Pankaj Raghav · Query Ownership

Base layer · Gurugram · 4 years at Google

Four years inside Google's Ads operation taught me how the auction actually thinks.

I worked on three of Google's advertiser processes, from auditing small-business accounts to leading the teams that served high-value clients. I learned the system from the side most agencies never see: what the product teams built it to reward, and where advertisers lose money without noticing.

What the auction rewards

Ad Rank weighs relevance and expected click-through alongside the bid. A tightly matched ad and landing page buys position more cheaply than budget does, so I build for relevance first.

What automation does quietly

Smart Bidding shifts spend across devices, audiences and hours based on the signals you feed it. Left unattended, it optimises toward whatever converts easiest. I decide which signals it gets.

What the dashboard never shows

Reported conversions are not the same as extra revenue. Brand clicks that would have come organically still count as paid wins. I measure what the ads add, then cut what they don't.

KSS Process

$3.5M → $10M per quarter

Pioneered a new process and built the team that took a declining business to nearly three times its quarterly revenue.

Led 18 executives and 7 campaign specialists optimising campaigns for Tier 2 and HV1 clients
Built the strategy with Google Singapore to migrate the partner-agency process to Google India

G-Tech Process

83% above global median productivity

Managed a team of 13 specialists that won Google's Top Team award.

Deep product training on AdWords, alongside Bing Ads and Facebook Ads
Consulted Tier 3 clients on their digital business strategy
Bridged sales strategy with Google's engineering team

SMB Process

5 campaign types audited

Audited Google's small-business accounts and rebuilt their keyword research and structure.

Search, Display, Video, Shopping and App campaigns
Growth analysis with Google Analytics, Keyword Planner, Tag Manager and Search Console

Subject Matter Expert and headed South Asia Pacific operations on Google's Ads teams along with experts from Google's partners Accenture and Cognizant.

The query ownership operating system

I study the business before I open the ad account.

Most campaigns fail because they start with keywords. Mine start with the moment someone searches, and everything that happens after they click. These are the steps, in the order I run them.

Step 1

Find the search that decides the market

Who is searching, at what moment, on what device. For a funeral service that is a son at 2 AM on a phone with 4% battery. For an industrial supplier it is a procurement manager comparing four vendors on a Tuesday.

Used on: Last Ride Funeral, MAM Engineering

Step 2

Re-Google every market in its own language

Translation gives you the words a marketer expects. I research what a Vietnamese construction manager or a Gulf wholesale buyer actually types, then build the account around those queries.

Arabic, Hindi, English, Bahasa and more

Step 3

Audit the chain past the click

I call the business from an unknown number. How many rings, who answers, do they close? For a service business the phone is the conversion, and most leads die there.

100+ service businesses audited

Step 4

Fix the offer when the market asks for it

Sometimes the product is the barrier. I have redesigned steel product lines for local building codes and worked on shelf-life research so the business can honestly be the best answer.

MAM Engineering, Co Co Trading India

Step 5

Measure what the ads add

I compare brand and non-brand demand, run kill-switch tests, and pull spend from anything that only intercepts customers who were coming anyway.

D2C brand: same revenue at one-third the spend

Step 6

Build one system and hand it to agents

Paid campaigns are structured as intent architecture alongside organic and AI Overview content, so both slots answer the same query. Then my own agents run and defend it around the clock.

45+ campaigns synced every 15 minutes

Case maps · one method at four zoom levels

The playbook that ranks one pincode in Delhi also wins a steel buyer in Ho Chi Minh City.

The method stays the same. What changes is the size of the map.

Zoom 1 · Pincode28.61° N, 77.21° E · one ghat at a time

Last Ride Funeral: one phone that rings in two rings

A vertical nobody wants to market, in one pincode, with zero brand recognition. Every competitor on page one used a lead form. Families searching at night don't fill in forms.

Killed the lead form

Each city page now shows a named ghat coordinator with his real photo, phone and WhatsApp: Deepak in Noida, Vinod in Gurgaon, Joginder in Delhi. Answered by a person within two rings, 24×7.

