Most SEO in Pakistan is quackery — and AI will expose it

The frame: SEO is the traffic police of the internet

The episode opens with Muzamil setting the table for an audience that has, in his words, always assumed SEO is “a technical thing for technical people.” His framing is deliberately simple. Search engine optimization, he says, is the traffic police of the internet — the system that decides who arrives where. Every small business, every lawyer, every doctor with a website and a digital persona depends on it, whether they know it or not.

He introduces Salman Baig with care. Salman has thirteen years in the industry, previously headed SEO at ProPakistani — “one of the biggest platforms in Pakistan” — and is now running SEO for the Alibaba Group across South Asia: five markets, Pakistan, Bangladesh, Sri Lanka, Nepal and Myanmar, anchored on Daraz. The conversation, Muzamil tells the audience, will be useful not just for SEO professionals but for anyone who wants to understand how the future of content, marketing and the internet itself will work.

How a Peshawar kid taught himself to game Google

Salman’s origin story is more honest than most. He grew up in Peshawar, went to F.G. School, joined the Institute of Management Sciences for university, and dropped out partway through because he was already earning more than his parents.

The first money came from something he is willing to call by its real name on camera. “It’s a very, very black-hat thing. I’m just sharing because I was a kid,” he says. He and a small online community would visit each other’s blogs and click each other’s ads to generate revenue. Opera Turbo was the trick of the moment — a click through it paid a dollar. He pulled in his first hundred dollars that way, paid through Liberty Reserve. Then one morning his AdSense account was shut down. “It was a proper shock,” he says. “I was in school, but I felt like I was unemployed.”

That shock pushed him into proper SEO. By 2013 he had reverse-engineered what he called the Timroid — timing on steroids — a ranking factor he believed Google weighted heavily. He noticed that any new product, even an iPhone that hadn’t been released yet, would already have someone ranking for it. The websites that posted first became the authority; everyone who came later followed. He tested it with the keyword “web hosting 2014,” ranked it in a single day, and was making two hundred dollars a day from one page. He wrote a short book about the technique. One reader, Salman recalls, built an Amazon affiliate blog using the method and flipped it for what he believes was around five million dollars — he is careful, on camera, to note he is unsure of the exact figure.

The lesson he draws from that period is harsher than the success story suggests. “From there I realised two things,” he tells Muzamil. “There is a game. And most people are dumb.”

What SEO actually is, in the ordinary person’s context

Muzamil asks Salman to bring the conversation back down for the audience: what is SEO, really, for the lawyer or the small-business owner with a website?

Salman’s answer is patient and structured. “If you want to sell something on the internet, you either run an ad or you go to where the people are. SEO is basically your relationship with Google. Organically — it doesn’t cost you money.” A few tweaks to the title, a sensible configuration of the content, an understanding of how search engines read a page, and a small operator can rank.

He is explicit about the asymmetry he sees. The lawyer, he argues, can beat ProPakistani on any given day for a lawyer query. The reason is what he calls topic distillation. “If I ask you what ProPakistani is about, you’ll say it covers this, that, the other. There is no one context. This guy is one context. He knows what he is selling and he is limited to what he is selling.” Search engines, in his reading, are biased toward the operator whose page satisfies a defined need cleanly. ProPakistani is a platform; the lawyer is an answer.

He extends the same logic to Daraz. “Daraz is a marketplace. Everyone is selling everything. If I come and I say I am the expert of this pen, and my website is only about this pen, I will beat Daraz any given day. That’s the edge an ordinary person has.” For most Pakistani niches, he adds, there are barely ten or twenty serious players. Reaching the first page is not the hard problem people think it is.

Technical SEO, and the small tweak that got him on Backlinko

Salman explains his own taxonomy of the field. There is simple SEO — content, configuration, the basics. There is technical SEO — becoming a crawler in your own head, looking at the site the way Google’s machinery does. And there is semantic SEO, which he believes is the most important and the most neglected.

The ProPakistani years sharpened the technical side. A small website can be hand-tuned. A large one — billions of pages, in his telling — forces you to work in systems and processes. A one percent change applied across millions of pages produces a disproportionate result. That work eventually got him featured on Backlinko, which he describes as the most authoritative SEO blog he knows. He laughs at how minor the actual tweak was. “If I tell anyone, they’ll say it’s nothing. But that’s the beauty of technical SEO. Small tweaks, huge impact.”

