SEO Playbook · Results

SEO results from the playbook

Results are credible only when the baseline, timeframe, intervention, evidence, and limits travel with the number. Explore published outcomes, how we measure them, and the use cases they can—and cannot—support.

Start free Pricing

amicited.com/seo-playbook/results

HZ-Containers · published outcome

Organic traffic 177× growth

Search impressions 19.3M

Content operation 1 person

Observation period 16 months

The case study retains Search Console and Ahrefs evidence. The traffic comparison uses the pre-FlowHunt period as its baseline.

Evidence contract

Compared with Named baseline

Observed for Fixed window

Changed during Recorded work

Supported by Retained source

A large number without these four fields is a claim, not a result we publish.

How we measure

How we measure these results

A result is a defined change from a recorded baseline during a stated observation window, supported by evidence we can reproduce. We separate visibility signals from clicks, qualified actions, and commercial outcomes.

✓Baseline — use the period before the intervention, a matched group, or another named comparator; never an unlabeled high-water mark.
✓Attribution window — state when work began, when measurement ended, and the lag expected before crawling, ranking, or citation systems respond.
✓Exclusions — remove or disclose paid traffic, brand campaigns, migrations, seasonality, tracking changes, and other events that can distort the comparison.
✓Evidence — retain exports or screenshots, query and page scopes, filters, dates, and calculation steps. Read how we measure results before interpreting any card below.

Headline evidence

How we measure results

These figures come from the two published case studies currently available. Different units describe different parts of the search journey, so they belong in separate tiles rather than a single blended score.

177×

HZ-Containers organic traffic growth versus its pre-FlowHunt baseline

19.3M

HZ-Containers search impressions over 16 months

7.37M

TarmacView search impressions in about eight months

40.6K

TarmacView organic clicks in the same evidence window

6,000+

TarmacView keywords ranking in Google's top 10

Person managing the HZ-Containers content operation

What these numbers do and do not prove

The HZ-Containers evidence includes a 16-month Google Search Console view, an Ahrefs trend, the traffic-growth baseline, and the documented one-person operating model. The TarmacView evidence includes its Search Console and Ahrefs views for a new domain over roughly eight months.

Supported interpretation

A systematic publishing program coincided with substantial organic discovery growth in both cases. The retained search-platform evidence supports the reported impressions, clicks, keyword footprint, and observation periods.

Unsupported interpretation

The figures do not establish that one template, element, prompt, or platform caused every gain. They are not forecasts for another domain, and search impressions are not interchangeable with leads or revenue.

From claim to result

What every result includes

A reader should be able to understand the comparison without asking what the chart leaves out. That is why every publishable result carries its scope, baseline, intervention, window, and evidence as part of the claim itself.

1. Scope defines what was counted

A domain total can hide that only one country, directory, device class, or query set changed. Name the URLs, markets, languages, query classes, engines, and conversion events included. Preserve the filter configuration with the source export. If the scope changes during the window, report both scopes or restart the comparison; silently widening it can manufacture growth.

2. Baseline makes “growth” meaningful

A result exists only in relation to something. Use a fixed pre-change period when the market is stable, a year-over-year period when seasonality is material, or a matched untreated cohort when pages can be compared fairly. Explain why that comparator was chosen. Do not select the weakest week after seeing the outcome, and do not compare a full month with a partial one.

3. Intervention names the work

List the pages launched, elements retrofitted, technical defects fixed, internal links changed, and publication dates. “Content optimization” is too broad to audit. When several changes overlap, publish them together and limit the claim to contribution: the program preceded the movement, but the evidence does not isolate one component as the sole cause.

4. Window respects system lag

Search crawlers need time to revisit pages, ranking systems need observations, AI answer engines can refresh sources on different schedules, and B2B revenue can trail discovery by months. State the start and end dates, but also explain the lag assumed. An early check is useful for diagnosing indexation; it is rarely enough to judge a long-cycle commercial outcome.

5. Evidence survives the dashboard

Dashboards change as data rolls forward, filters are edited, and vendors revise estimates. Retain first-party exports where available, screenshots that show dates and scopes, the calculation used for ratios, and any third-party corroboration. Record known discontinuities such as analytics migrations or consent changes. The source package should let a reviewer reproduce the displayed number later.

Exclusions prevent false precision

Paid campaigns, public relations events, brand launches, acquisitions, domain migrations, stock changes, outages, tracking repairs, and unusual seasonality can all move the same metrics as SEO work. Excluding them is appropriate only when the rule is declared and applied consistently. Otherwise disclose the event beside the result and narrow the conclusion rather than pretending the confounder was absent.

