How an Engagement Works

Every Ivan Sp engagement follows a fixed three stage process of diagnose, strategize, and build. Clients provide Search Console access and commercial priorities to establish the scope, exclusions, and timeline. Each stage uses specific tools to move from data analysis to verified implementation.

Stage One: Diagnose

The site gets audited with Screaming Frog, Google Search Console, Google Analytics 4, Ahrefs, SEMrush, and custom scripts. The purpose is identifying what actually limits rankings, which is usually one of three things: a technical fault, a content gap, or a structural problem.

Diagnosis also grades the source material available. Where a business cannot supply the specifics a page needs, that gap gets raised before writing starts rather than papered over with confident language.

Stage Two: Strategize

Findings become a prioritized action plan, ordered by impact rather than by effort. High-impact fixes come first, content strategy second, and programmatic scaling last.

You receive three things at this stage:

A prioritized roadmap you can act on with an existing team
A quick wins list covering changes that need little time and carry real upside
A six-month plan showing what compounds over a longer horizon

The roadmap is written to be implementable by someone else. A plan only Ivan Sp can execute is a weaker deliverable than one an internal team can run.

Stage Three: Build and Optimize

Where a system is needed, the system gets built. That work covers programmatic content pipelines, AI agents, schema automation, and internal linking tools, implemented in WordPress, through PHP, or inside the client’s own content management system.

Implementation support continues rather than ending at handover. A strategy delivered as a PDF and never applied produces no results, which is a common way search engagements fail.

Change Management on Sites That Already Rank

A site already ranking well for competitive terms gets a more conservative rollout than a site starting from zero. Winning pages carry downside risk from bulk changes, and that risk gets weighed before anything is edited.

Before new content publishes, target terms get checked against keywords the site already earns. Terms in the upper quarter of impressions count as protected territory and never get diluted by a competing new page.

SEO Engagement Scope and Agreement

Three things get defined before work starts: who on the client side approves changes, what falls outside the agreed scope, and what counts as a result. Vague versions of all three are a common reason engagements go wrong, and stating them early costs one conversation.

The Three SEO Engagement Stages in Order

Stage one, diagnose: Establish what is actually wrong, using the site’s own data.
Stage two, strategize: Decide what to build, ordered by commercial exposure.
Stage three, build: Implement, then verify the change took effect rather than that it shipped.
Why the order is fixed: A strategy written before a diagnosis is built on an assumption.
What the client supplies: Read access to Search Console and analytics, plus commercial priorities.
What is agreed up front: Scope, exclusions, timeline, and how the work will be verified.
Change management: Sites already ranking get changes staged rather than deployed at once.
Verification: Confirmation the fix worked, which is a separate step from confirming it deployed.

Ask Your Current Provider Two Questions

Ivan Sp suggests two questions that reveal how an engagement is actually running.

Ask which specific change caused your last ranking improvement.
Ask how that was verified, and what evidence supports the answer.
Ask what is explicitly excluded from your current scope.
Ask what happens next month if nothing moves.

A provider who cannot attribute a result deployed everything at once. A provider with no ready answer on exclusions has not written a scope.

Where the Standard Engagement Advice Breaks Down

Did your last provider’s plan arrive before or after they saw your data?

The usual advice: Come to the first call with a plan, since a provider who knows the answer already is one who knows the industry.
What holds up in practice: A plan written before seeing the data is a template. A provider arriving with recommendations has described what they usually sell rather than what your site needs. The two coincide by accident.

Picture a first call where the provider presents a twelve-month plan built from the industry and the website’s public pages. The plan is detailed and confident.

Nobody has opened Search Console. The site’s largest problem is an index coverage gap that appears in no public view, and no item on the plan addresses it.

I arrive at the first call with questions rather than a plan. The recommendations come after the data, which is the only order that produces a plan about your site.

A plan built before a diagnosis commits the budget to whatever the provider usually sells. Discovering the mismatch takes a quarter.

A twelve-month plan written from your homepage is a template with your logo on it. You will pay for it monthly.

