How to Measure SEO Results — KPIs, Attribution, and Real Outcomes

Most SEO reports tell you everything except whether the engagement is working. Forty-five slides of impressions, domain authority deltas, and “average position improved by 4.7” — zero answers on revenue.

This page is the antidote. The five-number summary that fits on one page — sessions to high-intent landing pages, qualified leads, cost per organic lead, top ranking gains that drove traffic, top ranking losses that need diagnosis. The KPI matrix below that, segmented by funnel stage and query type. Three anonymized Arizona outcomes with real dollar numbers. The 2026 AI Overview citation KPI nobody else is tracking yet, with the actual citation-log mechanic.

If you have been reading agency reports that leave you with the same question every month — did this actually work — this is the framework that answers it in ten minutes.

The three anonymized outcomes on this page are real Rule27 Arizona engagements. The home-services client added $5.2M in annual revenue over a nine-month engagement, against a $67,500 cumulative retainer — a 77:1 return inside year one. The Phoenix dental practice grew local pack impressions 412% across three East Valley locations in six months, with measurable new-patient appointment attribution once CallRail was installed mid-engagement. The legal services client earned the #1 AI Overview citation on a commercial divorce-attorney money keyword inside 90 days of a focused AI search optimization layer.

What unites the three is the measurement discipline. Each engagement started with documented baselines, ran a consistent attribution model, separated leading from lagging indicators, and closed with a year-end ROI reconciliation against signed-day numbers. None of the three relied on cherry-picked metrics to look good in a monthly report. The numbers came from the client's own GA4, Search Console, and CRM — every figure published here is auditable inside their own dashboards.

The Arizona context matters because the playbook variation is local. Heat-seasonal demand cycles for home services. Snowbird-driven patient flow for healthcare. The Phoenix Business Journal and AZBigMedia editorial relationships that compound the citation surface for AI search. National agencies running a generic playbook with a Phoenix landing page do not have the local citation graph that produces these outcomes. Geographic credibility is the precondition for results in this market specifically.

Most SEO reports tell you everything except whether the engagement is working. Forty-five slides of impressions trends, domain authority deltas, and “average position improved by 4.7” — zero answers on the only question that should anchor every SEO conversation: did the work move the business forward, and what changes next month?

This is not because the data is hard to find. GA4, Search Console, CallRail, your CRM, and a handful of free dashboards already publish every number you need. The problem is that tool-vendor blogs and agency reports are commercially aligned with making seo results look like a sophisticated discipline that requires their software to interpret. So you get the listicle of fifteen KPIs that all sound important, none of which tell you whether to renew the retainer.

The honest answer is the matrix on this page. Five business numbers in one summary view. Twenty operational KPIs below that, grouped by funnel stage and query type so the right metric is the one you read for the right decision. Three anonymized Arizona outcomes with real revenue numbers, because abstract benchmarks are not proof. And a section on the 2026 KPI almost nobody is tracking yet — AI Overview citation, the new featured snippet equivalent for the era where 40% of commercial queries surface an AI answer above the blue links.

The two questions every SEO report should answer

There are only two. Did the work move the business forward? And what changes next month based on what we just learned? Everything else is decoration.

If the report cannot answer the first question in plain English in the first paragraph, the metrics are wrong or the agency is hiding behind them. If the report cannot answer the second question, the agency is running on autopilot and nothing is being learned from the data they are paid to collect.

Google's own former search developer advocate Maile Ohye said the timeline on camera years ago: SEO needs four months to a year to show real benefits. That is the floor for the first question. Asking did the work move the business in month two is asking for results before the work has had time to compound; the honest answer is the foundation is laid, here are the leading indicators, the lagging indicators arrive in month four. Asking the same question in month nine and getting hand-waved is a different problem entirely.

The matrix below is what we use internally to answer both questions for every Rule27 client. It is not proprietary. The best operators in this industry use a version of it. The tool-vendor blogs we cite at the bottom of this page each get part of it right. We have not seen anyone publish the full matrix in one place, so we did.

The five-number summary that fits on one page

The monthly summary view at Rule27 has five numbers. They tell you within ten minutes whether the engagement is working and what to discuss on the next strategy call.

