The power of the autopilot: understanding System 1 decision-making

Most buying does not involve a decision in any meaningful sense

Someone picking up the same washing-up liquid they always buy has not weighed the options. They have repeated an action that worked before, in a shop laid out the way it was last time.

That observation is worth keeping. Nearly everything else marketing has built on top of it needs checking, because System 1 is the most borrowed idea in the industry and one of the most overstated.

My position: the useful conclusion from this research is defensive rather than offensive. It tells you to stop making your buyer think when they do not need to, and it does not give you a way to make people buy without noticing. Most marketing writing on this topic promises the second thing.

Where the idea comes from, and what Kahneman claimed

Daniel Kahneman set out System 1 and System 2 for a general audience in 'Thinking, Fast and Slow' in 2011. System 1 runs fast, automatically and without a sense of effort. System 2 is the slower, deliberate, effortful mode you notice yourself using. Kahneman credits the terms to the psychologists Keith Stanovich and Richard West.

The part that marketing dropped is Kahneman's own warning about the terms. He is explicit in the book that the 2 systems are, in his words, 'fictitious characters'. They are expository devices, not parts of the brain, and he tells the reader to treat the names as nicknames for characters in an argument.

That matters because the marketing version usually reifies them: System 1 becomes a place in the head with a door you can knock on. A critical review of dual-process theories documents how much weaker the general 2-system account is than its popular version. Fast impressions and slow reasoning interact continuously rather than taking turns, which is the position I take in the science of consumer behaviour.

The part of this canon that did not survive

A large amount of System 1 marketing advice rests on social priming research: the studies where a subtle cue changes behaviour without the person noticing. Words about old age slowing people's walking. Holding a warm drink making someone judge a stranger as warmer.

Much of that literature failed to replicate. Kahneman was among the first to say so publicly. In an open letter to priming researchers in September 2012, he called for a replication effort and warned that he could see a train wreck looming. The chapter of his own book that presented priming findings has not aged well, and he has since acknowledged as much.

I am not saying that fast, automatic processing is a myth. It is not. I am saying that the specific claim marketers took from this literature, that small cues can reliably steer behaviour without awareness, is the part that broke. If your marketing plan depends on subliminal influence, it depends on the weakest available evidence.

What survives, and what to do with it

This is the table I would keep. The middle column is the honest status of each claim rather than the version in a conference talk.

The claim	How well it holds	What to do about it
Familiar things are easier to process and tend to be preferred	Well supported, and one of the most replicated findings in the area	Keep your brand and message stable for longer than feels interesting to you
Being recalled at the buying moment beats being persuasive at it	Well supported by advertising effectiveness data	Build distinctive assets and attach them to real buying moments
Easy-to-read content is judged more favourably than hard-to-read content	Reasonably supported, effects modest	Fix the writing before the psychology, as in plain English
Defaults and ordering change what people choose	Supported, and effect sizes depend heavily on context	Set defaults that serve the buyer, and expect a smaller effect than the case studies promise
Emotional advertising outperforms rational advertising long term	Supported in the IPA databank, with the caveat that it is award entries	Treat it as a budget allocation argument, not a licence for vagueness
Subtle primes steer behaviour without awareness	Largely failed replication	Do not build anything on it
System 1 can be targeted directly and separately from System 2	Not a real claim, the systems are expository devices	Stop using it as a strategy

The 3 rows at the top are where a small business should spend its effort. They are also the least exciting, which is why they get less coverage than priming.

What this leaves you with

Fluency is the one usable idea here, and it does not require a media budget.

If your buyer has to work out what you sell, they are in effortful mode, and effortful mode is where doubts get raised. Making your offer easy to understand is not a soft accessibility concern. It is the closest thing to a System 1 tactic that survives scrutiny, which I set out at more length in plain English as a conversion tactic.

Consistency is the second. Changing your visual identity, your explanation and your positioning every 18 months resets the familiarity you were building. Most small businesses do this because they are bored of their own material long before their market has registered it. Attaching a stable identity to specific buying moments is the practical version, covered in how to make your brand top of mind.

The third is reducing the number of things a buyer has to work out at once. That is a design and structure question rather than a psychological trick, and it belongs with the friction work in designing decision interfaces.

The ethical line, and where it is now a legal one

Writing about automatic processing tends to slide into how to exploit it. The distinction I use is whether the technique still works when the buyer can see it.

A clearly labelled default that saves someone effort survives being explained. A countdown timer that resets on reload does not. In the UK that is no longer only an ethical question: the unfair commercial practices provisions of the Digital Markets, Competition and Consumers Act 2024 apply to practices from 6 April 2025, and I set out what that covers in the role of emotions in purchase decisions.

There is also a practical reason to stay on the right side of it. Techniques that work by preventing someone from thinking tend to produce customers who later think about it, which is where refunds, chargebacks and bad reviews come from.

Take one page and count the things a buyer has to work out before they can act. Reducing that count is the whole of the usable version of this research.