How the research is done
Six stages. The same six whether the engagement is a one-question memo, a committee briefing or a multi-market comparative study.
Every stage runs on real, sourced series
Stage 02, shown · India CPI · World Bank
Nothing in the six stages below is built on a vendor summary or a stylised chart. The evidence is the published series itself, with its breaks left visible. Indian consumer price inflation is the cleanest example on the site: the series carries a definitional change part-way through, so its two halves are not the same measurement. It is India here because that is the market the published work reads most closely; the same standard applies to any market the question names, developed or emerging.
Indian consumer price inflation
Annual %, 1998-2025
Source: World Bank Open Data, FP.CPI.TOTL.ZG, retrieved 5 September 2026. India’s headline CPI basket and base year were revised during this period; the series is drawn continuously here, which is exactly the smoothing a report from this practice would flag rather than repeat.
The six stages
From the question to the decision
-
01
Define the question
Most enquiries arrive as a topic rather than a question, and a topic cannot be answered. The scoping call turns it into something specific enough to be wrong about, by working backwards from the decision behind it. If that cannot be done, the engagement does not start.
-
02
Establish the evidence
Official data, primary sources, policy and regulatory documents, filings and market evidence, pulled and cleaned in-house for every market in the frame. Where a series has been revised, redefined or interrupted, that is recorded rather than smoothed over, because those breaks are usually where the interesting answers sit.
-
03
Challenge the consensus
The prevailing view is written down explicitly, along with what it rests on. Often it rests on a single dataset, a definitional choice, or a comparison nobody has revisited in years. Knowing which one tells you how much weight the consensus can carry.
-
04
Build the analysis
Macroeconomic, financial and structural variables are connected into one argument rather than presented as parallel sections. The subject market, India most often in the published work but any market the question names, developed or emerging, is set against the peers competing for the same capital, so the finding is comparative from the start and not benchmarked as an afterthought.
-
05
Test the conclusion
The finding is attacked before it is delivered. What alternative explanations produce the same result? Which parts survive if the weakest input is removed? Anything that only holds under one reading of the data is labelled as such in the deliverable.
-
06
Make it decision-relevant
The work closes on what you should now understand differently and what follows from it, including the conditions that would change the view. A conclusion you cannot act on, or cannot defend when challenged, has not finished being researched.
Stage zero: choose the comparison
Before the six stages begin
The comparison set is chosen, not inherited.
A good analysis of any market does not reach for China and the United States by reflex. The peer set follows the decision. A manufacturing question runs across Vietnam, Mexico, Indonesia, India and China. A domestic-demand question runs across India, Indonesia and Brazil. A commodity question runs across Brazil, South Africa, Indonesia and Chile. A market-access question runs against the whole classified set, and the classifications themselves differ by index provider, which is part of the analysis rather than a footnote. Wherever the decision runs against a developed market, whether the capital could stay at home, the plant could stay where it is or the benchmark is a developed-market index, that developed-market anchor is written into the set.
The peer group is written into the scope before work starts, with the reasons. If the comparison set changes mid-engagement, that is recorded as a scope change, the same discipline the data gets.
Standards
Six, on every engagement
-
Sources are recorded and disclosed
Every figure in a report can be traced to its origin. If a number came from an estimate rather than a published series, the report says so and shows how the estimate was built.
-
The data appendix is reproducible
The appendix is written so that someone on your side can rebuild the work. That is a deliberate constraint on the analysis, not a courtesy at the end of it.
-
Methodology is stated openly enough to be challenged
Definitions, peer sets, date ranges and adjustments are stated in full. Research that cannot be argued with is not independent, it is marketing.
-
One revision round is included
Delivery includes a call to challenge the findings and one written round of follow-up, so the report can absorb what you know that the team did not.
-
Where the evidence is thin, the report says so
Some questions cannot be answered to the confidence a decision requires. Saying that plainly is more useful than a well-formatted guess, and it is what makes the confident findings worth reading.
-
What would change our view
Important published conclusions carry the observable conditions that would weaken or invalidate them. A conclusion without a falsifier is an opinion wearing a chart.
Independence, in practice
The practice sells research and consulting and nothing else. There are no financial products behind the analysis, no execution business, no distribution agreements and no external capital with a view on the conclusions. Independence here is a matter of structure rather than a claim in a brochure.
The practical test is simple. Where the subject market, India or any other, ranks badly against a comparator, the report says so first. That is what makes the favourable findings worth anything.
The method is also a matter of who runs it. Kakde Research is founder-led: Pranav Kakde scopes every mandate, reviews the work and stays on the client relationship, and a team of six analysts assembles, checks and maintains the evidence. The people you speak to are the people doing the work, which is why the six stages can be defended in person rather than by a department.
Have a question worth researching?
Describe the decision, the question behind it and the markets it touches. We will get back to you as soon as possible.