Organic sessions to high-intent landing pages. Not total organic traffic. Total organic traffic includes the long-tail informational queries that may never convert, the branded searches that were already going to convert, and the irrelevant impressions that came in from queries you do not actually compete on. Filter for sessions landing on service pages, pricing pages, contact pages, and the top three blog posts that actually feed conversions. That is the real organic number.

Qualified leads from organic. GA4 conversion events from organic search, filtered for the lead-quality definition agreed at kickoff (form completion with required fields, phone call duration over 60 seconds, chat session resulting in an appointment, etc.). For SMBs without a CRM, this is GA4 plus CallRail plus your form-fill records reconciled monthly. For B2B with a CRM, this is HubSpot or Salesforce filtered for source = organic search.

Cost per organic lead. Total monthly SEO retainer divided by qualified leads attributed to organic. This is the simplest unit-economic metric for whether SEO is paying for itself. A $5,000/month retainer producing 25 qualified leads is $200/lead. The same retainer producing 5 leads is $1,000/lead. The number is interpreted against the lifetime value of a closed customer, which is why the year-end ROI conversation matters.

Top 10 ranking gains that drove traffic. Not every ranking gain matters. A jump from position 47 to position 14 looks like progress on a Semrush export, but page 2 of Google receives roughly 1.5% of clicks. The report should list the keywords where ranking gains resulted in measurable session lift on the corresponding landing page. Those are the wins worth celebrating.

Top 10 ranking losses that need diagnosis. Equal weight to losses. Rankings drop for reasons — Google core update, competitor publishing better content, technical regression on the page, intent shift in the SERP. Every loss is a question the next month's work should answer. An agency that does not surface the losses is hiding diagnostic work it should be doing.

Those five numbers go on one page. The exec who pays the bill reads them in ten minutes. The strategy call walks through the why behind each. Everything else is supporting detail in the dashboard, available on demand but not the headline.

The KPI matrix by funnel stage

The five numbers above are the summary. Below them, in the operational dashboard, the KPIs split by funnel stage. Each stage answers a different question for a different audience inside the client organization.

Awareness KPIs

Question this stage answers: are we reaching the right audience at all?

Impressions in Search Console are the leading indicator. They show how often pages from the site were displayed in search results, segmented by query. The number on its own is not meaningful — a million impressions on irrelevant queries is worse than ten thousand impressions on commercial-intent queries. Impressions matter as a segmented number, not an aggregate.

AI Overview citation count is the new awareness KPI for 2026. Semrush's AI visibility tracker, Ahrefs's Brand Radar, and manual SERP probing each surface a different slice of the citation log — which queries inside which engines (Google AI Overviews, Bing+ChatGPT, Perplexity, Gemini) are citing your pages. We run the citation log monthly for clients on the AI search engagement layer. The number that matters is the count of cited queries that match commercial intent. Citations on informational long-tails are a leading indicator; citations on “best [your service] [city]” are the lagging indicator that the AI search work is producing buyer awareness.

Share of voice across the target keyword basket is a tool-reported aggregate (Semrush, Ahrefs, Sistrix). It estimates what percentage of the total clicks across your target keyword set are landing on your site versus competitors. A reasonable benchmark for an established player is 15-25% of the addressable share of voice in a category; for a new entrant, the goal is incremental month-over-month gains, not a target percentage.

Branded search volume, tracked in Search Console under brand queries, is the lagging indicator of awareness work. It is the slowest-moving metric on the dashboard and the most expensive to fake. If branded volume is up year-over-year, the awareness work is compounding. If branded volume is flat after twelve months of SEO investment, the awareness work is not reaching the right audience or the audience is forgetting the brand between visits.

Consideration KPIs

Question this stage answers: when we reach the audience, do they engage?

Clicks from Search Console are impressions converted to visits. The relationship between impressions and clicks is mediated by click-through rate, which is the most diagnostic single number in this stage. CTR by query intent matters more than aggregate CTR. A commercial-intent query at position 3 with a 2% CTR is a problem — the title tag, meta description, or SERP snippet is failing to compete. A long-tail informational query at position 8 with a 12% CTR is performing above baseline.