He gives Muzamil a concrete example of why semantic understanding matters. In late October and early November, Google quietly modified one of its content-evaluation criteria. The effect was sudden: Reddit and Quora traffic spiked tenfold while authentic, single-author sites — doctors writing about health, for instance — collapsed. Salman’s inbox flooded. He went live on Facebook and said, on tape, that he believed Google would roll the change back. Within minutes, Google tweeted that they would not. He held his position. His reasoning was structural. “Google will never take the risk of giving fake information in an era where ChatGPT and AI are their direct competitors. They will not let ChatGPT give a better answer.” The next day, Google rolled the change back.

Cost of retrieval, and why you should help Google save money

This is where the conversation steps into the part of SEO almost nobody on Pakistani YouTube discusses. Salman introduces cost of retrieval — a concept he credits to a Turkish friend in the industry named Koray — as a frame for understanding why some pages rank and others don’t.

The idea is plain once it is said out loud. Google has to pay, in compute, to figure out what every page means. If your page lays the information out in a structure that requires less analysis, Google’s cost of serving you is lower. Two pages with equivalent information will not be treated equivalently if one is cheaper to read. “Google will go for the lesser cost,” Salman says, “because they have to look at the cost.”

He builds outward from this to a broader argument about why most SEO consultants in the region are quacks. “An SEO guy normally just tells you: change the title, put this keyword here, that’s it. If you ask him why, he has no idea. There is no fundamental layer. It’s just upper-upper quackery.” Muzamil agrees, and the two spend a stretch of the conversation circling the same point from different angles: the people doing SEO in Pakistan are not, for the most part, students of search engines. They are students of last year’s tutorial.

ChatGPT is not the end of Google — it’s a layer on top of it

Muzamil reframes the conversation toward what he believes the audience is really worried about. Google’s business is eyeballs and ads. What happens when the eyeballs move to a chat box?

Salman’s reply is unusually calm, and he is unembarrassed about why. “When AI came, everyone was like, Google is done. My question is: do you really think we’re the ones figuring out what Google should do? Google figured this out ten years ago.” He points to the patents Google has been quietly registering for years — generative integrations, ranking refinements — and says they were never caught off guard.

To test the argument, Muzamil opens the ChatGPT app on camera. He asks for the top lawyers in Islamabad. ChatGPT 3.5 returns a stale list, including a lawyer Muzamil notes “doesn’t exist anymore.” He rephrases — please share the website of a good lawyer. The model browses the web. Muzamil narrates what he is seeing in real time. “It’s going to Google. It’s searching the top-ranked sites. It’s pulling from their websites.” He turns to Salman. “Effectively, my understanding is — you’re not disrupted. There is a layer on top of you. You just have to now understand the layer on top of you.” Salman agrees: “Exactly. A hundred percent.”

The point Salman wants to land is one he has clearly been waiting to make. “The kind of SEO I do is not for the search engine. It’s for the users and for natural language processing. The whole concept of AI is NLP. So when the medium changes from Google to AI, you are still at the top of it — because you understand the baseline.”

The agency, the audits, and the move to Dubai

Muzamil steps back to the biography. After ProPakistani, Salman ran an agency offering what he calls SEO audits. The pitch was unconventional. Rather than executing the work, he would build a six-month plan for the client — a framework — and let the client’s own development, content and SEO teams implement it. The implementation, he tells Muzamil, is the easy part. The plan is the leverage. “Every guy came back and thanked me and paid me extra for the next website. Every guy brought referrals.”

The agency was making good money, but Salman walked away from it for the reason he keeps returning to throughout the episode: he could not see what was left to learn. “I had hit the saturation point. I was doing the same thing over and over.” He kept a manager in place to run what was left and moved to Dubai. Eventually he joined Alibaba — specifically the Daraz operation across the five South Asian markets — as head of SEO. Working on sites with what he says are around thirty million pages, he tells Muzamil, has been “out of the world.” The granularity of e-commerce SEO at that scale is the next ceiling he wants to push against.

The roadmap for someone starting out — and why “brain drain” is a mindset problem

The most generous stretch of the conversation comes when Muzamil asks Salman to lay out a roadmap for a Pakistani kid who wants to do what he did. Salman does not waste the opportunity.

