Six things that changed how I work in 2026
Each of these is a position, not a summary. Some of them cost me money to learn.
Automated bidding and asset generation removed most of the levers marketers used to pull by hand. What is left is the input: the offer, the positioning, the creative and the definition of success. Platforms perform well when those are sharp and badly when they are vague, which means strategy stopped being the soft part of the job and became the whole job.
People now ask an AI assistant, search inside TikTok or YouTube, read a community thread, and only then type your name into Google. Ranking first for a keyword matters less than being present and citable at every one of those stops. I plan visibility across the whole path now, not the single moment of search.
AI answers quote a handful of sources rather than listing ten links. Getting quoted rewards clean structured data, clear entity signals and content that answers a question outright in the first paragraph. This is the single biggest change to technical SEO in a decade and most sites have done nothing about it.
Consent requirements and automated platforms mean nobody sees the full journey any more. The businesses coping well stopped chasing perfect attribution and moved to first-party data, trends over time, and lead quality rather than lead count. Certainty is gone; direction is still perfectly readable.
When targeting and bidding are handled by the machine, the creative is what is left to control. It decides reach, cost and conversion at once, and the old split between brand creative and performance creative has collapsed into one asset that has to do both.
Audiences now recognise machine-written filler instantly, and platforms reward the engagement that filler does not earn. Real expertise, named voices and evidence of delivery outperform volume. Publishing less and meaning it is now the cheaper strategy.
Positions I will defend
- Fix measurement before spending anything
- Optimise to be quoted, not just ranked
- Write for one reader, not for volume
- Report what failed in the same document as what worked
- Refuse tactics that need explaining to a customer
Each of these costs something. Fixing tracking delays a launch. Writing for one reader lowers output. Reporting failures loses the comfort of a good-news deck. I think all three trades are worth it, and I would rather argue about that upfront than discover we disagree in month four.
Practice, and the study behind it
I am a doctorate candidate researching ethics and measurement in digital marketing. It is not a credential I mention for decoration; it is why I hold the positions above.
How businesses use data, personalisation and automated persuasion without crossing lines their own customers would object to, and how that can be measured rather than asserted.
Regulation, platform policy and buyer scepticism are converging on the same point. Tactics that cannot survive being explained to the customer are becoming liabilities on the balance sheet, not just ethical questions.
No dark patterns, no manufactured urgency, no bought reviews, no consent tricks. Written scope before delivery. Reports that state what did not work. These are research positions applied to client work.
[PLACEHOLDER - list articles, talks and papers here as you publish them, each with a date and a link. This section is where evidence of recognition accumulates; leave it empty rather than filling it with anything unverified.]