GA4 engagement rate has replaced bounce rate as the default engagement signal. Benchmarks vary by page type: 55-65% engagement rate is healthy for blog posts, 70-80% for service landing pages, 80%+ for product pages with clear next-actions. Engagement rate below the benchmark suggests the page is not delivering what the query intent expected.

Scroll depth, tracked through GA4 enhanced measurement or a GTM custom event, indicates whether the content depth is being consumed. Pages with low scroll depth in the 25-50% range tend to have an intro problem — the first screen is not setting up the reader for the deeper content. Pages with scroll depth concentrated in the 75-100% range are doing the writing work, regardless of whether the page is converting.

Time on page, paired with scroll depth, distinguishes engaged readers from confused ones. High time + high scroll is engagement. High time + low scroll is confusion (the reader is searching the page for something not present). Low time + low scroll is a misaligned query intent — the wrong audience arrived, or the title oversold the content.

Action KPIs

Question this stage answers: when the audience engages, do they convert?

Qualified lead count by source-medium is the headline number. The qualification definition is set at kickoff and should be consistent across reports. A qualified lead might be a form completion with required fields, a phone call over 60 seconds, an inbound email from a domain in your ICP, or a chat session that produces an appointment booking. Whatever the definition is, the report should use the same definition every month so trend lines are comparable.

Organic conversion rate is qualified leads divided by organic sessions on the same time window. For SMBs, conversion rate ranges widely — 1-3% for top-of-funnel content pages, 3-8% for service landing pages with clear CTAs, 8-15% for high-intent pages like pricing or contact. A page below its benchmark is either a CRO opportunity or a query-intent mismatch.

Cost per organic lead is the unit-economic metric. Total monthly retainer divided by qualified organic leads. This number is interpreted against customer lifetime value to determine whether SEO is paying for itself. For service businesses with average customer value over $5,000, a CPL up to $250-$500 is usually profitable. For e-commerce with average order value under $100, the same CPL is unsustainable; the playbook needs to compound traffic volume rather than chase narrow intent.

Form-completion funnels, tracked in GA4 funnel exploration, show where prospects drop out of the lead capture. If the form is being viewed but rarely completed, the form is the friction point. If the page is being visited but the form is rarely viewed, the page is not selling the offer. These funnels are how you distinguish a content problem from a CRO problem.

Revenue KPIs

Question this stage answers: when prospects convert, do they generate revenue?

Closed-won attribution to organic source-medium is the CRM-side number. HubSpot and Salesforce both support multi-touch attribution natively; the SMB workaround is to tag the original source on lead capture and carry it through to the closed-deal record. The dollar value of closed-won deals attributed to organic is the cleanest evidence that SEO is paying for itself.

Customer lifetime value (LTV) from organic leads, separated from LTV from paid or referral leads, indicates whether organic acquisition is also producing better customers. Organic leads frequently have higher LTV than paid leads because they self-qualified through deeper content consumption before contacting. If your organic LTV is lower than your paid LTV, something is upside-down in the lead-quality definition or the qualification gating.

Return on SEO investment is the simplest profitability check. Total annual revenue attributed to organic, divided by total annual SEO investment (retainer plus tooling plus content production). A healthy ratio is 5:1 or better for B2B services, 3:1 or better for e-commerce. Below 2:1, the engagement is not paying for itself and the conversation should be about why — wrong keyword targets, wrong page templates, conversion rate too low, retainer too high for the scope.

Incremental organic revenue, the year-over-year delta in organic-attributed revenue minus the year-over-year delta in total SEO spend, is the cleanest measure of marginal SEO ROI. It separates the “traffic was already there” baseline from the “this year's work produced this much new revenue” signal.

Vanity metrics: the four numbers to stop reporting

Four metrics dominate bad agency reports because they are easy to grow and hard to challenge. Each has a structural problem that makes it less informative than the agencies presenting it pretend.

