How to Know If Your SEO Is Working: A Practical 2026 Measurement Guide

SEO is working when qualified search visibility turns into meaningful business outcomes—not merely when a ranking tool shows more green arrows. The evidence usually appears in sequence: important pages are indexed, relevant impressions grow, searchers click, qualified visitors engage, key events increase, and organic revenue or leads eventually exceed the cost of the work.

That answer sounds simple. Measuring it is not. Search demand changes. Rankings move by device and location. AI Overviews can satisfy some searches before a click. Google Search Console and Google Analytics count different things. A lead may discover you through an article, return through a branded search, and convert days later through email. If you judge SEO from one chart, you can easily reward the wrong work—or cancel a strategy that is starting to gain traction.

This guide gives you a practical way to decide whether SEO is working, what to check in Search Console and GA4, how to read conflicting signals, and what to ask an agency before you approve another month of spend.

The short answer: look for a connected chain of evidence

Strong SEO produces movement across three levels:

1. Search visibility: priority pages are crawlable and indexed; relevant impressions, non-branded query coverage, and rankings improve.
2. Qualified acquisition: clicks and organic sessions reach the pages that can educate or convert the right audience.
3. Business impact: organic visitors complete valuable actions—qualified forms, calls, trials, purchases, bookings, subscriptions—and generate revenue or pipeline.

No single metric proves the whole case. A higher ranking with no impressions may target a query nobody searches. More traffic with no relevant conversions may reflect weak intent. More conversions with flat traffic, on the other hand, can be a genuine win: the site may be attracting a smaller but better-qualified audience.

Think of SEO as an evidence chain:

Indexed → visible → clicked → engaged → converted → profitable

The farther a signal sits to the right, the closer it is to business value. Early-stage signals still matter because they show whether the work is moving in the right direction before revenue has had time to follow.

First, define what “working” means for your business

Before opening a dashboard, write down the outcome SEO is expected to support. “Increase organic traffic” is not specific enough. Traffic is an input. The business probably needs something closer to one of these:

• Generate qualified consultation requests for two high-margin services.
• Increase non-branded ecommerce revenue while maintaining contribution margin.
• Grow demo requests from mid-market buyers in the United States.
• Increase calls, direction requests, and booked appointments from a defined local service area.
• Build an audience of relevant subscribers who later enter a sales sequence.

Then give each outcome a measurement rule. What counts as a qualified lead? Is a phone call valuable only after 60 seconds, or only after your team marks it qualified in the CRM? Will refunded orders be removed from organic revenue? Which geographies and product lines matter?

This prevents a common reporting trick: presenting whatever improved as if it were the original goal. If the assignment was to sell commercial roofing projects in Phoenix, a traffic surge to a nationwide article about DIY gutter cleaning is not success just because the sessions came from Google.

Choose one primary outcome and a small set of supporting metrics

A useful scorecard has one primary business KPI and a handful of diagnostic metrics.

For a lead-generation company, the primary KPI might be sales-qualified leads from organic search. Supporting metrics could include non-branded clicks to service pages, form-start-to-submit rate, tracked calls, and organic-assisted pipeline.

For ecommerce, the primary KPI might be organic contribution margin rather than gross revenue. Supporting metrics could include non-branded product-page clicks, add-to-cart rate, purchases, average order value, and new-customer revenue.

For a publisher, the primary KPI may be qualified subscribers or ad revenue, supported by search clicks, engaged sessions, returning readers, and newsletter signups.

You do not need 30 KPIs. You need enough information to locate the break in the chain when the primary outcome stalls.

Build a measurement setup you can trust

Bad measurement can make sound SEO look weak and weak SEO look impressive. Check the instrumentation before judging the strategy.

Use Search Console for what happens before the visit

Google Search Console reports how your site performs in Google Search: queries, pages, countries, devices, search appearances, impressions, clicks, click-through rate, and average position. It is the closest first-party view of how Google Search sends visibility and visits to your site. Google's Search Console metrics documentation^[1]Source 1Google Search Console Help. What Are Impressions, Position, and Clicks?View source ↗ explains how impressions, clicks, and position are counted.

Verify the correct property. A Domain property is usually preferable when you need all protocols and subdomains together. Confirm that the XML sitemap is current, priority pages are indexable, and the Page Indexing report is not showing a new pattern of accidental exclusions.

Search Console is not a revenue platform. It can show which query or page earned a click, but it cannot tell you whether that visitor became a customer. That is the job of analytics, your CRM, your ecommerce platform, and—when calls matter—a call-tracking system.

