Thinking, Fast and Slow in a Market With No Base Rates

A review of Thinking, Fast and Slow by Daniel Kahneman — and what changes when the outside view he prescribes has no data to stand on.

Someone asks how long the build will take. You answer in four seconds.

Six months. It sounds considered, because you have delivered projects before and the number arrives with a feeling of confidence attached. It goes into a proposal. Somebody signs it.

Then the client’s procurement cycle takes eleven weeks. A key approval sits with a person who is travelling. The integration partner’s sandbox credentials arrive a month late. Two of your engineers get poached mid-project, which you did not plan for and could not have. The thing ships in eleven months, and at some point in month eight you are in a meeting explaining a delay that was decided in those first four seconds, before anybody had done any thinking at all.

Kahneman’s book is about those four seconds. I read it as a founder, which is to say I read it as a catalogue of ways I have already lost money.

I wrote recently that Blitzscaling is a diagnostic, not a doctrine — a test you run before you spend, not a licence to spend. Thinking, Fast and Slow sits underneath that. Blitzscaling tells you whether the business qualifies. Kahneman tells you why you will get the answer wrong anyway.

What the book argues

Two systems. System 1 is fast, automatic, associative, always running. It produces impressions, intuitions and answers you did not consciously ask for. System 2 is slow, deliberate, effortful, and lazy — it mostly endorses whatever System 1 hands it, because thinking properly is expensive and we are all running low.

From that, Kahneman builds a catalogue of predictable errors. The ones that matter commercially:

• Anchoring — the first number in the room shapes every number after it, even when everyone knows it is arbitrary.
• Availability — you judge how likely something is by how easily an example comes to mind, which means recent and vivid beats frequent and true.
• Substitution — asked a hard question, you quietly answer an easier one and do not notice the swap. Will this market pay for this? becomes do I find this exciting?
• WYSIATI, what you see is all there is — System 1 builds a coherent story from the evidence in front of it and never flags the evidence that is missing. Confidence comes from the story hanging together, not from it being well-supported.
• The planning fallacy — plans built from the inside view, an optimistic path through your own intentions, ignoring the base rate of how similar projects actually went.
• Loss aversion and framing — losses land roughly twice as hard as equivalent gains, and the same choice framed two ways produces two different decisions.
• The illusion of validity — confidence is a feeling, not a measure of accuracy. Experts in noisy domains are frequently confident and reliably wrong.

The prescriptions are fewer and better than the diagnosis. Take the outside view: instead of asking how this project will go, find the reference class of similar projects and start from what actually happened to them. Run a premortem before committing — assume it is a year later and the thing has failed, and have the team write down why. Separate the quality of a decision from the quality of its outcome, because good decisions lose all the time in a probabilistic world. Where you can, replace judgement with a simple rule, because simple rules consistently beat expert intuition in noisy environments.

The honest caveat

Read it with one correction. The chapter on social priming has not held up. The studies behind it were small and underpowered, later reanalysis found almost none of them replicable, and in 2017 Kahneman publicly accepted that he had given weak evidence too much weight. He had flagged the risk to that field himself years earlier. He was a serious enough scientist to concede it in public, about his own bestseller, which is more than most authors manage.

What survived is the part you actually use: the two-system model, anchoring, loss aversion, framing, overconfidence, the planning fallacy. The foundation is fine. The edges needed trimming. Anyone quoting the priming studies at you in 2026 has read the book once and nothing since.

The four that actually cost money here

Every entrepreneur is exposed to all of these. Some are sharper in this market than Kahneman’s readers in Princeton would recognise.

1. The planning fallacy, with an external dependency stack

Kahneman’s planning fallacy is mostly about optimism regarding your own work. Here, most of the schedule risk sits with people who are not in your company at all: a regulator, a procurement committee, a bank’s integration team, a client’s finance department deciding when your invoice is due, the power supply, customs.

The inside view is wrong twice over. You underestimate your own work and you assume external dependencies behave the way they would in a system that rewards responsiveness.

The fix is Kahneman’s, applied locally. Do not forecast from the plan. Forecast from your own history: the last ten projects of this shape, what the gap was between signature and kickoff, how long approvals actually took. Once you have that recorded, quoting stops being an act of optimism and becomes arithmetic.

2. WYSIATI in a market with thin data

This one is structural, and it is the reason I chose this title.

Three good customer conversations feel like validation because the story they form is coherent. They are not validation; they are three conversations. In markets with mature research coverage you can check your story against published data. Here, for most sectors I work in, that data does not exist — no reliable churn benchmark, no survival curve for local SaaS, no published conversion rate for institutional sales in Ghana.

So the coherent story goes unchallenged, because there is nothing to challenge it with. WYSIATI is not a personal weakness in a thin-data market. It is the default condition.