Impressions in aggregate. A million impressions on irrelevant queries is worse than ten thousand on commercial-intent queries. Impressions are useful only as a segmented number filtered by query intent or page type. Aggregate impressions growth is sometimes a sign of useful work and sometimes a sign that Google is showing the site for queries it has no business ranking on. Without intent segmentation, the number does not tell you which.

Domain Authority. DA is Ahrefs's proprietary score, derived from their link graph estimate. It is not a Google ranking factor. Google has explicitly said so, multiple times, in public guidance. DA correlates with rankings because both are downstream of authority signals, but moving DA does not cause Google to rank you better. Agencies that report DA growth as a primary KPI are reporting on a tool's internal score, not on actual SERP performance.

Average position. Search Console's average position aggregates ranking across every query the site shows for. A site ranking position 3 for ten valuable queries and position 80 for ten irrelevant queries has the same average position as a site ranking position 41 for twenty mediocre queries. The number is mathematically meaningless without query segmentation. Agencies that lead with “average position improved by 4.7” are either uninformed or hoping the client is.

Backlink count. The total number of backlinks tells you nothing about the quality of those links. Ten editorial placements from Phoenix Business Journal, AZBigMedia, and ASU are worth more than five hundred forum-comment links. Backlink reports should be filtered by domain quality — DR over 40, traffic over a threshold, topical relevance to the linked page. The aggregate count is the lowest-information version of the backlink metric.

The fix in every case is segmentation. Impressions by query intent. Position by keyword group. Backlinks by domain quality. DA replaced entirely by the underlying authority signals that actually predict rankings. The five-number summary above is what survives the segmentation pass.

SEO attribution: tying organic to revenue

Attribution is where every honest SEO conversation eventually lands. The four common models each get part of the picture; the right answer is usually two models reported side by side, not one chosen as canon.

Last-click attribution gives full credit to the final touchpoint before conversion. In a world where SEO frequently introduces the brand and direct search or branded paid closes the conversion, last-click systematically undervalues SEO. A B2B prospect who reads three pieces of organic content over six weeks, then types your brand into Google and converts on the homepage, gets credited as a direct or branded paid conversion under last-click. The organic touches that built the consideration are invisible.

First-click attribution gives full credit to the first touchpoint. This usually overstates SEO's role, but it captures the awareness work that last-click hides. We typically report first-click alongside last-click for the same time window to show the spread — if last-click attributes $50K to organic and first-click attributes $180K, the truth is somewhere in the middle and SEO is generating more revenue than the last-click number suggests.

Time-decay attribution weights touchpoints closer to the conversion more heavily, while still giving partial credit to earlier touches. This is the default model we recommend for B2B engagements where the sales cycle is longer than 30 days. It captures the SEO-introduces-brand pattern without overstating the first organic touch.

Multi-touch attribution in a CRM (HubSpot, Salesforce, Pipedrive) is the gold standard for SMBs with a defined sales process. Every touchpoint is tracked from first lead capture through closed-won, with attribution credit distributed across the path. The setup work is non-trivial — UTM tagging on every campaign, lead-source field hygiene, integration with GA4 — but the resulting data settles the which channel paid for itself conversation cleanly.

For SMBs without a CRM, the workaround is CallRail-plus-form-plus-chat triangulation. CallRail provides dynamic number insertion so phone calls are attributed to the landing page and source-medium that drove them. Forms carry hidden UTM fields. Chat sessions are timestamped and reconciled monthly against the GA4 source-medium for the session. The triangulated dataset is less elegant than a CRM but more accurate than GA4 alone, which loses attribution data after 14 months and aggregates phone touchpoints to direct.

The gap nobody talks about is AI search attribution. When a prospect asks ChatGPT “best [your service] in Phoenix” and your business is named in the answer, the prospect frequently visits the site by typing the brand into Google, which logs as direct or branded paid in GA4. The AI citation gets zero attribution credit. The workaround is the citation log itself — we maintain it monthly per client, and the correlation between citation events and branded search lift is the closest proxy to attribution available until the AI engines publish referrer headers (Bing+ChatGPT now does this partially; Google and Perplexity do not, as of mid-2026).