Use GA4 for what happens after the visit

In Google Analytics 4, identify the actions that matter and mark the appropriate events as key events. Typical examples include a completed lead form, a qualified booking, a purchase, a trial activation, or a subscription. Do not mark every button tap as a key event. Inflating the count makes the report look busy while hiding actual performance. See Google's current definition of a GA4 key event^[2]Source 2Google Analytics Help. About Key Events in Google Analytics 4.View source ↗.

Audit the implementation:

• Test forms, thank-you pages, checkout events, and phone-click tracking yourself.
• Exclude internal and developer traffic where appropriate.
• Preserve UTMs during redirects and cross-domain journeys.
• Check whether a consent platform or browser restriction is creating material gaps.
• Connect Search Console and GA4, while expecting the numbers to differ.

Search Console clicks and GA4 sessions are not interchangeable. One is a search-result interaction; the other is a measured period of onsite activity. Time zones, consent, tag failures, attribution, repeat visits, and calculation rules create differences. Compare their trends rather than forcing the totals to match. Google's own guide to using Search Console and Analytics together^[3]Source 3Google Search Central. Google Search Console and Google Analytics Integration Guide.View source ↗ makes the same distinction.

Connect revenue and lead quality

For lead generation, send the original source and landing-page information into the CRM. Then report not only submitted forms but also qualified leads, opportunities, closed deals, and revenue. If the average sales cycle is 90 days, this layer will lag behind clicks and forms.

For phone-driven businesses, use dynamic number insertion where legally and operationally appropriate. Separate a phone-button click from a connected call, and a connected call from a qualified call. For ecommerce, verify purchase values, refunds, taxes, shipping, and currency handling before calculating ROI.

Without this layer, you can measure search activity. You cannot confidently measure the commercial return from SEO.

The 10 signals that reveal whether SEO is working

1. Your important pages are eligible to appear

Indexation is a prerequisite, not a victory. Check representative money pages and newly published resources with URL Inspection. Confirm that Google selected the intended canonical URL, can access the page, and is not blocked by a noindex directive, robots rule, authentication wall, or rendering problem.

Do not celebrate simply because the total number of indexed URLs increased. A growing index can include parameter pages, duplicates, internal-search results, tag archives, or thin location pages. The better question is: Are the pages we deliberately want searchers to find indexed, while low-value duplicates stay out?

2. Relevant impressions are growing

Impressions are often the earliest visible sign of traction. A new page may first appear in low positions for several related queries, then accumulate impressions, then earn clicks as its ranking and search presentation improve.

In Search Console, compare at least two useful periods and segment the data. Look at non-branded queries, the target country, the dominant device, and the specific page group you changed. Year-over-year comparisons help with seasonality; recent-period comparisons help detect the effect of a new release. Use both when possible.

Rising impressions are encouraging only when the queries match the audience and offer. If irrelevant impressions expand, the page may be semantically unfocused. If impressions rise while clicks stay flat, do not immediately rewrite the whole article. First inspect position, query mix, SERP features, title, snippet, and search intent.

3. You rank for the right query set—not just one trophy keyword

Rankings remain useful, but average position is a diagnostic measure, not a profit-and-loss statement. It can blend countries, devices, result types, and queries with very different commercial value. A rank tracker adds consistency for a controlled keyword set, but it still cannot represent every searcher or every long-tail query.

Group tracked keywords by intent and business value:

• Problem-aware: “why is my warehouse roof leaking”
• Solution-aware: “commercial roof leak repair”
• Vendor-aware: “commercial roofing company phoenix”
• Branded: your company, products, and executives

Improvement across a relevant cluster is stronger evidence than one number-one ranking. Separate branded from non-branded performance. Branded clicks are valuable, but they may be driven by offline campaigns, PR, referrals, or existing customers rather than non-branded SEO discovery.

4. Organic clicks and sessions reach the right landing pages

Traffic growth matters when it reaches pages connected to the customer journey. In Search Console, review clicks by page. In GA4, open the landing-page report and filter to the organic search channel. Tag page types—service, product, category, comparison, educational, location, support—so you can see which part of the site is growing.

Here is a realistic hypothetical pattern:

A software company’s organic sessions rise 38%, but almost all growth comes from a glossary page attracting students. Demo requests do not move. The SEO activity generated visibility, but it did not yet create commercial impact. The next step is not to call the entire program a failure; it is to improve topic-to-product pathways, prioritize high-intent pages, and measure assisted conversions.

