How to invest in stocks in 2026: a plain guide

Open a brokerage account, decide whether you are buying the average or individual companies, put in an amount you will not need for five years, and read what you buy. That is the whole method, and it has not changed in decades. What changes each year is the ground you start on, and in September 2026 the ground is unusual: a ten-year Treasury pays 5%, the highest since 2007, a money-market fund pays about 3.5%, and the S&P 500 sits 2% below the record it set in August, with its companies' profits expected to grow by more than 25% for the third quarter running. Both of those facts matter to a first purchase, and neither tells you whether to make it.

This post is the six steps in the order I would take them today, the numbers behind each one with their dates, and one figure from our own tables that I wish someone had shown me before my first trade: what happened, one by one, to the 966 companies we cover over the last twelve months.

Step one: decide which kind of investor you are

There are two honest answers and most people are the first one.

The average. You buy a fund that holds the whole market, or the largest five hundred companies in it, and you own a slice of every business in proportion to its size. You are not choosing companies; you are choosing to own the economy and let the winners and losers cancel. An S&P 500 index fund from any large provider costs about 0.03% a year, which on $10,000 is three dollars. Nothing else in investing is that cheap or that honest about what it does.

The individual company. You read a company's filed accounts, decide what it is worth, and buy it if the price is below that. This takes hours per company, it has to be redone every quarter, and the reward is that you know what you own. I do this, with my own money, and I built a tool for it. It is also the honest thing to say about it: the second answer is a job and the first is not.

The mistake is the third answer nobody admits to: buying individual companies without reading them, on a tip, a chart, or a feeling that the name is big. That is not investing in a company. It is investing in the average with extra risk and no discount.

Step two: open an account

A brokerage account is a bank account that can hold shares. In the United States the large brokers charge no commission on stock and fund trades, require no minimum to open, and sell fractional shares from one dollar, so $50 buys a fiftieth of a $2,500 share (Fidelity, Schwab and Robinhood's published terms, read 18 September 2026). Outside the US the same firms or their local equivalents exist; the tax wrapper differs by country and is worth ten minutes on your tax authority's website before you deposit anything.

Two things to check on the day you open one. First, that the broker is a member of the investor protection scheme in its country (SIPC in the US), so your shares are yours if the broker fails. Second, where the uninvested cash sits and what it pays. In September 2026 that number is not small: the Federal Reserve raised its rate to 3.75% to 4% on 16 September, and the hundred largest money-market funds paid 3.51% on average in the week of that decision (U.S. News, read 18 September). The national average savings account paid 0.38% in August (FDIC, via Raisin). Leaving cash in the second when the first is one click away is a choice, and a costly one this year.

Step three: decide how much, and how often

The rule is the one every guide gives and few people follow: money for the market is money you will not need for five years. Not the emergency fund, not the deposit on a flat, not next year's tax bill. The reason is in the table below, not in a proverb: in any given year a large share of companies, and in some years the whole market, is worth less than it was, and the person who has to sell in that year sells at that price.

How often is easier: a fixed amount on a fixed day, monthly, automated. This is not a strategy for beating anything. It is a way of removing the question "is now a good time?" from a decision that is mostly about how long you stay in, and it means you buy more when prices are low without having to be brave.

Step four: know what this year looks like

None of these numbers say what will happen. They say what you are paying and what the alternative pays, which is the part of the decision you can actually see.

Interest rates. The Federal Reserve's target rate is 3.75% to 4% after its increase on 16 September 2026; its statement said inflation "remains elevated". The ten-year Treasury yields 5.00% as of 18 September, a level last seen in 2007. That is the hurdle: a share has to be expected to do better than a government bond paying 5% a year for ten years, with none of the bond's certainty, or there is no reason to own it.

Inflation. Consumer prices rose 3.4% in the twelve months to August; excluding food and energy, 2.4%, the lowest since March 2021 (Bureau of Labor Statistics, 11 September). A money-market fund at 3.5% is roughly keeping pace with prices; through most of the 2010s cash was losing to them.

The index. The S&P 500 closed at about 7,650 on 18 September, roughly 15% above a year earlier and 2% below its August record of 7,817 (Trading Economics, read 18 September). The companies in it are expected to report profits 28.9% higher than a year ago for the quarter ending in September, which would be the third quarter in a row above 25%, and the index trades at 19.1 times the next twelve months' expected earnings, just above its ten-year average of 19.0 (FactSet Earnings Insight, 18 September).

Put those together and the picture is a market at a normal multiple of unusually fast-growing profits, priced against a bond that pays more than it has in nineteen years. The optimist reads the earnings line; the pessimist reads the bond line. I read both and conclude that this is not a year to skip step five.

Step five: read what you buy, and here is why

Here is the number I would have wanted. Take the 966 companies we cover for which our price table holds a close in the week to 17 September in both 2025 and 2026, and line up their twelve-month changes.

The index rose about 15%. The company in the middle of our list rose 0.5%. Half of the 966 fell. One in four fell by more than 20%. One in ten fell by 40% or more, and one in ten rose by 51% or more. The middle half of the list landed between a 20% loss and a 22% gain. The index's 15% is an average weighted by company size; the list above weights every company the same. Both are true, and only the second tells you what one name can do.

That is what "the market went up 15%" means when you own one name instead of the average: a coin toss on the sign, and a wide spread on the size. It is not an argument against owning companies. It is the reason the second kind of investor reads the accounts before buying, because the difference between the top tenth and the bottom tenth of that list was not luck in every case. Some of it was on the balance sheet a year ago: debt due, cash running out, a margin already shrinking, an auditor's note that nobody read. I know, because I did not read one once, and it cost me real money. That story is in what Tattooed Chef cost me.

Reading a company means three things, each with a date on it: what it earns, on the income statement it filed; what it owes, on the balance sheet; and what it is priced at against those. Not a price target, not a rating, not a summary. The filed number and the filing date. A tool that shows you the number without the date is showing you a headline, and the post before you pay for a stock research tool is a ten-minute test for telling the two apart.

Step six: keep going, and keep the receipts

The first purchase is the easy one. What decides the outcome is what you do in the year the table above is red for you. Two things make people sell at the bottom: needing the money, and not being able to stand the number. Step three is the defence against the first. The defence against the second is a written note, one sentence per holding, dated, saying why you bought it and on what number. When the number changes, reread the sentence. When the price changes and the number has not, do nothing, and that is most of investing.

What I cannot tell you is whether to buy anything this month, this year or at all; that depends on your money, your time, and your country's tax rules, and anyone online who tells you otherwise without knowing those three things is selling something. What I can offer is the reading. Pick one company you already know, the one whose product is in your house, and ask for its free report: its earnings, its debt and its price, every number with the filing it came from and the date, as a PDF by email, usually within minutes. No card, one per person. If you read it and understand what you own, you are the second kind of investor. If you would rather not, you are the first, and the index fund is waiting, three dollars a year.

Sources for the figures above: the Federal Reserve's statement of 16 September 2026; the ten-year Treasury yield and the S&P 500 level, record and one-year change from Trading Economics, read 18 September 2026; the Consumer Price Index for August 2026 from the Bureau of Labor Statistics, released 11 September; the S&P 500 earnings growth estimate and forward P/E from FactSet's Earnings Insight of 18 September 2026; the index fund cost from the published expense ratio of the largest S&P 500 ETFs, read 18 September; the twelve-month changes of 966 companies from our own price table, last recorded close in the week to 17 September of each year, no company excluded.

Educational and proprietary. This explains what our research does, not the exact formula behind it, and it is not personalised investment advice. See the full disclaimer.