AI Search Is Splitting Your Traffic: How to Measure ChatGPT and Perplexity Referrals in GA4
4x
Growth in AI referral traffic across client sites this year
15 min
To build the GA4 view that isolates it
2x+
Conversion rate of AI referrals versus average, consistently
0
Standard reports that show this today
There is a new traffic source growing in your analytics right now, and GA4’s default reports are hiding it inside the referral bucket. Visitors arriving from ChatGPT, Perplexity, Gemini, and Copilot click through when an assistant cites you, and across my client base that stream has roughly quadrupled this year. Small absolute numbers still, but the trend line and the visitor quality demand their own measurement.
Because here is the finding that matters: AI referrals convert at roughly double the site average, consistently, across clients. It makes sense. A visitor arriving from an assistant’s recommendation was pre-sold by the most trusted interface they use. You want to know exactly how many arrive and what they do.
Build the View in GA4, Step by Step
1
Know the referrer signatures
AI referrals arrive with identifiable referrers: chatgpt.com and chat.openai.com, perplexity.ai, gemini.google.com, copilot.microsoft.com, claude.ai, and you.com. Some assistant clicks arrive stripped as direct, so whatever you measure is a floor, not a ceiling. Floors still show trends.
2
Create the channel group
In Admin, under Data display, edit your channel groups: add an AI Referrals channel matching source against a regex of those domains. From that point on, standard reports break the stream out cleanly beside Organic Search and the rest.
3
Build the exploration for history
Channel groups apply going forward. For history, build an Exploration filtered on session source matching the same regex: sessions, engaged sessions, conversions, and landing pages, trended monthly. This is your baseline chart, and it will be a hockey stick at small scale.
4
Study the landing pages above all
Which pages do assistants send people to? That list is your citation footprint made visible: it tells you which content AI systems consider reference-worthy, and it is the template for what to produce more of. This single report has reshaped content plans for my clients more than any keyword tool this year.
Reading the Numbers Honestly
Metric
How to read it
Volume
Small today almost everywhere. Single digit percentages. The trend is the story, not the total
Conversion rate
Consistently above site average. Pre-sold visitors behave like referrals from a trusted friend
Landing pages
Comparisons, definitive guides, and original data get cited. Thin service pages do not
Month over month
The only chart worth presenting. Screenshot it quarterly; the compounding makes the argument for you
“Nobody reported on organic search traffic in 1999 either. The brands that measured it early built the playbooks everyone else bought later. AI referrals are that moment again, with a fifteen minute setup cost.”
Ram Kr Shukla, SEO and Growth Consultant
Once the measurement exists, it changes decisions: the landing page report feeds your content plan, the conversion data justifies AI visibility investment to whoever owns the budget, and the baseline lets you actually evaluate whether citation work, the kind I documented in my AI recommendations research, moves your numbers. Measurement first, then optimisation has something to answer to.
Want the full AI visibility measurement stack installed?
I set up AI referral tracking, citation monitoring across assistants, and the quarterly baseline test as part of every AI SEO engagement. Know your numbers before your competitors know theirs.
Google Search Console: 7 Reports Most Marketers Never Open (And What They Reveal)
7
Reports beyond Performance and its averages
Free
Every one of them, sitting in your property now
5 of 6
Ranking emergencies I diagnosed started in these reports
15 min
Monthly routine to check all seven
Most marketers use Search Console as a clicks chart: open Performance, look at the line, close the tab. Meanwhile the reports that actually diagnose problems, the ones I open first on every audit and every emergency call, sit unvisited one menu below. Five of the six ranking emergencies I documented recently were solved in these reports, not in the Performance chart.
The Seven, and What Each One Confesses
1
Page indexing, the exclusions side
Everyone checks how many pages are indexed. The diagnosis lives in why pages are excluded: crawled currently not indexed rising means a quality or demand problem, discovered not crawled means crawl budget or internal linking, and a sudden spike in noindex is a deploy accident announcing itself.
2
Crawl stats, hidden under Settings
Crawl requests trending down, response times trending up, or a spike in 404 fetches: this report shows how Google experiences your server. A slow, error-prone crawl experience quietly reduces how much of your site Google bothers with.
3
URL Inspection’s rendered HTML
Not a report, a tool, and the definitive answer to what Google actually sees on a page: the rendered HTML, the resources it could not load, the canonical it chose versus the one you declared. My client-side rendering post exists because of what this tool reveals.
4
Performance, filtered to regex and compared periods
The default view averages everything into mush. Regex filters isolate query families, brand versus non-brand, questions versus commercial. Period comparison shows exactly which queries lost clicks after a change. The report everyone opens, used the way almost nobody uses it.
5
Enhancements and rich result reports
Where schema breakage announces itself: valid items dropping after a plugin update is the silent click-through killer I keep finding weeks after the damage started. Calendar this one after every site update.
