The red flags in a research note cluster in four places. Earnings quality asks whether profits turn into cash. Accounting tells ask whether revenue is being pulled forward. Valuation tricks ask whether the model has been flattered. Analyst bias asks whether any real downside has been modelled at all. The professional habit that catches all four is the same one: do your own work, and trust no number you have not tested.

This closing module pulls the series together into a defensive reading checklist. It covers the signs that a note is weaker than it looks, and the habits that separate disciplined analysis from taking someone’s word for it.

Read defensively

A research note is an argument, and every argument arrives with something behind it. Sometimes that is an incentive, sometimes a relationship with the company, and often something far more ordinary: an analyst who has covered a business for six years and grown fond of the story they have been telling about it.

Reading defensively does not mean assuming bad faith. Most notes are written honestly by people who want to be right, and treating every one as a deception will make you miss good work. It means something narrower and more useful. Treat the conclusion as a claim to be tested rather than a fact to be accepted, and know where weak notes tend to give themselves away.

One structural point is worth knowing before you start. Across the industry as a whole, sell side ratings skew towards the positive, and BUY and HOLD together far outnumber SELL. There are unglamorous reasons for this. An analyst who publishes a SELL may find management less willing to take their calls, and the firm has commercial relationships to consider. None of that makes any individual note dishonest. It does mean that a SELL rating carries more information than a BUY, simply because it was harder to publish, and that a HOLD is sometimes doing the work a SELL would do elsewhere.

Earnings quality: when profit and cash diverge

The single most important tell in analysis is the relationship between reported earnings and cash flow, which you met in module 2 as FCF conversion. This is where it earns its keep.

High quality earnings turn into cash. Low quality earnings do not. So a growing divergence between rising earnings and flat or falling free cash flow is among the strongest warning signs there is. It says that profit exists on the income statement and not in the bank.

The reason this signal is so powerful is that it is hard to sustain. Almost every technique that flatters reported profit works by recognising something early or deferring something else, and both of those eventually reverse. Cash does not reverse. So a gap between the two lines can be opened for a year or two, but it has to close, and it usually closes suddenly and in the wrong direction.

What you are watching for is the trend rather than the level. One weak year is often just a company building inventory or paying for a factory. Three consecutive years of profit climbing while cash flattens is a pattern, and patterns are what you are reading for.

Accounting red flags

A small number of accounting signals reveal revenue being pulled forward, or quality quietly eroding underneath a healthy looking headline.

• Days sales outstanding rising faster than revenue. DSO = (receivables ÷ revenue) × 365, and it tells you roughly how many days customers take to pay. If sales are climbing while customers take longer and longer to settle, two explanations compete: revenue is being booked aggressively, or demand is softening and the company is offering easier terms to keep volumes up. Neither is good, and both show here before they show in the revenue line.
• Revenue growing much faster than bookings or backlog. Backlog is tomorrow’s revenue. If today’s reported sales are running well ahead of the forward order book, growth is being borrowed from a future that is emptying out as you watch. This is one of the few places where you can see a slowdown before it arrives.
• Acquisition driven growth without integration. Buying revenue is easy and it counts in the growth rate exactly like revenue that was earned. A string of acquisitions with no evidence the businesses are being absorbed can mask organic growth that has stopped. Look for whether the company discloses organic growth separately. Reluctance to split the two is itself the answer.

Notice the shape these share. Each one compares a headline number against something slower moving underneath it, and the warning is in the gap rather than in either figure alone. That is the same habit module 2 asked for when it told you to read the dashboard rather than a gauge.

Valuation red flags

The valuation section has its own tells, and you have met all of them in modules 3 and 4. What follows is the checklist version.

• A terminal growth rate above long run GDP. Unsustainable by definition, since the company would eventually exceed the whole economy, and it sits on the largest single component of a DCF.
• A discount rate set suspiciously low. Understates the risk and compounds through every forecast year and through the terminal value.
• An aspirational peer group. Comps padded with richer companies so the median multiple rises. Read the names before the numbers.
• A bear case above the current price. If even the downside scenario implies a gain, no genuine downside has been modelled and you are looking at three versions of winning.

These four share a property that makes them worth memorising. Each is a single input, buried where a reader is least likely to look, that moves the final answer a long way. Nobody has to write a dishonest sentence to use any of them. That is exactly why the checklist has to be applied deliberately rather than trusted to catch your eye.

Governance and incentive red flags

Some warning signs sit around the company rather than inside its numbers, and they colour how much weight the numbers deserve.

An aggressive acquisition pace with no evidence of integration is one, and it is worth watching for a second reason beyond masking organic growth: serial acquirers accumulate goodwill on the balance sheet, and goodwill written down later is an admission that the price paid was too high.

Management incentives tied to metrics that can be gamed are another. If the bonus depends on adjusted EBITDA, expect to see a good deal of adjusting. If it depends on earnings per share, buybacks become more attractive than investment, because buying back shares lifts EPS by shrinking the denominator without the business improving at all.

A persistent unwillingness to discuss risks is the third. Listen to how management handles the difficult question on an earnings call. Answering it plainly is a good sign. Answering a different, easier question is a signal in itself.

The question underneath all three is the one to keep asking: who benefits if this story turns out to be true, and what happens to them if it does not?

The professional habit: do your own work

The thread running through the whole series is that a research note is a starting point rather than a verdict. It is somebody else’s argument, generously shared, and it deserves the same treatment you would give any argument you cared about being right on.

The disciplined habit is to rebuild the load bearing parts yourself. Sketch your own rough DCF to see whether the assumptions survive contact with your own. Assemble your own peer group and see whether it looks like theirs. Write your own bull, base and bear cases before reading theirs. And always, on every company, cross check reported earnings against cash flow.

You do not need the analyst’s model. You will not have their data, their access or their hours, and trying to replicate the whole thing is how people give up on the habit entirely. What you need is enough of your own work to know whether their conclusion survives independent scrutiny. A DCF on the back of an envelope with your own growth and discount rate will not give you a price target worth acting on, but it will tell you very quickly whether theirs required heroic assumptions to reach.

That is the whole discipline, and it is why this module is last. Modules 1 to 5 taught you what the parts of a note are and how each one is built. This one asks you to build them yourself, badly and quickly, so that when you read someone else’s you know what it took.