The reverse can also happen. Organic sessions fall 8% after low-value pages are consolidated, while demo requests rise 14%. If tracking and demand are stable, that can represent a healthier acquisition mix.

5. Search click-through rate improves where you have a real opportunity

CTR equals clicks divided by impressions. It is affected by ranking, query, brand recognition, device, search features, title, snippet, and the presence of ads or AI results. There is no universal “good SEO CTR.”

Find pages with meaningful impressions and positions where a stronger search presentation could plausibly earn more clicks. Review the actual queries. Does the title answer the dominant need? Is the promise specific? Does the page look current without using a cosmetic date change? Could Google be selecting an unhelpful snippet because the opening is vague?

Do not use a sitewide CTR decline as proof that titles failed. A page can gain many new impressions at lower positions, which lowers aggregate CTR even while total clicks increase. That is normal expansion, not necessarily deterioration.

6. Visitors complete meaningful onsite actions

Engagement metrics help explain traffic quality, but they must match the page’s job. A visitor can read a concise answer, click a phone number, and leave quickly after a successful visit. Another visitor can spend eight minutes confused on a poorly structured page.

Use task-specific signals:

• On an article: scroll depth, useful internal-link clicks, video completion, resource download, newsletter signup.
• On a service page: CTA clicks, form starts, completed forms, calls, bookings.
• On ecommerce pages: product views, add-to-cart events, checkout starts, purchases.
• On support content: successful self-service, reduced repeat search, or a resolved support path.

Bounce rate and engagement time can point to a problem. They do not diagnose it alone, and they should not be presented as direct proof that Google “likes” or “dislikes” a page.

7. Organic conversions and conversion quality improve

This is where SEO moves from visibility to value. Report conversion counts and rates by landing page, page type, device, country, and branded versus non-branded entry where your data allows.

For leads, examine acceptance and close rates. Ten forms from bots or job seekers are not better than three requests from ideal customers. For ecommerce, examine new-customer mix, margin, refunds, and repeat behavior. For subscriptions, distinguish a low-commitment signup from an activated or retained user.

Use both last-click and assisted views. Last-click reporting is simple but can under-credit early discovery. A person may first discover a detailed guide through non-branded search, later return by searching the brand, and finally convert through an email. The original article influenced the journey even if it did not receive the final conversion credit.

8. Revenue, pipeline, or customer value grows relative to cost

The cleanest business test is incremental value, not raw traffic. At a basic level:

SEO ROI = (organic profit attributable to SEO − SEO cost) ÷ SEO cost × 100

Use profit or contribution margin when possible, not gross revenue. Include agency or staff time, content, development, tools, digital PR, and other material costs. Avoid claiming every organic sale as incremental SEO value; some branded demand would have existed without the current program.

Consider a clearly labeled hypothetical example. A company spends $12,000 over a quarter on SEO. During the same period, tracked organic leads create $36,000 in gross profit. If the attribution is sound, simple ROI is:

($36,000 − $12,000) ÷ $12,000 × 100 = 200%

That calculation is easy. Proving which profit is incremental is harder. Compare against a sensible baseline, account for seasonality, annotate major campaigns, and, for mature programs, use tests where practical: phased rollouts, page cohorts, matched markets, or time-series analysis.

9. Local search produces real-world actions

For a local business, website sessions are only part of the story. Track Google Business Profile^[4]Source 4Google Business Profile Help. Google Business Profile Performance Metrics and Insights.View source ↗ discovery queries and interactions alongside organic landing-page performance. Calls, website clicks, bookings, messages, and direction requests can matter more than a conventional website conversion.

Use location-specific rank tracking as a diagnostic because proximity changes local results. Do not rely on one incognito search from the office. Review performance across the service area and tie actions back to qualified calls, appointments, or store visits where measurement allows.

10. Your visibility extends into AI search features and diverse result formats

In 2026, a ranking can improve while classic clicks remain under pressure because the search results page itself is changing. Google’s AI Overviews and AI Mode can surface supporting links, and Search Console has reporting for generative AI performance. Track this visibility alongside Web search rather than inventing a separate “AI SEO” success metric.

The fundamentals still apply: the page must be indexable and eligible to appear with a snippet; important content should be available in text; internal links should make it discoverable; the page should offer a strong experience; and supporting images or video should be genuinely useful. Those are the same foundations Google lists in its guidance for AI features in Search^[5]Source 5Google Search Central. AI Features and Your Website.View source ↗.