6
Links, internal and external
Which of your pages Google considers most linked, internally and externally. The internal list regularly contradicts what teams believe their architecture emphasises, and that contradiction is the internal linking to-do list.
7
Removals and manual actions
Empty for most sites forever, which is why nobody looks, which is why a stray removal request or a manual action sits undiscovered while everyone debugs the algorithm instead. Thirty seconds, monthly.
“The Performance chart tells you something changed. The reports underneath tell you what, where, and usually who deployed it.”
Ram Kr Shukla, SEO and Growth Consultant
The 15 minute monthly routine:
Page indexing: exclusions trend, any new exclusion reason appearing
Crawl stats: request trend and average response time
Enhancements: valid item counts versus last month
Performance with brand regex excluded: non-brand clicks trend
Links: top internally linked pages still match your priorities
Removals and manual actions: still empty
One URL inspection on your most important money page
A Worked Example: The Twenty Minute Diagnosis
A founder called about a traffic slide their agency had spent three weeks attributing to a core update. The Performance chart did look like a slope. But the monthly routine from this post found the truth in twenty minutes: Page Indexing showed excluded pages climbing week over week, the exclusion reason was alternate page with proper canonical tag, and the timing matched a platform migration. The new platform was generating parameter URLs that the canonical setup handled badly, and Google was steadily choosing the wrong versions.
No update, no penalty, no content problem: a canonical configuration error, visible in a free report nobody had opened, fixed in an afternoon by the same developer who caused it. That is the recurring lesson of these seven reports. They rarely tell you something is wrong before the traffic chart does, but they tell you what is wrong, which is the difference between three weeks of theory and one afternoon of fix.
How to make the routine stick in a team:
Calendar it monthly with a named owner, not a shared intention
Log five numbers each run: exclusions, crawl requests, response time, valid rich results, non-brand clicks
Screenshot the exclusion reasons breakdown, because trends matter more than totals
Run it within 48 hours after every deploy, migration, or plugin batch update
Escalate on trend breaks, not absolute numbers, since every site’s baseline differs
Questions I Get on This Topic
GSC data seems to disagree with GA4. Which is right? Both, about different things. GSC measures search impressions and clicks at Google’s edge; GA4 measures sessions after consent, blockers, and JavaScript. Directional agreement is what you want; numeric identity is impossible. Diagnose search problems in GSC, behaviour in GA4.
How long is GSC data retained? Sixteen months, which is why the monthly log matters: it becomes your only view beyond that horizon, and year over year comparisons are where slow structural decay becomes visible.
Is the API worth setting up? Once you are logging monthly by hand, yes: the API removes the sampling and row limits of the interface and feeds dashboards. But the habit comes first. Automation of a report nobody reads is decoration.
These seven reports are the monitoring layer of my technical SEO service, and the starting instrument of every SEO audit I run.
Want an expert eye on your Search Console data?
Every engagement I run starts inside these seven reports. If your traffic moved and nobody can say why, this is where the answer is sitting.
Programmatic SEO Done Right: How to Scale Pages Without Getting Penalised
2M
Pages I have managed at marketplace scale
1
Question that decides survival: is each page useful alone?
10x
Faster with AI, which cuts both ways
3
Structural rules that separate assets from spam
Programmatic SEO, generating pages from data and templates instead of writing them one by one, built some of the biggest organic footprints on the web. I learned it managing marketplace sites with millions of URLs, long before AI made generation trivial. And that is exactly the new problem: AI made scaling easy, so the graveyard of penalised programmatic sites is filling faster than ever.
The difference between a programmatic asset and thin content spam was never the page count. It is whether each generated page would deserve to exist if you built it by hand. That standard is passable, and here is how the survivors pass it.
The Three Structural Rules
1
Unique data per page, not unique words
Spinning synonyms across a template fools nobody since Google’s helpful content systems matured. What works: each page assembled around data that genuinely differs, prices, specs, availability, local details, real comparisons. If two pages would answer a visitor identically, they should be one page.
2
A query with real intent behind every URL
Programmatic pages must map to searches people actually make: city plus service, product plus comparison, tool plus integration. Generating pages for query patterns nobody searches produces index bloat that drags the whole domain. Validate the pattern’s demand before generating the ten thousandth page.
3
Crawl architecture that carries the weight
Ten thousand pages need hub structure, clean pagination, and internal links that give every page a path from the money pages. Generated pages dumped into a flat sitemap with no internal linking are orphans at scale, and Google treats them accordingly.
Where AI Fits, and Where It Breaks Things
AI role
Verdict
Assembling pages from your structured data
Ideal. This is what the technique always was, with better tooling
Writing unique intros against real data per page
Works with editorial gates and the brief discipline from my AI content workflow
Generating both the data and the words
The graveyard. Invented data plus template prose is the exact pattern quality systems now catch
Deciding which pages deserve to exist
Never. Demand validation and pruning stay human decisions
“Programmatic SEO fails at the same question every time: would this page deserve to exist if a human had to build it? AI changed the cost of generating pages. It did not change the standard they are held to.”