Real Arizona SEO results: three anonymized outcomes with numbers

Abstract benchmarks are not proof. The three outcomes below are real Rule27 engagements, anonymized for confidentiality but with the actual numbers that anchored the year-end ROI conversation. Each one closed a different reporting question.

AZ home-services client — $5.2M annual revenue added in 9 months

The baseline going into the engagement was 1,400 organic sessions per month, average order value of $4,800 (a mix of installation and recurring service contracts), and a conversion rate of 1.8% on the service landing pages. The agency they had fired was running a generic content calendar and had not touched the Google Business Profile in fourteen months.

The nine-month transformation broke into three phases. Months 1-3 were GBP rebuild and technical SEO baseline — correct primary category, NAP cleanup across 32 AZ citation directories, Core Web Vitals fixes, and schema markup deployment on the priority pages. Months 4-6 were the long-tail content build — 28 city-by-service combination pages targeting the Phoenix metro suburbs, optimized for both blue-link rankings and AI Overview citation patterns. Months 7-9 were authority work — four editorial placements in AZBigMedia, Phoenix Business Journal, and an ASU faculty research page citing the client's safety data.

The end-of-engagement numbers: 7,800 organic sessions per month (5.6x growth from baseline), conversion rate up to 2.4% (CRO work alongside SEO), $5.2M in incremental annual revenue attributed to organic via multi-touch in their HubSpot. The retainer was $7,500/month, or $67,500 over the nine months — a 77:1 return on SEO investment within twelve months of signing.

The single highest-leverage intervention was the GBP rebuild. The previous agency had set the primary category to a tangentially-related category that under-indexed the client against the actual SERP queries. Within three weeks of the category correction and NAP cleanup, local pack visibility on the priority queries was up 280%.

Phoenix dental practice — +412% local pack impressions in 6 months

The practice had three locations across the East Valley and was ranking on page 2 or worse for “[service] near me” queries at all three. They had no organic-attributed revenue tracking in place; the assumption was that all leads came from referrals and paid search. The engagement question was whether SEO could become the third channel.

The playbook was location-specific. Each of the three locations got its own dedicated landing page with location-specific Core Web Vitals optimization, distinct GBP listing rebuild, and location-targeted content covering the questions specific to that geography (snowbird-season patient flow in the seasonal market, family-density questions in the family-suburb location, etc.). The schema markup was LocalBusiness plus MedicalClinic, deployed correctly for the first time.

Local pack impressions across the three locations rose from a baseline of roughly 12,000/month aggregate to 61,400/month by month six — a 412% increase. Profile clicks went from 540/month to 2,830/month (5.2x). Direction-request clicks (the closest proxy to actual visit intent) went from 180/month to 740/month. The practice installed CallRail mid-engagement and began attributing new patients to organic for the first time — by month six, organic was driving 18 new-patient appointments per month, with an LTV of approximately $1,800 per patient. Annualized, that is $390K in new patient revenue attributable to the SEO engagement, against a $42K/year retainer — a 9.3:1 return.

The lesson from this engagement was the distinction between local pack impressions (cheap to grow with GBP work) and qualified-patient appointments (the lagging indicator that took six months to fully bloom). A monthly report that surfaced only impressions would have looked great in month three and dishonest about whether the engagement was working. The report that surfaced both — impressions as leading indicator, appointments as lagging — is what built trust through the slower months.

AZ legal services client — #1 AI Overview citation on commercial money keyword

This client was a mid-sized Phoenix family law firm whose marketing question was “how do we appear in ChatGPT and Google's AI Overviews when prospects ask for a divorce attorney in Phoenix?” Their blue-link rankings were already strong (positions 4-7 on the priority commercial queries), and the partner had read enough about AI search to know the surface was about to matter more than the blue links.

The engagement was a focused 90-day AI search optimization layer on top of their existing SEO retainer. The work split into three streams. Stream one was Article + FAQPage + Attorney schema deployment on the practice-area pages with verified author and publishing dates. Stream two was the fan-out cluster build — thirteen supporting pages covering the conversational questions a prospect asks before asking “best divorce attorney Phoenix.” Stream three was citation source seeding — publishing original Arizona-specific statistical research (anonymized case-load data) that became citation-worthy primary source material for Google's and Perplexity's models.