Kept everything else identical

Same site, same ads, same budget. One variable changed, so the result is clean.

2×conversion rate in six weeks

< 2 ringsanswer time, every hour of the day

Zoom 2 · CityDelhi NCR · 45+ campaigns

Owning both slots for Delhi's highest-intent searches

The city level is dozens of neighbourhood queries feeding one master brand. I ran paid and organic against the same query map, so paid demand and organic content reinforced each other instead of competing for budget.

Stopped buying broad match

Competitors in this vertical were being drained by click fraud. I moved to tightly controlled match types and intent-led ad groups.

Built a click-fraud agent

It watches every click in real time and blocks fraudulent sources before they can keep billing the account. Wasted spend fell and ROAS rose.

Organic and paid on one map

Paid campaigns build the brand searches and engagement that strong organic pages convert, so every organic page is written for the queries the ads prove out.

Organic position today

cremation services in delhi#1

funeral service provider in delhi#1

cremation ground near me#1

45+Google Ads campaigns running alongside, across Delhi NCR

18 monthsfrom zero brand to #1

Zoom 3 · NationTamil Nadu → India → export markets · 2017 to 2024

Co Co Trading India: a B2B brand built on search

As Digital Business Head for seven years, I built the team and the brand in a B2B segment where buyers rarely search for brands at all. Google Ads, SEO and media management did the work a sales force usually does.

Started from the buyer's 6 AM search

A Gulf wholesale buyer opens his phone and types something like "fresh Indian coconut supplier bulk". The question is whose name appears, and whether that supplier can ship inside his shelf calendar.

Made the answer true

Eighteen months of research into extending the shelf life of Indian agro products, so the business could serve Gulf, EU and Southeast Asian buyers it was ranking for.

7 yearsleading digital business, team and B2B brand

4 languagessearch and supply architecture in Arabic, Hindi, English and Bahasa

Zoom 4 · GlobeSydney → 10.82° N, 106.63° E and beyond

MAM Engineering: Australian steel, sold where nobody knew Australia made steel

World-class precision steel and high-rise structures, with buyers 8,000 km away in Southeast Asian and Middle Eastern procurement offices. The usual route is a business-development team and trade shows: three years and a seven-figure budget.

Changed the product

Two product lines redesigned for Vietnamese and Indonesian building codes, at a new price point that fit those markets.

Re-Googled four languages

Query research in each buyer's language, then search campaigns and technical content that answer procurement questions directly.

Zero trade shows in year one

The whole budget went into search and content.

3new export markets opened

3×average deal size versus the Australian domestic average

0trade shows needed in year one

Welix Healthcare

International expansion of a pharmaceutical brand, with content rebuilt to be the source Google's AI Overviews cite.

Decoviour

A nonprofit connecting Indian street artists with global buyers. Prabhu Shankar of Neelsandra, Bengaluru, made his first real sale after 15 years of painting.

The audit room

Most Google Ads accounts I audit are leaking money. Check yours in a minute.

These are three of the signals I check first. Enter your own figures from the Google Ads search terms, segments and conversions reports.

Estimated recoverable spend ₹2,29,000 per month · about ₹27.5 lakh a year

LeakingIrrelevant matches

LeakingBrand interception

WatchDevice bidding

How this estimate works: irrelevant-match spend is counted in full. Brand interception is estimated at 30% of spend multiplied by the brand share of conversions, because brand clicks cost less than non-brand clicks. It is a starting point for a real audit, not a replacement for one. Across client accounts, these three checks have cut more than ₹40 lakh of wasted spend every month.

Case file · mid-size D2C brand

I switched off a client's Google Ads for 30 days. Revenue went up.

The account spent ₹18 lakh a month at a reported 4.2× ROAS, and any standard dashboard would have said to keep pushing. Brand search volume was 3.4 times non-brand volume, so most "ad conversions" were customers already looking for the brand. During the 30-day pause, direct traffic held, organic traffic rose 22% and total revenue came in 6% higher. Ads returned on non-brand, high-intent queries only.