Ram Kr Shukla, SEO and Growth Consultant
Start smaller than you want to: one pattern, a few hundred pages, indexed and measured for a quarter. Scale what earns impressions and conversions; prune what does not, ruthlessly. The discipline to delete generated pages is rarer than the ability to generate them, and it is what keeps the domain healthy.
Considering programmatic at scale?
I have run this at marketplace scale and rebuilt it after others’ penalties. Pattern validation, data architecture, and the guardrails, before you generate page one.
Cart Abandonment: Why 70 Percent Leave and the Fixes Ranked by Impact
70%+
Typical abandonment rate across e-commerce
48%
Cite unexpected costs as the reason, survey after survey
Ranked
Fixes below ordered by measured impact
3
Recovery emails, the only after-the-fact fix that scales
Seven in ten shoppers who add a product to cart leave without buying. Most brands treat that as weather: unfortunate, universal, nothing to be done except send a recovery email. But abandonment is not one problem, it is a stack of specific frictions, and they are not equal. Fixing them in impact order is the difference between a quarter of tinkering and a measurable revenue lift.
This ranking comes from the checkout audits inside my e-commerce engagements: session recordings, funnel step data, and exit surveys, across brands from wellness to fashion.
The Fixes, Ranked by Measured Impact
1
Kill the cost surprise
Half of abandonment traces to costs appearing late: shipping, taxes, fees revealed at the final step. Show total cost as early as possible, display shipping thresholds on product pages, and if margins allow, build shipping into prices. The single highest-impact fix on this list, and usually the least technical.
2
Let guests buy
Forced account creation is a commitment demand at the moment of least patience. Guest checkout with an optional account offer after purchase recovers most of what registration walls lose. The account can be one click on the thank you page.
3
Shorten the form to its skeleton
Every field is a toll. Address autocomplete, one name field, no phone number unless delivery genuinely requires it, and card scanning on mobile. Count your fields, then justify each one to a hostile audience.
4
Answer the trust question at the money moment
Security signals, return policy, and delivery time visible at the payment step, not buried in the footer. Hesitation at the card field is a trust question the page failed to answer.
5
Make mobile the primary checkout, not the port
Most carts are built on mobile. Test the whole flow on a mid-range phone over mobile data: keyboard types matching fields, buttons above the fold, no layout jumps as elements load. What annoys you mildly on WiFi kills conversions on a train.
6
Then, and only then, the recovery sequence
The three act email flow from my D2C email automations post: helpful within the hour, objection-answering next day, incentive on day three only if margin allows. Recovery is real revenue, but it is the mop, not the fix for the leak.
How to Find Your Own Ranking
Evidence source
What it tells you
Funnel steps in GA4
Which exact step loses the most: cart to checkout, checkout to payment, payment to done
Session recordings at the loss step
What people do right before leaving: hesitate, rage-tap, hunt for costs
Exit survey, one question
Ask leavers what stopped them. Crude, and reliably more honest than your assumptions
Mobile versus desktop split
A wide gap means the problem is the phone experience, not the offer
Sequencing matters because the fixes compound: a shorter form on a faster mobile page with visible costs multiplies, not adds. One client stack of the first three fixes lifted completed checkouts by a third before a single recovery email was improved.
A Worked Example: The Wellness Brand Checkout
A concrete sequence from a recent engagement. A D2C wellness brand came in with 78 percent abandonment and a recovery email as their only countermeasure. The funnel data showed the biggest loss between checkout start and payment: 61 percent of starters never reached the card field. Session recordings showed the same choreography again and again: shopper reaches the shipping step, sees the delivery charge for the first time, pauses, opens a new tab, and never returns. They were comparison shopping the shipping fee, not the product.
The fix stack: shipping cost surfaced on the product page next to the price, a free shipping threshold banner sitewide, and the account creation prompt moved to the thank you page. Checkout completion rose 31 percent in six weeks. Only then did we touch the recovery emails, which promptly performed better too, because the flow now re-entered a checkout that did not repeat the original offence.
The mistakes that keep abandonment high even after fixes:
Testing checkout changes on desktop when 80 percent of carts are mobile
Adding trust badges nobody recognises instead of the return policy people actually read
Discounting in recovery email one, training customers to abandon deliberately
Measuring cart-to-purchase as one number instead of step by step, which hides where the leak actually is
Celebrating recovered carts while the leak that created them stays unfixed
Questions I Get on This Topic
Should I use exit-intent popups on the cart page? Tested honestly, they recover very little at checkout stage and irritate plenty. The shopper leaving over a cost surprise is not stopped by a popup; the one leaving to compare prices comes back on merit or not at all. Fix the reasons, then let the email flow do the polite follow up.