Do not buy “secret GEO hacks.” Google's current generative AI optimization guide^[6]Source 6Google Search Central. Google's Guide to Optimizing for Generative AI Features on Google Search.View source ↗ says there is no special AI schema requirement, no need to rewrite everything into tiny chunks, and no Google Search benefit from an llms.txt file. The durable advantage is non-commodity material: original analysis, first-hand evidence, clear sourcing, expert review, and details that a generic summary cannot supply.

How to check whether SEO is working in Google Search Console

Use this monthly workflow. Save the filters or export the data so the comparison stays consistent.

1. Set a meaningful date comparison. Start with the last 28 days versus the previous 28 for recent movement, then check year over year for seasonality. For small sites, 90-day windows may be more stable.
2. Separate brand and non-brand. Use a query filter or a carefully maintained regex for the brand name, misspellings, products, and executive names. Review both groups.
3. Match the market. Filter to the target country and compare mobile with desktop. Do not let irrelevant international impressions hide local performance.
4. Review pages by business role. Inspect priority service, product, category, comparison, location, and supporting editorial pages separately.
5. Read clicks, impressions, CTR, and position together. One number without the others invites the wrong diagnosis.
6. Inspect queries for each priority page. Look for intent drift, cannibalization, new long-tail coverage, and queries sitting just outside strong click positions.
7. Check search appearance and AI performance. Identify whether rich formats, video, products, or generative AI visibility contribute to the trend.
8. Annotate releases and incidents elsewhere. Search Console is not a project log. Maintain dates for migrations, templates, content updates, outages, campaigns, and major Google updates.

Export data when the interface is too limiting. Larger teams can combine Search Console, GA4, CRM, and cost data in Looker Studio or a warehouse. The dashboard should preserve the definitions behind every chart.

How to check whether SEO is working in GA4

Search Console tells you what happened before the click. GA4 helps you understand what happened after it.

1. Open Reports → Acquisition → Traffic acquisition and review Organic Search sessions and key events.
2. Open the Landing page report, filter to Organic Search, and compare the pages where sessions began.
3. Add business-relevant dimensions: device, country, new versus returning user, or content group.
4. Compare key-event counts and rates, not only sessions.
5. For ecommerce, review purchases and revenue; reconcile them with the commerce platform.
6. Use Explorations for a path from organic landing page to product, form, booking, or purchase.
7. Review attribution reports to understand assisted journeys, while keeping one agreed model for executive reporting.

If organic traffic suddenly drops in GA4 but Search Console clicks remain stable, suspect measurement before rankings. Check tags, consent behavior, channel grouping, redirects, cross-domain configuration, and site releases. If both fall, then investigate demand, indexing, rankings, SERP changes, competitors, and technical health.

How to interpret conflicting SEO metrics

SEO reports rarely move in a neat straight line. Use the pattern, not your first emotional reaction.

What you see	Likely interpretations	What to check next
Impressions up, clicks flat	New lower-position coverage; weaker CTR; more zero-click or AI-result exposure; intent drift	Query mix, position bands, SERP features, title/snippet, device
Rankings up, impressions flat	Search demand is low or falling; tracker watches unrepresentative terms	Search Console impressions, Trends, broader cluster, target market
Clicks up, conversions flat	Informational growth; mismatched intent; weak offer or UX; broken conversion tracking	Landing pages, page type, key-event QA, CTA path, lead quality
Traffic flat, conversions up	Better qualification, CRO, or high-intent visibility	Conversion rate, query mix, margin, assisted conversions
GA4 organic down, Search Console stable	Analytics or consent issue; session definitions; channel misclassification	Tagging, consent, redirects, source/medium, release log
Search Console and GA4 both down	Demand shift, technical problem, lost rankings, SERP change, seasonality	Indexing, site changes, Trends, competitors, queries and pages
Indexed pages up, performance flat	New pages lack demand or quality; index bloat; weak internal linking	URL sample, page purpose, impressions, canonicals, internal links
Backlinks up, outcomes flat	Links may be irrelevant, low quality, too recent, or pointing to the wrong pages	Referring-domain relevance, target URLs, referral traffic, rankings

How long should SEO take before you judge it?

There is no responsible universal deadline. A technically blocked site can improve within days of a high-impact fix being crawled. A new site in a competitive market may need many months to build coverage, reputation, and demand. A mature site can see a measurable response to a title change or internal-link improvement much faster.