3. Anchoring, in every pricing conversation you will ever have

Anchoring is the bias with the most direct effect on revenue, and the good news is that the defence is simple and completely learnable.

Do not put the first number on the table. Draw out the other party’s figure, or the budget envelope, before you price. When an anchor does land, do not negotiate against it — reset the conversation to scope and total value, so the discussion is about what is being bought rather than how far your number can be pushed down.

I have held this discipline for years and it works. Kahneman just explains why it works, and the mechanism is more unsettling than the tactic: knowing a number is arbitrary does not protect you from it. Experienced negotiators anchor too. The only reliable defence is procedural — decide the rule before you are in the room.

4. Sunk cost, in a portfolio

Run more than one venture and this becomes the expensive one. You keep funding the thing because of what it has already consumed, and you dress the decision up as commitment.

Kahneman’s framing helps: the money is gone either way. The only question is where the next cedi earns most. But the bias here is not purely cognitive — it is social. In a market this relationship-dense, shutting something down is a public act with reputational cost, and that cost is real, not imagined. What the book gives you is the ability to see the two costs separately, and to stop pretending that a decision about face is a decision about the business.

Where the prescription breaks

Kahneman’s single best piece of advice is the outside view: find the reference class, start from the base rate.

Across most of what I build, the base rate does not exist. There is no published survival curve for Ghanaian B2B software companies. No reliable benchmark for what proportion of government pilots convert to funded contracts. No dataset on collection periods by institution type. The reference class is unwritten.

Two consequences follow.

The first is that importing foreign base rates is worse than useless, because they arrive wearing the authority of data while describing a different market. US SaaS benchmarks assume card payments, contract enforcement and sales cycles measured in weeks. Applied here they will make a healthy business look broken and a doomed one look fundable.

The second is the actual prescription: you have to build your own base rates, and nobody is coming to do it for you. That is unglamorous, and it is probably the highest-leverage administrative work available to an African founder. Log every deal: source, cycle length, outcome, reason lost. Log every project: quoted duration, actual duration, where the time went. Log every hire against how they performed a year on. Three years of that and you have something no competitor can buy — a private reference class for a market with no public one.

There is a second place the book’s advice needs adjusting. Kahneman recommends broad framing: treat each risk as one of many similar bets you will make, and loss aversion loses its grip. That advice quietly assumes you get many bets. A founder with one venture and one round of family capital does not have a portfolio, and for them loss aversion is not a bias — it is an accurate reading of a one-shot game where a loss is not recoverable.

Broad framing is a privilege of structure. Running several ventures, I can take portfolio-level risks a single-venture founder would be irrational to take. That is an argument for building the structure, not for lecturing people about their psychology.

Which instincts I trust

The most useful thing in the book is not a bias at all. It is the test Kahneman worked out with Gary Klein for when intuition can be trusted. Expertise is real, but only under conditions: a sufficiently regular environment, enough repetitions to learn it, and feedback that is fast and clear enough to correct you.

Apply that honestly to your own judgement and it sorts into two piles.

That last row is the whole argument in one line. The same instinct is trustworthy or dangerous depending entirely on whether you have been keeping score.

Four practices follow from it, and they cost nothing:

• A premortem before any significant commitment. Twenty minutes, the whole team, assume it failed, write down why. It surfaces the objection the junior person was never going to raise in front of the client.
• A decision journal. What I decided, what I expected, why, what I would need to see to be wrong. It is the only defence against hindsight bias, which will otherwise quietly convince you that you knew all along.
• Judge decisions, not outcomes. A deal that closed badly may have been a good decision; one that landed may have been luck. Praising outcomes teaches your team to gamble.
• Rules over instinct where the environment is noisy. Decide the bidding threshold, the payment terms, the discount floor once, in a calm room. Then apply them under pressure without renegotiating them.

The verdict

Read it. Not in a weekend — it is long, it is repetitive in places, and the middle section on prospect theory is heavier going than the rest. Read a chapter at a time, against decisions you are actually facing.

It will not make you rational. Kahneman is explicit that studying these biases does not immunise you against them, including him. What it gives you is the vocabulary to catch the error in someone else’s reasoning, and occasionally in your own, early enough to matter. Most of that value is institutional rather than personal: teams that share the language argue better, and a premortem works even when everyone in the room is biased.

Pair it with Annie Duke’s Thinking in Bets for the operating version, and Tetlock’s Superforecasting for the evidence that calibrated judgement is trainable rather than innate. Kahneman gives you the diagnosis. Those two give you the practice.

And then do the boring part. Start the log. In a market with no base rates, the entrepreneur who has been keeping records for three years is not smarter than the competition. They are just the only one in the room who can tell the difference between a good decision and a good story.

I build software products and companies from Accra — enterprise platforms through Digiits, and ventures across compliance, marketplaces, media and agribusiness. If you are working through a decision like this one, I’m always open to a conversation.