His starting point is uncomfortable. “Blame the mentors and the people in Pakistan who don’t teach them in a way that lets them hit the peak. People come to learn SEO to make money. They do a great job — they make money. But the mindset to study the subject? Nobody has that here.”

Then he gives the plan, in order. First, learn basic SEO — what search is, how it works, what a user wants, what a keyword is. Then learn natural language processing, because every search engine and every AI is built on NLP. Understand how a sentence is broken down — entities, adjectives, sentiment analysis, stemming — so you know what Google is actually seeing when it reads your page. Then learn machine learning, specifically the cost-of-retrieval angle. Then learn data analysis — not a PhD, he says, a good Excel sheet is enough — because nothing in SEO should be guesswork.

He frames the payoff bluntly. “If Google hired me tomorrow to build a new search engine, I could help them build it. Because I am not an SEO guy anymore. I am the search engine guy.” On top of all that, he reads Google’s patents. “Google is never open about their algorithm. They tell you the surface — we are integrating AI in your search — but the patents tell you what is actually coming.”

Muzamil pushes him on the brain-drain question. Salman’s framing is precise. “My concern is — either everyone leaves, everyone hits the peak. Or nobody does. Why is it only selected people who leave?” His answer is mindset, not resources. “I come from a middle-class family. I funded my own education from school. The information is not limited. The resources are not limited. The only difference is mindset.”

Where websites and content are headed in the next ten years

The closing stretch is the most disagreement the two have all episode, and it is productive. Muzamil’s thesis is that websites as we know them are heading for a graveyard. Content farms, faceless aggregators, photocopy-machine SEO operations — the whole layer that profited from the last decade of search is going to be cleaned out. AI, in his reading, will accelerate the wipeout, because AI-generated bulk content has no brand equity, no retention, and exists only to harvest a single click before the user leaves and forgets.

Salman agrees with the wipeout, but not with the death of websites. “Websites are going to be there. They are the baseline of any database. But the utility of going to a website will change.” His own writing process has already moved, he says, to a semantic mode — every word checked against how an NLP classifier will read it, so the meaning he intends is the meaning the machine extracts. “I am not just writing for Google anymore. Maybe tomorrow OpenAI builds a search engine. Bing. Microsoft. I am the SEO for them as well. Because I am learning the baseline, not what Google is doing.”

Muzamil pushes the brand-equity point with a small piece of editorial archaeology. He remembers when ProPakistani led the country on tech and telecom coverage — live broadcasting the 3G auction when nobody else was, owning thought leadership on Jio’s launch. Today, he says, he could not tell you what ProPakistani is for. “Is it a fashion blog? A food blog? A mobile-phone blog?” The race for clicks worked during the pump, but a pumped audience does not retain.

Salman closes the argument with a story he tells with some self-awareness about how dark it is. A friend of his wrote a Python script that scrapes a Google topic, generates a video, and uploads it to YouTube — two thousand videos a day, from one person. “If this is not controlled,” Salman says, “there is no utility to the internet anymore.”

Pakistan in 2050, and the gift of being introduced to digital

Muzamil ends the way he ends most of these conversations: extrapolate. Where is Pakistan in 2050?

Salman gives an answer he warns most people will not like. “Pakistan in 2050 is going to be where India was ten years ago.” India, he says, was the call-centre economy a decade back and is now the innovation leader of the region. Pakistan is just entering the call-centre and back-services phase now. A few companies might climb higher; the majority, he believes, will sit at the bottom of the value stack.

Muzamil offers a sharper reading of what made even that possible. For two decades, he argues, Pakistan ran on quackery — property flipping, import trading, currency control, a society where the loudest quack was the richest man. The shift to IT and freelancing, he says, was made possible because the dollar got squeezed and the property game stopped paying. Even the Ponzi-adjacent crypto manias of recent years, in his framing, had one redeeming consequence: they technologically introduced an entire generation to digital systems for the first time.

Salman agrees with the framing more than he expected to. “The introduction had to happen. That introduction has happened. The rest — what works, what doesn’t — they’ll figure it out on their own.”

By the end of the conversation, both men are converging on the same point from opposite directions. The decade ahead will be a hard correction for everyone who learned the surface — the SEO quacks, the content farms, the platforms that traded brand equity for clicks. The people who survive it are the people who learned the subject. As Salman puts it, with the cleanest line of the episode: “Earlier I used to tell people — please, learn it as a subject. Now I don’t even have to say it. It’s my way or the highway.”