Judge different work on different clocks:

• Days to weeks: tracking repairs, crawl access, indexation of a changed page, obvious technical errors.
• Weeks to a few months: impressions, long-tail query coverage, CTR response, movement for existing pages.
• Several months or longer: competitive rankings, sustained non-branded traffic, qualified pipeline, revenue, authority-building effects.

These are observation windows, not guarantees. Agree on expected leading indicators before the work begins. If an agency says “SEO takes time” but cannot show completed work, baseline data, page-level movement, or a testable plan, patience is being used as a shield.

A practical monthly SEO scorecard

Keep the executive view short. Add a diagnostic appendix for the team doing the work.

Layer	Metric	Segment	Comparison	Owner
Business	Qualified organic leads, profit, pipeline, or contribution margin	Product/service and market	YoY plus target	Marketing + sales/finance
Conversion	Organic key events and rate	Landing-page type, device	YoY and recent period	Analytics/CRO
Acquisition	Non-branded organic clicks and sessions	Market and page group	YoY and recent period	SEO
Visibility	Impressions, query coverage, priority rank distribution	Non-brand cluster	YoY and release baseline	SEO
Eligibility	Indexability and critical technical errors	Priority templates	Release baseline	Engineering/SEO
Modern search	Generative AI and rich-format visibility	Topic/page group	Trend	SEO/content

Every report should answer four questions in plain language:

1. What changed?
2. Why do we believe it changed?
3. What business effect did it have—or what leading signal moved?
4. What will we do next, by whom, and by when?

That is more useful than a PDF filled with tool screenshots and no decisions.

How to tell whether an SEO agency or consultant is doing a good job

Good SEO work is visible even before every outcome matures. You should receive a strategy tied to business goals, a prioritized backlog, a record of shipped changes, access to first-party data, and explanations that connect work to expected signals.

Ask these questions:

• Which pages, audiences, and business outcomes are we prioritizing—and why?
• What changed this month? Show the URLs, tickets, content briefs, technical fixes, or outreach evidence.
• Which leading indicators should move first?
• How are branded and non-branded results separated?
• How are leads qualified and revenue attributed?
• What failed or underperformed, and what did we learn?
• Which recommendations depend on third-party metrics, and how were they validated against first-party data?
• What will you stop doing because the evidence is weak?

Red flags include guaranteed number-one rankings, secrecy about changes, reports dominated by Domain Authority or total ranking keywords, bulk low-quality links, unexplained traffic from irrelevant markets, a rising indexed-page count presented as success, and conversions that no one has tested.

Google explicitly advises businesses to ask prospective SEOs what results they expect, in what timeframe, and how they measure success. It also warns that third-party tools do not have access to Google's internal ranking data. A trustworthy practitioner will not treat a proprietary score as Google's verdict. See Google's current guidance on hiring and evaluating an SEO^[7]Source 7Google Search Central. Do You Need an SEO? Hiring and Evaluating Guidance.View source ↗.

Common measurement mistakes to avoid

Comparing the wrong periods

Monday versus Sunday, a 28-day period versus a 31-day period, or December versus November can create false narratives. Use same-length comparisons and year-over-year views where seasonality matters. Annotate promotions, PR, migrations, outages, and tracking changes.

Treating correlation as causation

Traffic rose after an article was updated. That does not prove the update caused all of the growth. Demand, competitors, links, algorithmic changes, or other campaigns may have contributed. Use page cohorts, controlled rollouts, and corroborating query data to strengthen the inference.

Reporting sitewide averages

Sitewide numbers can hide the success of one product line and the decline of another. Segment by market, intent, page type, brand, and device. Page-level evidence is usually more actionable.

Using third-party authority scores as KPIs

Domain Authority, Domain Rating, Authority Score, and similar metrics can help compare link profiles within the same tool. They are not Google metrics and are not business outcomes. Use them as directional diagnostics, never as proof that SEO is working.

Assuming every backlink is progress

One editorial link from a relevant industry resource may matter more than hundreds of scraped directory links. Review relevance, editorial context, target page, referral activity, and patterns that could indicate manipulation.

Confusing content volume with content value

Publishing 20 articles is output. Earning qualified visibility, links, subscriptions, assisted conversions, or sales is an outcome. A smaller library of distinctive, maintained resources can outperform a large set of repetitive pages.

