A research desk built
the slow way
No brokerage arm. No distribution commissions. No issuer mandates. One revenue line, one job — to be right often enough, and honest every time we are not.
It started with a question nobody would answer straight
In 2014, our founding analysts were on the wrong side of a recommendation that came with no model, no assumptions and no explanation. When they asked what growth rate the target price implied, the answer was that the number came from the research team. That was the whole answer.
The frustration was not that the call went wrong — calls go wrong, that is the nature of markets. The frustration was that there was nothing to learn from. No chain of reasoning to inspect, no assumption to argue with, no way to tell whether the thesis had broken or the market was simply being slow.
SkyGrowthWealth Research exists to close that gap. Every note we publish carries the model that produced it, the assumptions inside that model, and the specific evidence that would tell us we are wrong. You are not asked to trust a conclusion. You are given the materials to test one.
That approach makes us slower than the sell side and considerably less prolific. We publish perhaps a dozen substantial documents a month where others publish a hundred. We think that is the right trade.
The desk today
Figures describe the scale of the research operation. They are not performance claims, and are pending verification before publication.
Where the money comes from
Subscriptions and commissioned mandates. That is the complete list. We hold no brokerage licence, distribute no products, and accept no payment from a company we cover — which is the only version of "independent" that means anything.
Make good reasoning available
To put institutional-grade research method — full models, forensic screens, stated invalidations — in the hands of investors who were previously offered only conclusions.
Research judged on its evidence
A market in which research is evaluated by the quality of its reasoning rather than the confidence of its delivery, and where revising a view in public is a mark of rigour rather than weakness.
Risk is decided first
Position size, invalidation level and drawdown budget are written before the thesis is. An idea that cannot survive that ordering is not an idea worth publishing.
Six steps, in this order, every time
The sequence matters as much as the content. Screening before modelling stops us falling in love with a story; stressing before publishing stops us shipping a thesis that only works in the base case.
Screen
A quantitative pass over the liquid universe — returns on capital, cash conversion, accounting flags, balance-sheet stress and valuation percentile — narrows thousands of listings to a working shortlist.
Model
Shortlisted businesses get a full three-statement model with segment detail. We then run it in reverse to establish exactly what growth and margin the current price already assumes.
Verify
Primary work — channel conversations, distributor pricing, capacity filings, regulatory documents — tests whether the model reflects the business or just the disclosure.
Stress
Each thesis is run against a downside case, a liquidity shock and a competitive-response scenario. If it only works in the base case, it does not ship.
Publish
The note goes out with the model attached, the invalidation level stated and the position-sizing guidance included. Nothing is held back for a premium tier of the same product.
Maintain
Coverage is reviewed every quarter and on every material event. When evidence moves against us, the revision is published — not buried in a later note.
Most of what we read never reaches you
Of everything screened in a typical quarter, a small fraction survives to a full model, and a fraction of that reaches publication. The funnel below shows a representative quarter. The discipline of throwing work away is the part that is hard to buy.
Sample distribution for illustration.
Six commitments we can be held to
These are operating rules rather than slogans — each one describes something we actually do differently, and each one costs us something.
Be right slowly rather than loud quickly
We would rather publish a week late with the model finished than first with a guess. Research that cannot be checked is just opinion with a logo on it.
Publish the reasoning, not just the rating
Every conclusion ships with the assumptions behind it. If you disagree with an input, you can change it and see what happens to the answer.
Say when we are wrong, in writing
Thesis breaks are published with the same prominence as initiations. A research desk that never revises is not being careful, it is being quiet.
Risk is the first page, not the last
Position size, invalidation and drawdown budget are decided before an idea is written up — not appended to it afterwards.
Independence is structural, not stated
No brokerage arm, no distribution commissions, no issuer mandates. The subscription is the only revenue line, which is the only way that promise holds.
Fewer, better documents
We publish less than our peers on purpose. Volume is easy to manufacture; relevance is not.
What is different, stated plainly
Four things separate this desk from a typical research subscription. None of them are marketing claims — you can verify each one from a single report.
The model is attached
Not a summary table. The working three-statement model, in a spreadsheet, with our assumptions visible and editable.
The invalidation is stated
Every thesis names the evidence that would break it, written before the evidence exists rather than rationalised afterwards.
Reversals are published loudly
Thesis-break notes get the same prominence as initiations. We keep the original note online next to the revision.
Sizing comes with the idea
A view without a position size is incomplete. Every note carries conviction-weighted sizing guidance and a concentration cap.
What we will and will not claim
Financial research invites overstatement. Here is the line we hold.
What we do claim
- Every published view carries its model and assumptions
- Our revenue comes only from subscriptions and commissioned mandates
- Analysts disclose personal positions in covered securities
- Reversals and errors are published, not removed
- Research is written to be checked by the reader
What we will never claim
- Guaranteed, assured or "risk-free" returns of any kind
- That past results predict future performance
- That our research is personalised investment advice
- Accuracy statistics we cannot independently substantiate
- Endorsements, awards or credentials we do not hold
Judge the desk by one report
Download an open-access note, read the reasoning end to end, and decide whether this is how you want your research written.
Research is educational and informational. It is not personalised investment advice.