Is COD availability part of abandonment in India? Significantly, yes. For Indian D2C, payment method availability, COD, UPI, cards, at the moment of payment is a trust and convenience factor on par with shipping cost. If your COD rules are conditional, state them early, not at the final step.
What is a good abandonment rate to aim for? Under 65 percent is strong for D2C. But the absolute number matters less than your trend and your step-by-step loss profile. A 70 percent rate where the loss is spread evenly is healthier than 68 percent with a cliff at the payment step.
Want your checkout audited step by step?
Cart and checkout CRO is core to my e-commerce work: funnel analysis, session recordings, and the fix list ranked by impact for your specific leak, not the generic one.
Schema Markup in 2026: The Only 6 Types Most Businesses Actually Need
6
Types that earn their maintenance cost
30+
Types most plugins offer that you can ignore
2
Jobs schema now does: rich results and AI comprehension
15 min
To validate everything you currently have
Schema markup suffers from a completeness problem: there are hundreds of types, plugins offer dozens, and teams either implement nothing or implement everything badly. Broken schema is worse than none; I covered a client whose rich results vanished sitewide from one invalid plugin update in my rankings emergencies post.
Schema also quietly picked up a second job. It used to earn rich results in Google. Now it also tells AI systems, the assistants and the Overview builders, exactly who you are and what you offer in a format they parse without guessing. Both jobs are done by the same six types for most businesses.
The Six That Matter
1
Organisation, on every page
Name, logo, URL, and sameAs links to your real profiles. This is your entity anchor: the machine-readable statement of who is behind the website. For consultants and personal brands, Person schema does this job alongside it.
2
Service or Product, on money pages
What you sell, described in structured form. For e-commerce: Product with price, availability, and ratings, which earns the rich result that moves click-through even without a rank change. For services: Service with provider and area.
3
FAQ, where real questions live
On pages answering genuine questions buyers ask. Google trimmed its visual FAQ real estate, but the structured question and answer pairs are exactly the fragments AI systems lift. Write real questions, not keyword-stuffed fakes.
4
Article, on content that argues expertise
With author, dates, and publisher. This connects your content to your entity, which is how expertise accumulates to a name instead of evaporating page by page.
5
BreadcrumbList, sitewide
Cheap to implement, improves the SERP display, and reinforces your site architecture to crawlers. The lowest-effort item on this list.
6
Review or AggregateRating, where earned
Ratings you genuinely collect, marked up where they live. Never fabricated and never sitewide decoration: unearned rating markup is the fastest route to a manual action of anything on this page.
What to Skip, and the Maintenance Rule
Skip the exotic types your plugin offers unless you demonstrably need them: Event without events, VideoObject without video, HowTo on pages that are not how-tos. Every type you add is a validation surface that can break silently in a plugin update. The maintenance rule that survives contact with reality: validate quarterly in Search Console’s Enhancements reports, and after every plugin or theme update. Fifteen minutes, calendared, owned by a named person.
“Schema used to be decoration for rich results. It is becoming your machine-readable identity. Six types, kept valid, beat thirty types nobody checks.”
Ram Kr Shukla, SEO and Growth Consultant
A Worked Example: The Silent Breakage and the Recovery
The client I mentioned in my rankings emergencies post is worth expanding here, because the sequence is so typical. A plugin update changed how review markup rendered: technically present, semantically invalid. Google stopped showing stars sitewide within days. Rankings barely moved, but click-through fell by nearly a third, and revenue followed. It took weeks to notice precisely because everyone was watching rankings, the metric that had not changed.
The recovery was mechanical once diagnosed: valid markup restored, Search Console validation run, rich results back within two crawl cycles. The durable fix was procedural, the quarterly validation calendar and a named owner. Schema does not fail loudly. It fails like a fridge light: you only notice when you finally open the door and check.
Marking up ratings that do not visibly exist on the page, a manual action magnet
Duplicate Organisation schema from theme, plugin, and manual code all firing at once
FAQ markup on invented questions no one asks, which wastes the one fragment AI systems lift
Schema in the page builder that disappears when the template changes
Validating in a testing tool once at launch and never in Search Console where field reality lives
Questions I Get on This Topic
JSON-LD or microdata? JSON-LD, without hesitation. It lives in one script block, survives template edits far better, and is what Google recommends. If a plugin outputs microdata woven through your HTML, that is a fragility worth migrating away from.
Does schema directly improve rankings? Not as a ranking factor in the direct sense. It improves how results display, which moves click-through, and it improves machine comprehension of who you are, which increasingly matters beyond Google. The revenue path is real; it just does not run through the position number.
How do I add schema on WordPress without another plugin? Most SEO plugins you already run handle the six core types adequately. The gap is usually configuration and validation, not tooling. Before adding anything, audit what your current stack already outputs; duplication is the more common disease than absence.
Want your structured data audited and rebuilt properly?