Ignoring SERP and demand changes

A stable ranking can receive fewer clicks when demand falls or the result page adds ads, video, forums, products, or an AI answer. Measure impressions, CTR, result format, and downstream value together.

A 30-day plan to stop guessing

Week 1: Define and verify. Choose the primary outcome, document metric definitions, test all key events, confirm Search Console coverage, and establish a baseline by page type and non-branded query cluster.

Week 2: Segment and diagnose. Find the pages with the most commercial potential, not merely the most traffic. Identify high-impression/low-click opportunities, pages that attract irrelevant queries, and conversion paths with obvious friction.

Week 3: Ship focused improvements. Improve search titles and openings where intent is clear, strengthen the content that differentiates the page, add useful internal links, fix tracking and technical blockers, and align calls to action with the visitor’s stage.

Week 4: Record and schedule the readout. Annotate every material change. Request recrawling only where appropriate. Set the first readout based on the type of change, and state which leading signal should move before conversions.

Do not change ten variables on every page at once. A smaller, logged rollout makes the result easier to interpret and the next decision easier to defend.

The standard that matters

SEO is not working because a dashboard looks active. It is working when the right people discover the right pages, choose your result, complete a useful journey, and create measurable value at a sustainable cost.

Start with the business outcome. Trace backward through conversions, qualified visits, clicks, visibility, and indexation. When a number moves, ask what part of that chain it represents—and what evidence would confirm the story.

If your current report cannot answer those questions, the next SEO task is not another blog post. It is fixing the measurement system.

Frequently Asked Questions

What is the clearest sign that SEO is working?

The clearest sign is sustained growth in qualified organic conversions, revenue, profit, or pipeline. Earlier signals—relevant impressions, non-branded clicks, and traffic to priority pages—show progress, but business outcomes prove value.

Can SEO be working if organic traffic is down?

Yes. Traffic can fall while conversions or profit rise if low-value visits decline and high-intent visibility improves. Confirm that tracking is sound, compare query and landing-page mix, and check whether demand or SERP features changed.

Can traffic increase while SEO is failing?

Yes. A site can attract more visitors from irrelevant queries, countries, or informational pages that never support the business. Evaluate landing pages, intent, lead quality, assisted conversions, and profit—not traffic alone.

How often should I measure SEO?

Monitor critical errors and tracking continuously, review operational metrics weekly, and make strategic decisions monthly or quarterly depending on traffic volume and sales-cycle length. Small datasets need longer windows to avoid reacting to noise.

Is Domain Authority proof that SEO works?

No. Domain Authority and similar scores are third-party estimates, not Google metrics. They can support competitor or link analysis within the same tool, but first-party search performance and business outcomes should lead the scorecard.

Is bounce rate an SEO ranking factor?

Do not treat GA4 bounce rate as a direct Google ranking factor. Use it as a contextual onsite diagnostic. A high bounce rate can represent dissatisfaction, a measurement problem, or a successful one-page answer; the page’s task determines the meaning.

Why are Search Console clicks different from GA4 sessions?

They measure different events. Search Console counts eligible clicks from Google Search, while GA4 counts measured sessions on the site. Consent, blocked tags, attribution, time zones, repeat visits, and implementation differences prevent an exact match.

How do I measure SEO for a local business?

Combine Search Console and GA4 with Google Business Profile interactions, call tracking, bookings, direction requests, and qualified offline outcomes. Segment by service area and do not rely on one manual search from one location.

How do I measure SEO visibility in AI Overviews or AI Mode?

Use Search Console’s current generative AI reporting and overall Web performance, then evaluate the landing pages and downstream actions those experiences generate. Focus on useful, original, crawlable content; no special AI schema or llms.txt file is required for Google Search.

When should I replace my SEO agency?

Consider a change when the agency cannot show what it shipped, will not provide data access, relies on irrelevant vanity metrics, uses risky link schemes, guarantees rankings, or repeatedly misses agreed leading indicators without a credible diagnosis and corrective plan.

Authoritative Sources

1. Google Search Console Help. What Are Impressions, Position, and Clicks?.

2. Google Analytics Help. About Key Events in Google Analytics 4.

3. Google Search Central. Google Search Console and Google Analytics Integration Guide.

4. Google Business Profile Help. Google Business Profile Performance Metrics and Insights.

5. Google Search Central. AI Features and Your Website.

6. Google Search Central. Google's Guide to Optimizing for Generative AI Features on Google Search.

7. Google Search Central. Do You Need an SEO? Hiring and Evaluating Guidance.