Amit Goel

I started as an engineer because I wanted to understand how things work at the level where they cannot lie to you. Code either runs or it does not. That instinct never left. It just got more expensive to indulge.

The career started at APT Software and then Rebaca Technologies in Kolkata, writing the kind of code that nobody glamorises at conferences but that everything else quietly depends on. Instant messaging infrastructure handling millions of messages a second. Broadband provisioning toolkits. Early apps for Palm Pilot and Pocket PC, back when that was considered cutting edge rather than a pub quiz answer. It was the sort of work that builds the engineering instincts no bootcamp can teach: how to make things fast, make things reliable, and live with the consequences when they are neither.

From there, NDS — the conditional access and content security company now called Synamedia — where I spent seven years building the invisible infrastructure behind digital pay-TV platforms across four continents. DirecTV in the United States. BSkyB in the United Kingdom. Canal Plus across Europe. Tata Sky and Airtel Digital TV in India. Embedded JVMs, browser engines squeezed onto set-top box hardware with almost no memory, the plumbing that made interactive television possible before anyone called it connected TV. Unglamorous work. Essential work. And the foundation for understanding media distribution from the transmission layer upward, which turns out to be the kind of knowledge that never goes stale.

That foundation matters more than it sounds. Advertising does not exist in a vacuum. It sits on top of a content distribution system, and if you do not understand the distribution system, you will misread the advertising layer every time. That gap shows up in many of the writings on the internet, constantly.

In 2012 I stopped being an engineer inside a large company and became a founder. Patterbuzz was a digital magazine platform, effectively the Spotify for magazines in India, built at a time when the market was not yet ready for content micropayments or SMS-based billing for digital subscriptions. We raised angel funding, signed up over twenty publishers, built two genuine first-in-India innovations in content unbundling and carrier billing, and learned the specific lesson that every founder learns eventually: being early and being wrong look identical from the outside. What it left me with was the education that only comes from building something with your own money. You learn what matters very quickly when the consequences are personal.

After Patterbuzz came Amagi, and the streaming era arriving in earnest. My job was to build the OTT ad platform from zero. Server-side ad insertion — the technology that stitches advertisements into a video stream at the delivery layer, invisibly, so the viewer never sees the join. The clients were Roku, Samsung, and Xumo. A billion monthly ad requests by the time I left. Four years on the supply side of programmatic advertising, watching how publisher inventory gets packaged, how SSPs and DSPs negotiate at the technical layer, and precisely where revenue disappears between what a buyer bids and what a publisher actually collects.

Then The Trade Desk, and the perspective reversed completely. Demand side now. Leading product strategy for retail media across global markets, partnering with major Fortune 500 retailers on how first-party data becomes addressable inventory at scale. Creative management and scaling how brands build, test, and optimise ad creative across channels and formats. Strategic roadmaps across APAC presented at the most senior levels of the company. The same auction I had spent four years watching from the sell side, I was now running from the buy side. Very few people have sat in both seats. That full-cycle view is the reason I trust my own read on where the money flows, where it quietly disappears, and why the official explanations for both are usually incomplete.

Currently CPO at Moving Walls, a global digital out-of-home advertising platform building toward IPO. The challenge is a familiar one in unfamiliar clothing: how do you build a real audience-based ad marketplace when the inventory is physical screens, the measurement is probabilistic, and half your buyers still think in gross rating points. It involves relitigating most of the arguments that programmatic display and CTV already had, a decade later and with less infrastructure to inherit. I also run an AI R&D team here, building MCP servers and agent orchestrators as a new product line, which is either visionary or deeply optimistic depending on which week you ask.

Most industry writing falls into one of two categories. Trade press that reports what companies say about themselves. Or thought leadership that companies pay people to produce. Both are fine if you want news, a headline, something to forward to a colleague to look informed in a meeting. No judgment.

But if you want to actually learn — to understand how this industry works, what breaks, who benefits, and why decisions get made the way they do — neither will get you there. The only format that does is long-form, data-driven, deeply reported writing that follows the numbers wherever they lead, without a sponsor to protect or a narrative to sell. I name companies. I cite data with dates and sources. I read the footnotes in the measurement methodology documents that everyone else skips. I write because the gap between what the industry reports and what the numbers actually show is wide, and somebody should at least point at it.