publish the commercial-intent pages. Comparison pages and alternatives pages. These rank faster than educational content because the SERPs are less contested by publishers, and they convert immediately because the searcher has already decided to buy.
- Clear blocking technical defects. Only the blocking ones.
By the end of the quarter you should have a small number of pages that convert, and a structure that tells you where everything else goes.
This is the work I do as SEO systems engagements: deciding the content types and link rules before anything gets published against them.
Quarter two: the first cluster, built properly
One pillar page on your core problem space, five to eight supporting pages, linked as a hub. Not three half-built clusters, one complete one.
The reason for completeness over breadth is that a cluster only starts working as a cluster once it is dense enough for the internal links to signal topical coverage. A pillar with two spokes is a page with two links. A pillar with seven spokes, each linking back and laterally, is a structure a crawler can read as depth on a subject.
Choose the problem space by asking which conversation you want to be the default source on. Not which keyword has the most volume. If your product solves onboarding drop-off, own the onboarding drop-off conversation completely before you write anything about adjacent topics.
The rules that make a cluster behave like a cluster rather than a folder are set out in the internal link graph piece.
Quarter three: utility and the second cluster
This is where a free tool earns its place, once you know the channel works. Something that replaces a spreadsheet your buyers currently maintain by hand, runs client-side, and produces output worth keeping.
Build it on its own URL with the method written out below the tool: what it assumes, where the weights came from, what it is not appropriate for. That reasoning is what makes the page linkable and citable rather than a widget with nothing to grab.
Start the second cluster in parallel, chosen for adjacency to the first so the two reinforce rather than compete.
Quarter four: depth over breadth
The instinct at this point is to add more clusters. Resist it for one more quarter and go back over what exists.
Pages published in quarter two were written before you had rankings data on them. You now know which ones nearly rank, which queries they picked up that you did not target, and where the SERP has shifted. Updating a page that sits at position eleven is materially cheaper than writing a new page that starts at nothing, and the return is faster.
Run a consolidation pass at the same time. Pages that overlap get merged, the weaker URL redirecting into the stronger. Two pages at position fifteen usually make one page at position six.
Note what this is not: another crawl of the whole site. The case against repeat audits explains why the second audit rarely earns its cost.
What to do when the plan meets reality
A four-quarter sequence assumes four quarters of stable conditions, which is not a thing that happens. Three disruptions are common enough to plan for.
The company pivots in quarter two
A competitor launches a large content programme
Budget is cut in half at quarter three
The principle underneath all three: the sequence exists so that each quarter builds on the last. Anything that breaks the dependency is worth stopping for. Anything that merely slows it down is not.
The one disruption that should end a programme rather than pause it: losing the internal owner with no replacement. Content produced without someone who can answer a product question is generic by necessity, and generic content in B2B does not convert regardless of how well it ranks.
The dependencies that decide the order
The reason a strategy is a sequence rather than a list is that most of the work depends on other work. Publishing in the wrong order does not slow you down, it wastes the effort entirely.
| This work | Depends on | If done early |
|---|---|---|
| Cluster content | Architecture and the keyword map | Pages compete with each other and cannibalise |
| Free tools | Enough domain strength to rank at all | A good tool nobody finds, and no links earned |
| Thought leadership | Commercial pages that convert the traffic | Traffic arrives with nowhere useful to go |
| Consolidation | Twelve months of data on what earns | You merge pages before knowing which was working |
The free tools row deserves emphasis, because building one is the most enjoyable work in the plan and the easiest to pull forward. A tool published in quarter one on a domain with no authority sits unfound. The same tool in quarter three, on a site with a cluster behind it, earns links from the first month.
The general rule: anything that depends on authority goes late, anything that builds authority goes early, and anything commercially urgent goes first regardless, because a programme that produces nothing usable for three quarters does not get a fourth.
Writing the plan down so it survives contact
A sequence that lives in one person’s head is not a strategy, it is an intention. The version that survives staff changes, budget arguments and a quarter of bad numbers is written down, and it is shorter than people expect.
Four things, on one page.
- The order, and what each stage depends on. Not the tactics, the dependencies. Anyone should be able to see why quarter three cannot happen before quarter two.