Schema architecture is part of every technical SEO engagement: what you have, what is broken, what is missing, and the six type implementation your stack actually needs.
Why Your Rankings Dropped and It Was Not the Algorithm: 6 Real Causes From Client Emergencies
6
Real emergencies, six different causes
1
Was actually caused by a Google update
48h
Typical time to find the real cause
100%
Had already blamed the algorithm
The call always sounds the same. Rankings crashed, traffic is down 40 percent, and someone has already diagnosed it from a headline: it must be the latest Google update. In the last two years I have taken six of these emergency calls. Exactly one turned out to be the algorithm. The other five were self-inflicted, invisible to the team, and fixable within days once found.
I am documenting all six here, anonymised, because the diagnostic path is more valuable than any individual fix. When rankings fall off a cliff, the date of the fall is your best friend: algorithms roll out over weeks and create slopes. Deployments happen on a Tuesday and create cliffs.
“Google updates create slopes. Deployments create cliffs. Look at the shape of your traffic drop before you blame the algorithm.”
Ram Kr Shukla, SEO and Growth Consultant
The Six Causes, In Order of How Often I See Them
1
The deploy that shipped a noindex
A staging configuration reached production during a replatform. Every category page carried a noindex tag for eleven days before anyone noticed, because the pages looked completely normal in a browser. Traffic fell 60 percent. Found in twenty minutes by crawling the site and sorting by indexability. Recovered in three weeks after the fix. The lesson: every deploy checklist needs an SEO line item.
2
The redesign that deleted internal links
A beautiful new design replaced a text-heavy footer and descriptive sidebar navigation with a minimal menu. Nobody realised those unfashionable links were carrying authority to 200 deep pages. Rankings for mid-tail terms slid over six weeks, which looked exactly like an algorithm slope. The redesign date gave it away. Restoring a structured footer nav recovered most of it.
3
Content cannibalization from enthusiasm
A funded startup published 14 posts in one quarter around the same keyword theme, each slightly different, none clearly primary. Google rotated between them, and the original page that ranked position 4 fell to page two. Consolidating 14 posts into 3 with redirects brought the primary page back inside a month. More content is not more SEO.
4
The robots.txt line nobody owned
An agency handover left a disallow rule blocking a parameter path that, after a platform migration, became the canonical path for the entire blog. The blog quietly stopped being recrawled. New posts were invisible, old posts froze. One deleted line, recovery within two crawl cycles. Audit your robots.txt quarterly. It takes four minutes.
5
Schema that expired and took rich results with it
Review snippets vanished sitewide, and click-through rate fell 30 percent even though positions barely moved. The culprit: a plugin update changed the schema output and made it invalid. Traffic loss without ranking loss is a presentation problem. Check the Enhancements reports in Search Console before anything else.
6
The one that really was the algorithm
A content site built on aggregated, lightly rewritten material dropped in a core update, and there was no technical rescue available. The honest answer was a two-quarter investment in original data and genuine expertise. That one hurt to deliver, but pretending a technical fix exists when the problem is content quality just burns budget. Recovery came, slowly, through the harder path.
The Diagnostic Sequence I Run Every Time
In this order, before touching anything:
Plot the drop precisely in GSC. Cliff or slope? Note the exact start date.
Ask what shipped within five days of that date: deploys, plugin updates, redesigns, migrations, agency changes.
Crawl the site and sort by indexability: noindex, canonicals, robots directives, status codes.
Compare rankings versus click-through rate. Positions stable but traffic down points to SERP presentation, not ranking.
Check GSC Enhancements and Page Indexing reports for sudden spikes in exclusions.
Only after all five: check whether a confirmed update overlaps your dates, and whether affected pages share a quality pattern.
The order matters because the self-inflicted causes are both more common and faster to fix. Starting with the algorithm theory means starting with the one cause you cannot act on this week.
This diagnostic sequence is exactly what my technical SEO service runs in emergency engagements, usually finding the real cause within 48 hours.
Rankings dropped and nobody can explain why?
I run emergency rendering-to-robots diagnostics for exactly this situation. Usually the real cause surfaces within 48 hours, along with the fix plan.
The Keyword Research Method That Finds Buyers, Not Just Volume
6%
Of pages produced 79 percent of conversions in my audits
4
Sources of buyer language most researchers never open
0
Keyword tools required for the first half
1 metric
That matters: revenue per ranking
Standard keyword research is a volume-sorting exercise: export ideas, sort by searches per month, pick the big ones. It reliably produces content that attracts readers who will never buy anything. I have audited the results dozens of times: a blog ranking for high-volume informational terms, a business wondering why none of it becomes pipeline.
Buyer-first research inverts the process. Instead of starting with what has volume, you start with what buying sounds like, then check volume last, almost as an afterthought. Volume decides how big the opportunity is. Intent decides whether it is an opportunity at all.
