Outlook

The ranking factors didn't change. AI changed which of them are still a moat.

A read on where iGaming affiliate media is heading, based on how search actually rewards sites and what AI has done to the economics of winning.

What actually ranks

The eight categories that decide it, from years of building against them.

After years of optimising sites for search traffic, my working model of what drives rankings comes down to eight categories:

  • Technical foundations
  • Intent fit
  • Content quality
  • Authority and trust
  • User satisfaction
  • Freshness
  • Context adjustments
  • Spam and trust dampeners

The first three, technical foundations, intent fit and content quality, are table stakes. Almost everyone serious gets them broadly right, and AI has made that easier, not harder. They're necessary and they win you nothing on their own.

It's decided in authority and trust and in user satisfaction. Those signals are slow to accumulate, hard to fake, and heavily weighted: topical authority built across a whole cluster rather than a single page, brand mentions, regulatory and compliance trust, and on the other side repeat visits, resistance to people bouncing back to the results, and engagement that says a visitor actually got what they came for. Much of this is measured at site and cluster level, not page level. You are building the reputation of a whole entity, which rewards patient investment and punishes portfolio sprawl.

The category most affiliates underrate is the spam and trust dampeners: footprint similarity across a portfolio, scaled-content classifiers, host-reputation baselines. As AI floods the space with templated pages, those dampeners only get more aggressive, and anything that looks like every other affiliate operation gets caught by them.

AI moved the moat

AI changed which of these you can defend.

What has changed is which of these you can defend. Content scale used to be a moat: covering every query, in every market, faster than the next affiliate. That is gone. A small team with the right tooling now produces that volume at pace, so raw output stops being a differentiator almost overnight.

The categories that hold up are the ones AI can't fake at scale: genuine authority and trust, real user satisfaction from a product people come back to and actually use, and a brand people look for by name.

The mechanism is simple. An LLM can write the review. It can't run the operation behind it: the support that actually resolves a stuck withdrawal, the years of the same reviewers being right often enough that players and Google both learn to trust them, the community that keeps coming back, the name a player types in without a search. Those are accumulated, not generated, and the gap between the operators who have them and the ones who don't widens as producing everything else gets cheaper.

Who is still here in five years

The properties that actually survive the shift.

Not the content machines. Not portfolios of thin, heavily optimised landing pages. Those were built for a moat that no longer exists.

The ones that last will have done four things: built trust and brand into real assets rather than SEO inputs; earned direct, returning traffic that doesn't route through a search engine; become a source that LLMs cite and recommend; and built actual products that add value to a user rather than just monetising them.

How I'd run a portfolio for it

Two time horizons, run at once.

Two horizons at once. We are still optimising for traffic algorithms across several sources, and there will always be fast, smart, sustainable ways to win with them in the near term. That game is still valid and worth playing hard.

What matters is what you do with the return. It funds the long-term bets: heavy investment in the moats AI can't spin up quickly, the ones that will still matter in five years. So the right portfolio does both deliberately. It reads the current algorithms closely and maximises ROI from them now, and it reinvests that into what outlasts them. Pick the two or three of those eight categories where being genuinely best-in-class compounds, resource them properly, and let the near-term wins pay for it.

The moat behind the moat

Underneath all of it, the team.

All of it is downstream of the team. AI gives everyone the same scale and the same access; the people making the calls are what still separates operators. The best culture attracts the best people, and that compounding is the moat behind the moat.

These organisations get leaner and more senior every year: fewer people, weighted heavily toward strong individual contributors rather than layers of management. The investment I'd make now is for that shift, ahead of needing it. Bring in the most senior individual contributors in the space, and the best from outside it, while there's a window to do it deliberately, and build the pipeline underneath them. The teams that win the next phase are small, senior, and very good.