Results by business type

Business type changes the unit of analysis. A catalog needs page-cohort and product-discovery evidence; a consultancy needs qualified-demand and buying-committee evidence. Use these facets to choose the right measurement design, not to borrow another company's target.

E-commerce →

Measure product and category cohorts: indexed coverage, non-brand discovery, AI product citations, product-detail engagement, add-to-cart behavior, and contribution after stock and price changes are disclosed.

SaaS →

Follow category and competitor-query visibility through citations, qualified visits, trial starts, demos, assisted pipeline, and retention signals. Separate branded demand from new category discovery.

Local service →

Segment by location and service area. Track local discovery, map and organic actions, calls, booking starts, completed appointments, and lead quality without combining markets that have different demand.

Marketplace →

Measure both sides of the market: indexable inventory, long-tail coverage, listing freshness, buyer actions, seller acquisition, and completed matches. Expired inventory and empty result pages require their own exclusions.

Media, publisher, and affiliate →

Separate reach from value: query and AI citation coverage, engaged reading, return visits, newsletter actions, outbound merchant clicks, and revenue. Disclose commercial relationships and major algorithm-driven volatility.

B2B services →

Use account and opportunity evidence where volume is small: expert-topic discovery, service-page engagement, proof consumption, qualified inquiries, sales acceptance, and influenced pipeline across long buying cycles.

Results by playbook component

“We did SEO” is not an intervention. Record whether the team launched a page cohort, retrofitted an element, or completed a process phase, then choose a comparison that matches that unit of change.

Component changed	Why it can matter	Measurement design	Decision the result supports
Post-type rollout	A repeated document shape lets a team satisfy the same class of intent consistently across a page cohort.	Freeze the launch list; compare cohort-level indexation, query coverage, citations, clicks, engagement, and qualified actions with the baseline and an unchanged group where possible.	Continue the rollout, revise the specification, narrow the eligible cohort, or stop.
Element retrofit	An element can remove a specific comprehension or extraction problem without requiring a full rewrite.	Record changed URLs and dates. Compare the same pages before and after adding features such as comparison tables or direct answer blocks, while noting other edits.	Standardize the element, change its acceptance rules, or reserve it for contexts where it resolves a real reader need.
Process phase	Publishing more cannot compensate for pages that crawlers cannot access, duplicate architecture, or work aimed at the wrong demand.	Capture defect counts and affected URLs before and after the phase, then observe downstream indexation and visibility only after the expected recrawl lag.	Clear the production gate, continue remediation, or investigate a different constraint before creating more pages.

For process work, start with the technical baseline audit and AI accessibility audit. For production, use keyword and prompt research to define demand and a topical map to prevent overlap before pages are commissioned.

Published case studies

Case studies

Each card names the business model, starting point, work applied, timeframe, and measured outcome. The component links describe the closest playbook specifications; they do not assign isolated causal credit.

HZ-Containers

Business type: e-commerce shipping-container sales and rentals across several European markets.

Starting point: hundreds of products, several languages, a small team, and insufficient capacity to write unique catalog and educational content manually.

Applied: systematic product descriptions, glossary and educational articles, category and market landing pages, multilingual production, and prompt-led targeting. The nearest specifications are the product page and category page.

Timeframe: 16 months in the published Search Console evidence window.

Outcome: 177× organic traffic growth versus the pre-FlowHunt baseline, 19.3M impressions, more than 1,000 traffic-driving queries, and a one-person content operation.

HZ-Containers evidence

TarmacView

Business type: specialist B2B aviation services for drone-based airport-lighting inspection.

Starting point: a brand-new domain with no established authority, a narrow expert audience, dense regulatory terminology, and eight language markets.

Applied: technically precise topic coverage, AI-first structure, demand-led targeting, and multilingual publishing. The nearest specifications are the ultimate guide and glossary term.

Timeframe: roughly eight months from the new-domain starting point.

Outcome: 7.37M search impressions, 40.6K organic clicks, more than 6,000 keywords in the top 10, and a 9.5 average Google position across the reported footprint.

TarmacView evidence

When publishing the next result, use the case study specification. It requires a sourced baseline, timed intervention, evidence ledger, attribution limits, client approval, and a next step that fits the reader's decision stage.