Start Where Buying Language Actually Lives
1
Your sales calls and inbound emails
The phrases prospects use in their first message are search queries with names attached. Collect twenty of them and you have a keyword seed list no tool would ever surface, in the exact vocabulary of people who spend money.
2
Your GSC queries, filtered ruthlessly
Search Console already shows queries where you get impressions. Filter for commercial markers: best, vs, alternative, pricing, for, near, service. These are buyers finding you by accident. Make it deliberate.
3
Competitor comparison and pricing pages
Whatever your competitors write comparison pages about is where their buyers hesitate. Their page titles are a map of purchase-stage queries they have already validated with their own money.
4
Review sites and community threads
The exact sentences in reviews and community complaints become long-tail commercial queries within months. Reading them is keyword research from the future.
Then Classify Before You Count
Intent tier
Marker patterns
What to build
Ready to choose
alternative, vs, pricing, best X for Y
Comparison and alternatives pages. Build these first, always.
Defining the solution
how to fix, tool for, service for
Solution pages and deep guides with hard CTAs
Feeling the problem
why is X happening, what is X
Cluster content with email capture, not trial CTAs
Never buying
free, template, DIY, salary
Deliberate choice only: list building or nothing at all
“Sort by volume and you optimise for readers. Sort by intent and you optimise for revenue. The spreadsheet looks worse. The pipeline looks better.”
Ram Kr Shukla, SEO and Growth Consultant
Only after classification do the tools come out, to size the tiers and check difficulty. You will ship a plan with smaller numbers on it than the volume-first version, and it will make more money. The 20,000 visit blog producing 10 trials that I wrote about earlier was a volume-first plan executed perfectly. That is the trap.
A Worked Example: The B2B Software Client
A workflow software client came to me with a keyword plan built the traditional way: 40 informational topics sorted by volume, none below 1,000 monthly searches. We ran the buyer-first process instead. Twenty sales call transcripts produced phrases like approval workflow for finance teams and alternative to spreadsheets for purchase orders, terms with search volumes the tools listed as 30 to 70 a month, some as zero.
Six months later, the pages built on those tiny terms produced more demo requests than the entire existing blog. The zero volume page ranked first for a phrase the tools claimed nobody searched, and it converted at 9 percent, because the seventy people a month who did search it were all buyers. Volume tools measure the past at low resolution. Sales language measures buyers in the present at full resolution.
The traps in buyer-first research:
Taking sales phrases too literally: extract the pattern, not the exact quote
Classifying by what the keyword contains instead of what the SERP shows, since Google’s results reveal the intent it has already decided on
Ignoring zero volume terms that your own GSC proves get impressions
Building every commercial page at once instead of validating three first
Letting the informational tier grow back to 80 percent of the calendar because it is easier to brief
Questions I Get on This Topic
How do I get sales language if I have no sales team? Your inbox, your contact form submissions, and your competitors’ review pages. Ten inbound enquiries contain more buyer vocabulary than most keyword exports. For newer businesses, communities and forums where your buyers complain are the same source, one step earlier.
Do volume tools have any role left? Yes: sizing and prioritising within a tier after intent classification, and difficulty checks before committing to a big build. The failure mode is letting them lead, not using them at all.
How many keywords should the final map contain? Fewer than feels comfortable. A typical engagement ships 30 to 60 mapped terms across the four tiers, each assigned to a page with a job. A 500 row spreadsheet is a research artefact, not a plan.
Want a buyer-first keyword map for your business?
Every SEO engagement I run starts with this method: sales language mining, GSC excavation, and an intent-classified map your content team can execute against for a year.
llms.txt Explained: Should Your Website Have One Yet?
2024
When the proposal first appeared
1 file
Markdown at your domain root
Low
Cost to implement, under an hour
Early
Adoption stage: signal, not standard
Every few years the web grows a new plain text file at the domain root. robots.txt told crawlers where not to go. sitemap.xml told them where everything was. The newest proposal, llms.txt, wants to tell AI systems what your site is about and where your most important content lives, in clean markdown they can digest without fighting your page templates.
The honest status in mid 2026: it is a proposal with growing but uneven adoption, and no confirmed guarantee that the major AI providers consistently consume it. So should you bother? My answer for most clients is yes, and the reasoning has less to do with the file than with what producing it forces you to do.
What llms.txt Actually Is
A markdown file at your domain root containing a short description of your site and a curated, annotated list of your most important pages, optionally with companion markdown versions of key content. Think of it as a manually curated sitemap written for language models: not everything, just what matters most, described clearly.
The Case For and Against, Honestly
For doing it now
For waiting
Costs under an hour and cannot hurt anything
No guarantee major AI crawlers consistently read it yet
The curation exercise itself reveals how legible your site is to machines
Standards this young can change shape and need redoing
Early signals in an emerging convention tend to age well
Time might be better spent on schema and content structure first
AI crawlers that do respect it get your best content, framed your way
A neglected, outdated llms.txt is worse than none
“The file takes an hour. The thinking it forces, deciding which twenty pages define your business and describing each in one clear sentence, is worth the hour even if no crawler ever reads it.”
