A pile of tools and a system that keeps running are two different things.

You now hold nearly every tool to analyze a SaaS company. But tools aren’t ability — a carpenter’s worth isn’t how many he owns, it’s whether he can build good furniture, reliably, again and again. This final chapter integrates twenty-two chapters into a research system you can own and run for a lifetime, in three steps.

Step one: screen with an eight-dimension scorecard (TAM, growth quality, retention, moat, profit & cash flow, management, AI opportunity/threat, valuation). But hear the limit clearly: this is not a formula that spits out “which to buy” — that would repeat the Rule-of-40 mistake of letting one number erase the important information. Its value is in the filling out (it forces you to confront the evidence on every dimension and exposes the blanks in your knowledge), and you read the weakness, not the total: a high total with one fatally low score is a warning, not a green light.

Step two: turn research into a process. One-time research goes stale fast — companies, competition, AI, and price all move. So run a quarterly tracking sheet, and for every company write down and maintain three things: the thesis (why it’s a good/bad investment, in one line), the key assumptions it depends on, and the kill criteria — the evidence that would prove you wrong. Kill criteria are the moral core of the whole system, because they force you to answer, in advance, “how might I be wrong?”

Step three: fight your own brain. The last enemy is cognitive bias, and three are lethal in SaaS. Narrative bias — captured by a great story; the antidote is to make every conclusion verifiable by hard operating data. Survivorship bias — studying only the winners and overestimating success; the antidote is to actively study the failures. Valuation anchoring — hostage to a stock’s old high or your cost basis; the antidote is to value from zero every time, on today’s fundamentals.

What ties the three steps together is a refusal to treat any judgment as final. The scorecard gets re-scored as new data arrives; the thesis gets re-tested every quarter; the kill criteria are checked, honestly, for whether they’ve quietly been triggered. Most investing mistakes aren’t analytical failures at the moment of purchase — they’re the failure to update when the facts changed and the story didn’t. The system’s whole purpose is to make updating a habit, not an admission of defeat.

Combine the three into a loop — screen → deep-dive + value + write thesis/kill criteria → track → revise → repeat — and you no longer depend on anyone’s tips or any hot narrative. The system’s deepest value isn’t helping you analyze companies; it’s helping you fight yourself. That’s the real gift of the book: not a list, but a system; not an answer, but an ability.

This is Chapter 23 of “Compounding in the Cloud.”