The Guidance Line That Mattered More Than the Beat

The beat was worthless before it printed. That is the conclusion. I spent one bad evening earning it. The rest of this is the working.

Four o’clock. Revenue above consensus. Earnings per share above consensus. Both by enough that every headline in the next thirty seconds said clean. The stock jumped in the thin after hours market and sat there looking pleased with itself, and I bought some of it inside a minute because I had read two numbers and formed a view.

Forty minutes later, halfway through the call, it was trading below the regular session close.

The quarter you are reading about is over

I had understood this as a sentence. I had never felt it anywhere lower down. The three months being reported have already happened to the business. The price in front of you has spent those three months adjusting to the market’s guess about them, and by the time the release lands, the guess and the answer are close enough that the difference is small change.

So a beat settles a question the market had mostly stopped asking.

What is genuinely open is the next few quarters, and the company answers that in a different part of the same document, in smaller type, as a range.

On that night the range was revenue for the coming quarter of $1.19bn to $1.23bn. Analysts had been carrying $1.24bn. The midpoint is (1.19 + 1.23) / 2 = 1.21, and (1.24 - 1.21) / 1.24 = 2.4% below. A shade under two and a half percent. Nobody put it in a headline. Next to it sat one sentence about a slower ramp in a product line, which the chief financial officer repeated on the call in plainer language and then declined to quantify when asked twice.

Two large numbers about the past and one small range about the future, and the small one moved the price by more than both of the others combined.

The order I read in now

My old order was headline earnings per share, headline revenue, the year over year comparison, then the stock reaction, then the rest of the release if I still cared. That order works if the question is whether the quarter was good. It is useless if the question is what the stock is worth tomorrow morning.

Guidance first. The midpoint of the guided range against what the market was carrying. As a percentage. Before I look at anything else.

Then the reason given for the guidance, in the company’s own wording, because a timing problem and a demand problem arrive wearing exactly the same tone of voice and the difference between them is the entire trade.

Then cash flow, because a quarter can beat on earnings while cash conversion quietly deteriorates, and the gap between the two is where the interesting arguments live.

Headline earnings per share comes near the end of my pass. It is a fact I need. It is rarely the fact that decides anything.

The other thing I got wrong that night

I had an opinion inside sixty seconds, in a session where the spread was wide and there were a few hundred shares a side, on a document I had read a fifth of. The direction turned out to be wrong. The execution was worse than the direction.

So the rule is simple. The first move after a release is an opinion formed by whoever reads fastest, and the call has not started. I read the release properly while that move happens. If there is still a trade afterwards, there is still a trade. The spread will be a fraction of the width in the morning.

Underneath this sits a habit I would defend anywhere. Anything that reprices a stock overnight has a public document behind it, and reading the document is slower than reading the headline about it and is the only way to find out which sentence did the work.

The method for getting through a release quickly is in how to read an earnings report in 20 minutes, and the mechanics around the date, including what happens to options premium and why the drift afterwards is real, sit in trading earnings season. The checklist that puts one quarter back into the context of a whole business is how to analyze a stock.

I still get the direction wrong on earnings nights, regularly. What I no longer do is get it wrong in the first sixty seconds on the basis of the one number the company knew you would read first.