Ram Kr Shukla, SEO and Growth Consultant
If You Do It: The Right Way in Five Steps
Implementation checklist:
Write one paragraph describing what your business is and who it serves, in plain language with your standard entity description
Curate 15 to 25 URLs maximum: services, key comparisons, best guides, about. Curation is the point; dumping the sitemap defeats it
Annotate every link with one sentence on what the page covers
Keep it synchronised with reality: assign an owner and a quarterly review
Pair it with the fundamentals that AI systems verifiably do use today: clean HTML, schema, and server-rendered content
Priority check before you start: if your content is client-side rendered or your schema is broken, fix those first. llms.txt is a refinement on top of machine-readable foundations, not a substitute for them.
Want your site AI-legible from the foundations up?
AI readiness is part of my AI SEO engagements: rendering, schema, entity consistency, and yes, a properly curated llms.txt at the end of it.
Why Most Landing Pages Convert Under 2 Percent, and the Anatomy of Ones That Do Better
2%
Where most B2B and service pages sit
8%+
What disciplined pages reach with warm traffic
5 sec
How long you have before the back button
1
Job per page. More is fewer conversions
The average landing page converts a low single digit percentage of its visitors, and most teams accept that as a law of nature. It is not. It is the compound interest of a dozen small frictions: a headline about the company instead of the visitor, three competing calls to action, proof buried below the fold, and a form that asks for information nobody wants to give.
I have rebuilt enough landing pages across SaaS, e-commerce, and services to know the difference between 2 percent and 8 percent is rarely one dramatic change. It is anatomy: the same organs, arranged in the order the visitor’s brain actually wants them.
The Anatomy, Top to Bottom
1
A headline that finishes the visitor’s sentence
They arrived with a problem in mind. The headline names the outcome they want, in their words, not your category jargon. The five second test: would a stranger know what you do and why it matters to them?
2
One promise, one call to action
Every additional CTA divides attention and clicks. One primary action, repeated down the page, with everything else demoted to quiet text links. Deciding what the page is not for is the design decision.
3
Proof before claims get expensive
A specific number, a named client, a short quote, placed immediately after the promise. Proof works when it arrives before scepticism, not after.
4
The objection paragraph nobody writes
Every offer has a silent objection: too expensive, too complicated, not for companies like mine. Name it and answer it on the page. Sales teams handle objections; pages that convert do the same.
5
A form that matches the commitment
Ask for what the next step genuinely requires and nothing else. Every field beyond name and email needs to justify its existence in conversion terms. Progressive profiling exists for the rest.
6
A close that restates the exchange
End with what they get, what happens next, and how little they risk. Confusion at the point of action is the quietest conversion killer there is.
The Diagnostic Order When a Page Underperforms
Symptom
First suspect
High bounce, low scroll
Headline and above-fold mismatch with the traffic source’s promise
Good scroll, no clicks
Weak or competing CTAs, proof arriving too late
Clicks but abandoned forms
Form length, or trust collapsing at the commitment moment
Converts on desktop, dies on mobile
Speed, layout collapse, or a form that fights the keyboard
Traffic source matters more than most CRO advice admits: an 8 percent page fed warm branded traffic and a 2 percent page fed cold prospecting clicks can be the same page. Diagnose against the source, not the average.
A Worked Example: The Demo Page Rebuild
A B2B SaaS client’s demo request page converted at 1.9 percent from paid traffic, and the team was convinced the traffic was the problem. The recordings said otherwise: visitors scrolled the full page, hovered the form, and left at the same field, company size, positioned second in a seven field form. Combined with a headline about the product’s AI engine rather than the visitor’s approval bottleneck, the page was doing everything in the wrong order.
The rebuild followed the anatomy: headline rewritten to the outcome, cut approval time from days to hours, one named client result placed directly beneath it, the silent objection, this looks like an enterprise tool, answered with a mid-page paragraph about setup time, and the form cut to three fields with the rest moved to the booking step. Same traffic, same offer: 5.8 percent within a month. Three times the pipeline from a page that took four days to rebuild.
The landing page mistakes that survive most redesigns:
Rewriting the design while keeping the old headline, the highest leverage element on the page
Social proof from logos nobody recognises instead of one specific number from one named client
CTAs that describe your process, request a consultation, instead of their outcome
Mobile treated as a smaller desktop instead of the primary experience
Testing button colours while a seven field form sits untouched
Questions I Get on This Topic
How much traffic do I need before A/B testing? Meaningful tests need hundreds of conversions per variant, which most pages never have. Below that, sequential testing against a stable baseline plus session recordings beats a statistically doomed split test. Confidence comes from converging evidence, not just p values.