By week eleven, the practice was the cited source in Google's AI Overview for the “best divorce attorney in Phoenix” query, with continued citation in Perplexity and ChatGPT (Bing+ChatGPT) on the same query. The blue-link ranking moved from position 4 to position 2 in the same window — a secondary benefit of the schema and content work. Branded search volume rose 31% in the same quarter, which closed the attribution loop: AI Overview cites the firm by name, prospect googles the firm by name, firm acquires the prospect through “direct” in GA4 (which was actually the AI citation chain).

The engagement quantified the previously-invisible AI search referral pattern. New-client matter intake associated with branded search up 22% quarter-over-quarter; the citation log showed 47 distinct AI Overview citation events across the three engines tracked in the same period. The retainer for this 90-day engagement was $9,000 — the firm calculated the resulting matter pipeline at $620K of qualified prospective business, with a typical 18% closed-engagement rate.

The lesson from this engagement was that AI Overview citation is measurable, repeatable, and revenue-relevant when the page architecture is engineered for it. The citation log is the proof. Without the log, every claim about “AI-optimized SEO” is hand-waving.

What to expect at month 3, 6, and 12

Real SEO results follow a curve that does not match the way most engagements are sold. The first sales call promises 30-day results because that is what wins the deal. The first six months produce the foundational work nobody can see, and the curve bends at month four to six, then steepens through month twelve.

Month 3 milestones

By ninety days into a competent engagement, the dashboard should show movement on every leading indicator and zero panic on the lagging ones.

• Indexation should be clean across the priority pages. New pages from the engagement are indexed in Search Console; old pages with technical issues are now resolved.
• Impressions in Search Console should be rising, segmented by query intent. The rise is mostly long-tail informational at this stage; commercial-intent impressions take longer.
• Local pack movement should be visible for businesses with a local component — first appearances on “[service] [neighborhood]” queries.
• GBP profile views and clicks should be up, attributable to the GBP rebuild.
• The first AI Overview citations should be appearing on the long-tail informational queries we engineered citation-worthy pages for.
• Qualified leads from organic may be flat or up slightly — the lagging indicator has not bloomed.

If month three shows none of the above, the engagement has a structural problem. The conversation should produce a written diagnosis. “Be patient” is not a diagnosis.

Month 6 milestones

Month six is the “is this engagement working” gate. Maile Ohye's lower bound timeline. The cheap pitches die here.

• First page-1 rankings on the priority commercial-intent keywords. Not all of them — the easiest two or three of the cluster.
• Organic sessions up 30-60% over baseline depending on starting position and vertical density.
• Qualified leads from organic attributable in GA4. The number does not have to be enormous yet; it has to be non-zero and trending up.
• CPL conversation can begin meaningfully — the unit economics start to be readable.
• AI Overview citation count rising on commercial-intent queries, not just informational.
• Branded search volume showing the first measurable uptick.

Month six is also the renegotiation gate. If the engagement is working, this is when you increase scope to compound the wins. If the engagement is not working, this is when you have the direct conversation about “which agency, not whether SEO.” Riding a non-working engagement past month six is how clients end up with twelve-month write-offs.

Month 12 milestones

A year in, the question changes from “is anything happening” to “is this paying for itself.”

• Pillar keyword candidacy on the head terms that anchored the engagement. The two-word commercial-intent queries should be on page 2 or low page 1.
• 50+ pages indexed and ranking for at least one term each. The long tail compounds; this is when the compounding shows.
• Branded search volume up year-over-year. The clearest lagging indicator of awareness work.
• ROI breakeven conversation in writing. Not always profit yet — SEO usually breaks even between month 12 and 18 depending on tier — but the unit economics should project forward with confidence.
• Year-2 plan documented. If the agency cannot show year two's plan in year one's review, they are flying blind.

The year-end report at Rule27 includes a reconciliation against the signed-day baseline: here are the five numbers from month one, here are the five numbers from month twelve, here is the delta, here is the cumulative incremental revenue attributable to the engagement, here is the cumulative retainer paid, here is the ROI ratio. That single page closes the year-one ROI conversation in plain English.