Monthly ad spendsame revenue

Before₹18L

After≈ ₹6L

ROAS

Before4.2×

After11×

During the 30-day pause

Organic+22%

Revenue+6%

Mission control · automation and programming

I use AI to run the accounts, and I write the agents myself.

I trained as a computer science engineer before I joined Google. More than half of the agents I've built work on paid media, because that is where waste hides fastest and defence has to happen in real time.

agents runningBuilt and maintained by Pankaj

click-guard● runningDetects competitor and bot clicks in real time and blocks the source before it keeps billing the account.

campaign-sync● every 15 minPulls 45+ campaigns into a queryable local database, so every question about the account is answered in seconds.

protected-registry● enforcingKeeps a registry of critical campaigns so nothing gets paused by accident or pushed over budget.

intent-rewriter● on demandRewrote 85 blog posts overnight to match a new intent map, and structures content to be cited by AI Overviews.

a2a-desk● listeningRuns Webflur's agent-to-agent server, so assistants like ChatGPT, Claude or Perplexity can commission work directly.

Sample feed showing the kind of events the agents handle. Illustrative, not live client data.

Typical agency

Once a day

Someone opens the Google Ads interface each morning and clicks through yesterday's numbers.

My accounts

Every 60 seconds

Agents watch spend, clicks and changes continuously, and act before the waste shows up in a report.

Webflur · webflur.com · Gurgaon and New York

Webflur gets businesses found, cited and hired by AI assistants.

Buyers now ask ChatGPT, Perplexity, Claude, Gemini or Google's AI Overviews before they click a link. I founded Webflur to build what those assistants need to find a business, cite it and transact with it: the positioning, the schema and a live agent endpoint. Last Ride Funeral is one of its clients.

Agent Engine

One agent-to-agent endpoint where AI assistants can discover, query and transact with the business.

Agent Card

A signed business profile for the Google, Microsoft and OpenAI agent ecosystems.

A2A API

Live answers to agent requests for quotes, availability and support, with no form or sales call in between.

GEO and AEO

Schema, llms.txt and passage-level content built to be quoted inside AI answers, and tracking of which queries select the business.

Why this belongs next to Google Ads: the ad wins the click, but the page's structure decides whether AI Overviews and assistants cite the same business above it. I build both, so the paid slot and the AI answer point to one name.

Free tool · AI Readiness Score

How visible is your website to AI search?

Paste any public URL. In about 15 seconds the tool scores the page from 0 to 100 across six pillars, then lists the top ten fixes with copy-paste code: JSON-LD schema, robots.txt lines for AI crawlers and a starter llms.txt. No signup needed.

Run a free scan on webflur.com

Example report · AI Readiness Score

Solid foundation, room to grow

72 / 100

The route here

From selling AdWords to building the agents that run it

2006 to 2012B.E. Computer Science · Visvesvaraya Technological University

Programming and digital marketing analysis, graded A.

2013Senior Executive · Getit Infomedia

Facebook, LinkedIn, email and SMS campaigns as an account manager.

2013 to 2014Assistant Manager, Digital Sales · Quikr

Led the team selling Google AdWords and Quikr listings to retailers.

From 2014 · 4 yearsSubject Matter Expert · Google Ads teams, South Asia Pacific

SMB, G-Tech and KSS processes. Top Team award and the $3.5M to $10M quarterly turnaround.

2017 to 2024Digital Business Head · Co Co Trading India

Built the team, the B2B brand and the export search architecture.

2024 to nowFounder and strategist · Last Ride Funeral, Webflur, Decoviour

Running the query ownership system across six businesses, with AI agents doing the execution.

Certifications

Google AdWords Advanced AdWords Display AdWords Shopping AdWords Fundamentals Google Ad 360 Meta Blueprint

What a founder said

“Whoever built this system is doing the best category-domination work in this country.”

A founder who found Last Ride Funeral's rankings, reverse-engineered the site and asked who ran the campaigns.

Client testimonial goes here: two or three lines from a client, with name, role and company.

Client testimonial goes here.

Whose name comes up? Every market is decided by one search. I make sure the answer is my client's name, in the paid slot and among the top in the organic one.