Use cases without named clients

Anonymized use cases

The scenarios below describe implementation and measurement designs for situations where a client name or outcome cannot be published. They are not case studies, benchmarks, or performance claims; no numerical gain should be inferred from them.

B2B manufacturer · 400-SKU catalog

Starting pattern: duplicated supplier copy, thin categories, inconsistent specifications, and no defensible way to compare models.

Playbook application: rewrite priority SKU cohorts, add normalized specification blocks and decision-grade comparisons, strengthen category copy, and control indexation before expanding.

Measurement: freeze treated and untreated URL cohorts; track valid indexation, non-brand query coverage, cited pages, qualified product views, and quote requests. Disclose inventory and price changes.

Vertical SaaS · crowded comparison demand

Starting pattern: strong branded traffic but limited discovery for category, alternative, integration, and workflow questions.

Playbook application: map evaluation prompts, publish fair comparison and use-case pages, add verifiable tables, and connect proof and product documentation.

Measurement: separate brand and non-brand cohorts; track prompt citations, comparison-query visibility, trial and demo assists, accepted opportunities, and the lag between first discovery and pipeline creation.

Multi-location service · inconsistent local pages

Starting pattern: near-duplicate location pages, unclear service areas, mixed contact data, and no page-level ownership for updates.

Playbook application: repair the location architecture, define unique local evidence, standardize service and booking information, and introduce a freshness review.

Measurement: segment every metric by location; compare local impressions, organic actions, calls, booking completions, and qualified jobs while accounting for seasonality and territory changes.

Specialist publisher · aging archive

Starting pattern: a large archive with declining freshness, overlapping articles, inconsistent sourcing, and affiliate pages whose commercial relationships are unclear.

Playbook application: consolidate cannibalizing URLs, refresh evidence, add explicit comparison criteria, strengthen authorship, and retire pages that no longer serve a distinct intent.

Measurement: label refresh, merge, and retirement cohorts; track index coverage, query and citation recovery, engaged reading, return visits, and outbound actions without hiding traffic lost from deliberate removals.

Interpretation rules

Turn a result into your next decision

The purpose of measurement is not to decorate a report. Before work begins, define what each possible pattern will trigger: scale, revise, wait, investigate, or stop.

Visibility rises, actions do not

Check intent fit, snippet or citation framing, page experience, offer relevance, and action tracking. More impressions may mean the page is being shown for broader but less useful demand.

Actions rise, visibility is flat

Inspect conversion-path changes, branded demand, returning users, and attribution rules before crediting the content program. The business improvement may be real while the proposed SEO explanation is wrong.

Nothing moves yet

Compare the elapsed time with the declared recrawl and buying-cycle window. Verify access, indexation, measurement coverage, and intervention quality before either waiting longer or declaring failure.

Human writers and AI agents

Hold agent-written content to the same evidence bar

Content produced by an agent is not automatically good, and saying otherwise would undermine every number on this page. It is measured exactly like everything else.

Same baseline, same window, same attribution

An agent-produced page gets a baseline before publication, a stated comparison window, and the same honest account of what else moved in that period.

Nothing about the authoring method changes the measurement contract. If anything it raises the bar, because volume makes it easier to mistake activity for progress.

Attribute to the component, not the tool

Results here are tied to the part of the system that produced them — a post type rolled out, an element retrofitted across an existing corpus, a phase run in the right order.

That is a more useful claim than crediting automation in general, and it is one you can actually test on your own site.

Consistency is measurable on its own

Because elements are typed, you can measure the corpus directly: share of pages carrying a sources block, spec conformance by post type, element coverage, freshness distribution.

These move before traffic does, which makes them the earliest honest signal that a content system is working.

What we would not publish

A win from a single page presented as a sitewide result. A window chosen after the fact because it looked good. A metric selected because it moved.

Those exclusions apply identically to human and agent output, and stating them is the reason the rest of the figures are worth reading.

Read the measurement rules before the outcomes. A result is only useful if you can reconstruct how it was produced and decide whether the same approach would work on your site.

Start with evidence

Start with your baseline

Define the scope, source, filters, observation window, exclusions, and decision rule. Then run the visibility check and preserve the first export. A result becomes defensible when another person can reconstruct it.

Read how we measure results

177× organic traffic growth with a one-person content operation HZ-Containers measured the change against its pre-FlowHunt baseline across a 16-month Search Console evidence window. It is a documented case result, not a benchmark or promise. Inspect the case evidence

Establish the baseline before you change the site

Free check · 7-day trial · no credit card