- What each quarter should produce, agreed in advance. Impressions at four months, clicks at eight, pipeline at twelve. Written before anything is published, because after the fact every number can be explained.
- What would make us stop. The condition, stated in advance, that means this is not working. A programme with no stopping condition never gets assessed, only defended.
- Who owns it. One named person who can answer a product question. Not a team, a person.
The last one determines the other three in practice. Content produced without someone internal who understands the product is generic by necessity, and generic content in B2B does not convert however well it ranks. That constraint, and what to do when it binds, is the subject of the content marketing piece.
The page is worth revisiting each quarter and rewriting each year. A plan nobody has looked at in twelve months is describing a company that no longer exists.
Where a year of this actually gets you
Worth being concrete about the destination, because a four-quarter plan with a vague endpoint tends to lose funding in quarter three.
A year of this, executed properly on a mid-market B2B SaaS site starting from close to nothing, produces a specific shape rather than a specific number.
What it does not produce, and should not be sold as producing: a large traffic number. A complete cluster on a mid-market category might be a few thousand sessions a month, and if those sessions are the right people that is a good year.
The thing worth measuring at twelve months is not the total. It is whether the commercial pages rank for the terms your sales team would choose, and whether anyone internal can now brief a new page without asking you. The first is the outcome. The second is whether it will keep happening.
The parts you can start before the plan is finished
Sequencing is the argument of this piece, and it has an obvious objection: waiting for a full plan before doing anything is its own failure mode. Three things are safe to start immediately, because nothing else depends on them and they are useful under any plan.
- Fix what is technically broken. Pages that do not render, redirect chains, a sitemap listing URLs that return errors. This is not strategy, it is maintenance, and it does not need a plan to justify it.
- Write down the two competitors that come up most in sales calls. Ten minutes with a salesperson. It is the input for the first commercial pages under any sequence you eventually choose.
- Claim your brand terms. If a competitor outranks you for your own product name, that is revenue leaving with no strategy required to stop it.
What is not safe to start early is anything that depends on a positioning decision or a keyword map. Publishing content before those exist is the single most common way a year of budget produces pages that have to be rewritten.
The distinction worth holding: maintenance can start today, architecture needs the map, and content needs both. Teams that conflate the three either wait too long to do the first or rush into the third.
What to report at each stage, and what not to
Reporting the wrong metric at the wrong stage is how good programmes get cancelled.
- Quarter one. Indexation, technical defect closure, commercial page rankings. Do not report traffic. There is not any yet, and promising some invites the wrong question.
- Quarter two. Impressions and average position on cluster terms. Impressions move well before clicks and prove the direction is right.
- Quarter three. Clicks, tool completion rate, assisted conversions. Now traffic is a fair measure.
- Quarter four. Pipeline influence, self-reported attribution, branded search volume. The lagging measures that actually justify the budget.
| Quarter | Report this | Do not report this yet |
|---|---|---|
| Q1 | Indexation, defects closed, commercial rankings | Traffic |
| Q2 | Impressions, average position on cluster terms | Pipeline |
| Q3 | Clicks, tool completion rate, assisted conversions | Revenue attribution |
| Q4 | Pipeline influence, self-reported source, branded search | – |
Setting this schedule in advance protects the programme from being judged on quarter four metrics during quarter one, which is the most common way B2B SEO dies.
Architecture and commercial-intent pages first, one complete cluster second, utility third, depth fourth. Agree in advance which metric gets reported in which quarter, so the programme is not judged on pipeline before it has had time to produce any.
The reporting schedule is the part of this I would fight hardest to keep. Agreeing in advance which metric gets reported in which quarter is what stops a programme being judged on pipeline during the quarter where there is not any yet.
I have never regretted spending quarter four going back over existing pages instead of adding new clusters. Moving a page from position eleven to position six is cheaper and faster than starting a new page at nothing, and almost nobody chooses it because it does not feel like progress.
Frequently asked
What should a B2B SEO strategy document contain?
How many blog posts per month should we publish?
When should we build the second topic cluster?
What SEO metrics should we report to leadership?
I take on a small number of engagements at a time. If you are past product-market fit and want to talk through what this would look like for your situation, the calendar is open.
Book a call