How the SERP competitors measure results (and where they miss)

The top of the “seo results” SERP is dominated by tool vendors and aggregator practitioner blogs. Each gets part of the framework right, and each has a structural limitation worth naming.

DashThis publishes a competent 10-KPI listicle with benchmarks. The framework is fine; the article is a reporting-software sales pitch. The recommendation to “use DashThis to track these KPIs” is structurally aligned with the publisher's interest, which makes it useful as a reference but not as objective methodology guidance.

AgencyAnalytics has the deepest content in the top 5 — a comprehensive guide that genuinely covers the territory. Same structural caveat: it is published by a reporting-dashboard software vendor whose business model is selling agencies the dashboard. The framework is good. The article is also a long-form ad. We recommend it as a reference document with that lens.

SEO Sherpa comes from a practitioner agency and reads with operator voice. Their KPI list is solid, with practical operator detail that the tool-vendor articles lack. The limitation is the absence of AI search KPIs — the 2026 dimension of the matrix is missing.

Cometly is an attribution-software vendor whose article focuses heavily on attribution mechanics. This is genuinely useful content if attribution is the gap in your reporting. Same vendor-alignment caveat: the article points to Cometly as the solution.

Vazoola is a link-building agency whose KPI guide leans heavily on link-related metrics. The framework is link-heavy in a way that reflects the vendor's expertise but understates the full KPI surface.

Traffic Think Tank covers 16 KPIs with community-style thoroughness. The list is comprehensive to the point of being overwhelming for a buyer trying to decide what to actually track. The five-number summary at the top of this page is the antidote.

ClickRank publishes an AI-first KPI angle that genuinely engages the 2026 dimension. The framework is fine; the article is short on operator detail.

Backlinko's “Ahrefs vs Semrush” page ranks for adjacent measurement-tool queries. Backlinko (Brian Dean) earns the rankings on research authority but does not publish client-results methodology; the focus is on tool comparison rather than reporting framework.

Ahrefs and Semrush themselves are the underlying data sources for most KPI tracking, regardless of which reporting layer sits on top. Both publish strong individual case studies; neither publishes a unified measurement framework for agency-client reporting.

First Page Sage is a competing agency that publishes case studies and methodology content. Their case-study format is useful as a reference for what real client outcomes look like in writing; the methodology depth is shallower than the agency-tool vendors.

The HOTH publishes case studies and an AEO measurement guide. The case studies are real; the measurement guide engages AI search but is structured around HOTH's own service offering.

The collective gap across all ten is the integration: a single page that ties KPI definitions, attribution mechanics, vendor-honest tool selection, query-type segmentation, and real client outcomes into one framework. That is what this page tries to be.

AI Overview citations: the 2026 KPI almost nobody is tracking

The biggest shift in SEO results measurement since 2024 is the rise of AI Overview citation as a measurable, repeatable KPI separate from blue-link rankings. SE Ranking's measurement shows AI Overviews now appear on roughly 57.9% of question queries and over 40% of commercial queries in the US. Pages that get cited inside these answers get a different kind of visibility — the prospect reads the AI answer and never clicks the blue link, but the brand is named in the answer.

Measuring the citation requires a different toolset than blue-link tracking. The three approaches in current use:

Semrush AI visibility tracker is the most accessible for agencies already in the Semrush ecosystem. It tracks which queries inside Google's AI Overviews and a sample of other AI engines cite the tracked domain. The data is sample-based and has known gaps, but it is a reasonable proxy for citation trend at the domain level.

Ahrefs Brand Radar does the same job from the Ahrefs side with overlapping but non-identical query coverage. The two tools used together produce a more complete citation picture than either alone.

Manual SERP probing is the gold standard for high-priority queries. We run a manual probe weekly for the top 20 commercial queries per client — query each in Google AI Overview, Bing+ChatGPT, Perplexity, and Gemini; capture screenshots when the client domain is cited; log the citation in a citation log spreadsheet. The labor is non-trivial; the data is the most accurate available.