Long page or short page? Match length to commitment. A newsletter signup earns a short page; a six figure engagement earns every section of a long one. The question is never length, it is whether every block answers something the visitor is actually weighing.
Should the page match the ad exactly? The promise must match word for word; the page then expands it. Most quality score and bounce problems trace to ads writing cheques the headline never mentions again.
Have traffic that refuses to convert?
I run landing page teardowns as part of every CRO engagement: heatmaps, session recordings, and the exact anatomy fixes ranked by expected lift.
Product Page SEO for E-Commerce: The 12 Elements That Move Rankings and Sales
12
Elements, each doing ranking and conversion work
2x
Typical traffic gap between optimised and template product pages
30 of 30
Audited stores had product page gaps
1 page
Template fix scales across the whole catalogue
Product pages are where e-commerce SEO stops being theoretical. A category page brings the shopper in; the product page has to rank for the long-tail buying query and close the sale, both at once. Yet in all 30 store audits I ran last year, product pages were the least optimised template on the site, because nobody wants to touch a thousand pages one by one.
The good news: you do not touch them one by one. You fix the template and the content model, and the improvement scales across the catalogue. Here are the twelve elements, in the order I fix them.
The Twelve Elements
1
A title formula with buying modifiers
Product name plus the attributes buyers actually search: brand, material, size, use case. Set the formula once in the template; let real search query data from GSC refine it per top seller.
2
Unique opening copy above the fold
Two or three sentences, written for the buyer’s decision, not pasted from the manufacturer. Duplicate manufacturer descriptions are the single most common product page failure, and the most fixable.
3
Product schema, complete and valid
Price, availability, ratings, and brand in structured data. This is what earns the rich result, and rich results move click-through even at the same position.
4
Real reviews, marked up and visible
Reviews are conversion fuel, unique content, and long-tail keyword coverage all at once. Surface them, paginate them properly, and mark them up.
5
The specification block as crawlable text
Specs in real HTML, not images or tabs that never render. Long-tail queries live in specifications.
6
Buyer questions answered on the page
The three to five questions support hears repeatedly, answered under the product. FAQ schema optional; answering them at all is the win.
Elements seven to twelve, the supporting structure:
Breadcrumbs with schema, matching your category architecture
Internal links to the parent category and two or three genuinely related products
Image file names and alt text that describe the product, not IMG_4021
Canonical discipline across variants: one product, one indexable URL
In-stock and out-of-stock handling that preserves the URL and offers alternatives
Loading speed under 2.5 seconds on mobile, because every element above loses to a page nobody waits for
“Category pages decide whether you get the shopper. Product pages decide whether you get the order and the long tail. Fix the template and you fix a thousand pages in one decision.”
Ram Kr Shukla, SEO and Growth Consultant
Sequencing advice from the audits: schema, titles, and the duplicate copy problem produce the fastest measurable movement. Start there, measure for six weeks in GSC filtered to product URLs, then work down the list.
A Worked Example: The Template Fix That Moved a Catalogue
The fashion e-commerce brand from my 120K visitors case study is the clearest demonstration of template leverage. Early in that engagement, product pages were manufacturer descriptions under a photo: identical copy to forty other stockists, no schema, specs locked inside image files. Individually fixing 800 SKUs was never going to happen.
The template pass took three weeks: a title formula with material and occasion modifiers, a two sentence unique opening generated from structured attributes and then human-edited for the top 100 sellers, Product schema wired to live price and stock, and the spec table converted to crawlable HTML. Within two months, product page impressions doubled, and long-tail queries the tools had never suggested, fabric plus occasion plus size phrasing, started appearing in GSC. The template did in three weeks what page-by-page editing would not have finished in a year.
The product page mistakes I still see on audits:
Deleting out-of-stock pages and losing their rankings instead of holding the URL with alternatives
Variant colour pages all indexable, splitting authority six ways for one product
Reviews behind a JavaScript tab that never renders server side, invisible to the crawl
Schema price out of sync with displayed price, which risks rich result loss
Zero internal links from content into products, leaving the catalogue to rank on its own
Questions I Get on This Topic
Should every product page have unique long copy? No. Two or three genuinely unique decision-focused sentences beat four paragraphs of padded uniqueness. Depth belongs on the category page and buying guides; the product page needs enough to differentiate and convert.
What about products with almost identical variants? One canonical product page per real product, variants as options on it. The exception is variants people specifically search for, colour or capacity with real query volume, which can earn their own indexable page and unique copy.
How do I prioritise 800 SKUs? You do not. Fix the template for all, then hand-polish the twenty pages where revenue concentrates. Product revenue follows a power law; your optimisation effort should follow the same curve.
Template-level fixes like these are how my e-commerce SEO service scales improvements across whole catalogues.
Want your product template audited element by element?
I run this twelve point assessment across your catalogue, benchmark against your top competitor’s template, and hand your developer a prioritised template fix list.