Five Things Commonly Assumed About How SEO Work Runs

Which of these matched your last engagement?

I meet these assumptions on most first calls.

Commonly assumed	What actually holds
A good provider knows the plan already.	A plan written before the data is a template. Recommendations that arrive before a diagnosis describe what the provider usually sells.
Changes should go live as soon as they are ready.	On a site already ranking, staged deployment is what lets you attribute a change. Shipping everything at once removes that.
Verification means confirming the work shipped.	Shipping is not the same as working. Verification means confirming Google’s behaviour actually changed.
The scope can be refined as you go.	An unwritten scope becomes a disagreement. Agreeing exclusions before starting is what prevents one.
Faster is better.	On a ranking site, a fast broad change makes cause and effect unrecoverable. Sequencing is a diagnostic tool.

The order these stages run in decides whether you can attribute anything at the end. Getting it wrong costs you the ability to learn from your own budget.

A Worked Example: Two Deployment Approaches

Consider one set of fixes deployed two ways on a ranking site. The comparison below is illustrative, written to show what sequencing buys. No client engagement is described here.

Property	All at once	Staged
Changes deployed in week one	14	3
Time to full deployment	One week	Six weeks
Rankings moved by week eight	Yes	Yes
Which change caused the movement	Unknown	Attributable to two of fourteen
What could be repeated on the next site	Nothing specific	Two specific changes
What happens if rankings fall	Fourteen candidate causes	Three, then one

I stage on sites that already rank. Both columns reached the same place, and only one produced knowledge.

Deploying fourteen changes at once is adding fourteen ingredients and tasting once.

The final row is the risk case. A fall after a single broad deployment leaves fourteen things to unwind under pressure.

How an Engagement Runs, Stage by Stage

I follow the same sequence on every engagement, whatever was discussed on the first call.

Hold a free thirty-minute call, with no plan presented.
Take read access to Search Console and analytics before forming any opinion.
Run the diagnosis on the site’s own data rather than on its public pages.
Report the findings ordered by commercial exposure rather than by severity label.
State plainly where nothing is wrong, since that is part of a diagnosis.
Agree the scope in writing, including exclusions and how verification will work.
Sequence deployments on any site that already ranks, so changes stay attributable.
Implement, then wait for the recrawl before claiming anything worked.
Verify by checking Google’s behaviour changed, rather than by checking the code shipped.
Report against the agreed measure, including the periods where nothing moved.

How This Engagement Is Scoped

What Gets Agreed Before Anything Starts

Scope: The specific deliverables, named individually rather than as a service category.
Exclusions: An explicit list, so nothing is assumed to be included.
Access: What the client supplies, and when each item is needed.
Sequencing: Whether changes deploy at once or in stages, and why.
Verification: How each change will be confirmed to have worked.
Timeline: Including the waiting periods that recrawls require.
Reporting: The measure, the cadence, and what happens in a month with no movement.
What is never agreed: A ranking, a traffic figure, or any outcome the practice does not control.

Tools and Primary Sources

The references below cover the data and the standards each stage uses. Each opens in a new tab.

Google Search Console supplies the diagnostic data stage one works from.
Google guidance on hiring an SEO sets out what Google recommends agreeing with any provider before starting.
Search Console URL Inspection is how a fix gets verified against Google’s own view rather than against the code.
Google Search Central documentation is the reference every recommendation is written against.

Common Questions About SEO Engagements

How long does a typical engagement take?

Fixed-price packages state their hours and usually complete within two to four weeks. Programmatic builds and agent projects run longer, and the timeline gets set after diagnosis.

What access do you need?

Administrative access to the website, plus Google Search Console and Google Analytics 4. Additional access depends on scope, and everything needed arrives as a checklist at the start.

What if the diagnosis shows the problem is not SEO?

The diagnosis says so plainly. Some traffic problems trace back to conversion, product, or pricing rather than search, and selling search work into those situations helps nobody.

Why does the diagnosis have to come first?

A strategy written before the data is built on an assumption about what is wrong. First diagnoses commonly find a fault that appears in no public view. No plan written from the outside would have addressed it.