The attribution gap that follows citation tracking is real and worth naming directly. When a prospect reads the AI Overview, sees your business cited, and visits your site by typing the brand into Google, GA4 logs the visit as direct or organic-branded. The AI citation gets zero attribution credit. The workaround is correlation — plot citation events against branded search volume against direct traffic, and the lift in branded-plus-direct following a citation event approximates the AI surface's contribution.

Bing+ChatGPT has begun passing referrer headers for some citation events as of mid-2026, which is the first crack in the attribution wall. Google AI Overviews and Perplexity still do not. The citation log itself is the closest thing to first-touch evidence available until the engines fix this.

We maintain the citation log as a monthly deliverable for clients on the AI search engagement layer. It is one of the deliverables in the standard report that did not exist eighteen months ago and is now central to proving the AI search work is working.

The "1 month results" trap

Anyone who tells you SEO will move the needle in 30 days is, with very high confidence, selling one of three tactics that produce ranking spikes inside the first quarter and traffic collapse inside the second or third. We have done recovery work on three Phoenix businesses who learned this the expensive way.

Private blog networks generate fast ranking lifts on commercial keywords by injecting manufactured backlink signals from a network of low-quality sites the agency owns or licenses. Google's Penguin algorithm has been penalizing PBN use since 2014. The penalty typically arrives in months 6-9. Recovery takes months of disavow work and an honest reconsideration request.

Doorway pages stuff a site with thin city-by-service pages designed to capture local intent at the keyword level without offering real differentiated content. They produce ranking lifts inside 30-60 days and trigger algorithmic demotion inside 6-9 months. The recovery work involves consolidating or removing the doorway content, which loses the temporary ranking lift but stops the demotion.

AI-spam content shipped at volume — generic LLM-generated content optimized for keyword density and shipped at 50-100 pages per month — produces a measurable ranking lift on long-tail terms inside the first quarter. Google's helpful content updates and the March 2024 core update specifically targeted this pattern. The pages get demoted at the site-level, not just the page-level, which means even good pages on the same domain lose visibility.

The three Phoenix recovery audits we have run shared the same upstream pattern: confident timelines on the sales call that contradicted Maile Ohye's published timeline, refusal to put the link or content tactics in writing, and contract structures that locked the client in past the point where the penalty surfaced. If an agency cannot explain in plain English what they are doing this month — not what category of work, but what specific actions — the work is one of the three traps above.

How Rule27 reports results

What we publish to clients every month is the five-number summary, with the operational dashboard available on demand. The summary is one page. The strategy call walks through it in 45 minutes. The numbers come from GA4, Search Console, CallRail, and the client's CRM — not from a custom reporting layer the client has to learn.

Dashboard access is direct. Looker Studio updated daily, GA4 and GSC and GBP under the client's Google account with Rule27 as Admin (never Owner), CallRail directly logged into. There is no proprietary reporting curtain. The same numbers we read are the numbers the client reads.

Quarterly we publish a citation log for AI search engagements: query, engine, citation event, screenshot, date. The log is the auditable record of the AI search work. We have never had a client question what was in the log because the log is the work, not a summary of the work.

The year-end ROI reconciliation is the document that closes year one. Five numbers from signed day, five numbers from month twelve, delta on each, cumulative incremental revenue attributable to organic via the agreed attribution model, cumulative retainer paid, ROI ratio. One page. That is the document on which renewals are signed or terminated.

What we will tell you on the first call is whether SEO is the right channel at all. If your vertical is too head-term-dominated for a 12-month window to break even, we will tell you to run paid first and SEO second. If your conversion rate is so low that no amount of traffic will close the math, we will tell you that and recommend CRO before SEO. The proof that SEO results are real and measurable is not the same as the proof that SEO is the right investment for your specific business this quarter. We will not sell the engagement that does not pass that second test.

The shortest path to seeing whether we are a fit is the free baseline audit linked at the bottom of this page. We will pull your current five numbers — sessions to high-intent pages, qualified leads, CPL, top ranking gains, top ranking losses — and deliver them in the same format we use for paying clients. 24-hour turnaround. Real PDF. No upsell. If the recommendation is “keep your current agency, here's why,” you get the document anyway.