A model signal can tell you that a pattern deserves attention. It cannot tell you whether the underlying business is healthy, whether the market has already priced in the good news, or whether an approaching event could overwhelm the pattern.
That is the job of fundamental research.
The goal is not to collect enough facts to justify the signal. The goal is to try to disprove it before risking money. A useful research process should make it easy to say no, not yet, or I do not understand this company well enough.
Keep the signal and the business case separate
Write down the model output before opening another research source:
- ticker and exchange
- publication date
- Buy or Sell label
- evaluation window
- the model drivers shown by WealthGlider
- what would make the signal no longer useful to you
Then start a separate business-research note. Do not rewrite the model explanation as a company thesis. Price behavior and business quality can agree, disagree, or operate on different time horizons.
For example, a financially strong company can still be overpriced or face a difficult quarter. A weak company can experience a sharp rally. Fundamental research helps identify that tension; it does not make the tension disappear.
Step 1: identify what the company actually does
Begin with primary sources:
- the latest annual report and quarterly report
- the company's investor-relations site
- the latest earnings release and presentation
- recent material-event filings
- the company's description of its segments, customers, and geographic exposure
The SEC EDGAR database is the starting point for U.S. issuer filings. A third-party summary can make a company easier to understand, but use the filing to verify consequential facts.
Answer these questions in plain language:
- What does the company sell?
- Who pays it?
- What causes revenue to rise or fall?
- Which segment produces most of the revenue and profit?
- Is demand recurring, cyclical, regulated, or dependent on a small number of customers?
If you cannot explain the business without borrowing its marketing language, keep researching.
Step 2: inspect financial direction, not one isolated number
Compare several reporting periods. Look for direction and consistency across:
- revenue
- gross and operating margins
- operating income
- operating cash flow
- capital expenditures
- free-cash-flow calculations and their assumptions
- cash and debt
- share count
- stock-based compensation
One quarter rarely tells the whole story. A revenue increase can accompany weaker margins. Positive earnings can coexist with poor cash conversion. Per-share results can improve while the share count expands.
Write down both the favorable and unfavorable interpretation of each important change. If management calls a cost temporary, identify what evidence and future date could confirm that claim.
Step 3: understand the balance-sheet constraint
A signal's evaluation window may be short, but financing risk can change the outcome quickly.
Check:
- cash available relative to normal operating needs
- short- and long-term debt
- upcoming maturities
- interest expense
- covenant or refinancing disclosures
- working-capital movement
- dilution, repurchases, or recent capital raises
The question is not simply whether the company has debt. It is whether the balance sheet limits management's choices under a less favorable scenario.
Step 4: connect expectations to valuation
Valuation is not a single multiple. It is a statement about what must go right.
Compare the company with appropriate peers and with its own history, while accounting for differences in growth, margins, cyclicality, capital intensity, and balance-sheet quality. Depending on the business, useful measures may include earnings, revenue, cash flow, book value, or enterprise value.
Document:
- the metric you used
- the time period behind it
- whether the denominator is historical or forecast
- which peers you selected and why
- the growth and margin assumptions implicit in your conclusion
Avoid a valuation method you cannot explain or reproduce. A precise output built on fragile assumptions is still fragile.
Step 5: build a dated catalyst and risk calendar
List events that could occur during or near the signal's evaluation window:
- earnings and guidance
- regulatory decisions
- product launches
- investor days
- financing or debt maturities
- litigation milestones
- shareholder votes
- lockup expirations
- major economic releases relevant to the business
Separate a known event from a prediction about the event. "Earnings are scheduled" is a fact. "Earnings will beat expectations" is a forecast requiring evidence.
For every positive catalyst, write down at least one credible failure mode. For every bearish concern, identify what evidence would weaken it.
Step 6: use third-party research as a map, not the territory
Third-party platforms can help locate earnings-call transcripts, news, financial data, ratings, and competing interpretations. They are most useful when they expose disagreement or point you toward a primary source you have not considered.
Seeking Alpha is one independent research platform you may choose to consult for company research, earnings-related material, ratings, screening, and investor commentary. WealthGlider does not currently have an affiliate relationship with Seeking Alpha and receives no compensation for this ordinary link.
Treat every third-party rating or opinion as an input to investigate. Check its date, methodology, assumptions, author disclosures, and underlying source. Do not copy premium data or commentary into your notes for redistribution, and do not confuse access to more opinions with completion of due diligence.
Other useful sources may include company filings, investor presentations, exchange notices, industry publications, and reputable reporting. The right source depends on the claim you are testing.
Step 7: write the strongest opposing case
Before reaching a conclusion, write a short case against your preferred interpretation.
If the model says Buy, ask:
- What could cause estimates or margins to fall?
- Which favorable assumption is already reflected in the price?
- What does the balance sheet make harder?
- What would a skeptical customer, competitor, or shareholder say?
If the model says Sell, ask:
- What improvement could surprise the market?
- Is the negative thesis already widely understood?
- Could liquidity, positioning, or a catalyst produce a sharp reversal?
- Which operating metric would invalidate the bearish case?
A thesis that survives only by ignoring contrary evidence is not a researched thesis.
A compact research record
End with a dated worksheet:
| Field | Your note |
|---|---|
| Signal reviewed | Ticker, label, publication date, evaluation window |
| Business in one sentence | What it sells, who pays, and the main demand driver |
| Financial direction | Revenue, margins, cash flow, debt, and share-count trend |
| Valuation expectation | What appears to be priced in and the assumptions used |
| Near-term events | Dated catalysts and risks |
| Strongest confirming evidence | Primary source and filing date |
| Strongest contrary evidence | Primary source and filing date |
| Unknowns | Questions you could not resolve |
| Reconsideration condition | The fact or event that would change your conclusion |
Know when to stop
Stop or defer the idea when:
- you cannot explain the business
- a key claim cannot be verified
- the financial statements and promotional narrative conflict
- the outcome depends on a binary event you cannot evaluate
- liquidity, leverage, or dilution risk is outside your comfort zone
- your conclusion requires dismissing every opposing fact
Finishing the checklist does not make a security suitable for you. WealthGlider does not consider your holdings, objectives, taxes, liquidity needs, or risk tolerance. A model signal and a research platform can organize questions; neither replaces your judgment or advice from a properly licensed professional who understands your circumstances.