7 Skills in In-depth Research on Companies:Insights from Peking University

Insights from a Himalaya Capital Managing Director on how to conduct in-depth research on companies, based on his experience teaching at Peking University's Value Investing Course

Q1

Due to course constraints, it's quite challenging to thoroughly research a company in just one semester. From your perspective, how long does it typically take to conduct an in-depth study of a company?

Answer:

That's a good question.

Generally speaking, companies vary greatly, and understanding a company thoroughly depends on your access to different aspects of the organization.

For example, I've known researchers who worked faster because they had family members in senior management positions at the company, giving them better access to information. However, in our typical investment research process, even half a semester or a full semester isn't really enough to thoroughly understand all aspects of a company.

Of course, for any company you research, there are plenty of existing research reports, whether they're detailed sell-side reports or other analyses. But for this course, I recommend finding one or two good entry points that will impact the company's core value and focusing your research there.

Look for aspects that existing research reports haven't deeply explored, and conduct thorough research on these one or two points. In fact, I think you don't even need to worry about determining the company's exact valuation or whether it's overvalued or undervalued. Frankly, the time frame probably isn't long enough to figure all that out with just the spare time from one course. However, you can definitely gain deep insights into one or two key factors that affect the company's core value.

I've seen research reports from previous students, and some of them were truly impressive. Therefore, diving deep into one or two specific aspects is likely more valuable than trying to cover everything superficially.

Q2:

One basic requirement of the course is to list facts first. I've been thinking about what constitutes a "fact". From my understanding, information has different depths. After hearing your explanation, I feel we need to collect different pieces of information and piece them together to understand the facts. But this can be very challenging. If we need to write a report, even if we're just doing basic research like collecting cigarette butts, how should we define "facts"? Is it the first-hand information you mentioned earlier? In other words, what exactly is a fact?

Because information can be continuously dug deeper - this is what I gathered from your course. We might see some second-hand information and feel we can trace it back to become first-hand information, then investigate further through field research. We might look at company announcements or annual reports and then investigate further. For a normal report, most information should be first-hand material, but I think it might be unrealistic in the short term to have a report where everything is verified or even deeper than first-hand sources.

Answer:

Distinguishing between facts and views is indeed crucial. I've made similar mistakes before, and many of our researchers and former students have also confused the two.

For our research, separating facts from opinions is extremely important. Whether we recognize this ourselves or demonstrate it in our reports, you need to clearly distinguish between facts and views throughout the research process. This doesn't mean we can't cite news or data from sell-side reports, but you need to clearly state that this information comes from Report A or News Source B - be clear about your sources. Of course, whether what they're saying is actually factual requires our verification and research. Even in announcements from companies with questionable reputations that I've dealt with before, not everything can be taken as fact - it all needs verification.

When we're working on projects and writing final reports, we can say which information we've collected from various sources - reports a, b, c, d, e, annual reports e, f, g - and clearly cite all sources. Then we can explain which points might not be entirely accurate based on others' statements, what research we conducted ourselves, and what conclusions we ultimately reached through our investigation.

For example, last year a group researching baijiu (Chinese liquor) wrote about how a company announced they would expand production to XX tons and couldn't expand further. We can actually research whether it's true that they can't expand beyond XX tons, or if they're just saying this to influence market perception and prices. Or when they actually want to achieve further expansion, we can investigate whether local climate conditions or physical/chemical conditions would support it - these are all areas worth researching.

There's a common misconception here - it's not that everything in our report must be facts. But we need to be clear, both to ourselves and others, about what is fact and what isn't. This is what sets our project research apart from other sell-side research. Most sell-side reports, because they need to provide in-depth analysis, mix opinions with facts. When we review these, we need to clearly understand what comes from others and what we've verified through our own research.

Q2 Follow-up: Should we understand it this way: when we see information, the research deepens step by step. We might not be able to get many facts initially, but at least we know what's first-hand versus second-hand information, and as our understanding deepens, we can increase the number of verified facts?

Answer:

Exactly. It's actually quite difficult to thoroughly research a company in one semester since students aren't doing this full-time. As students taking this as just one course or part-time study, it's unlikely you can dedicate 100% of your time to it. So I think it's better to pick one or two key points for in-depth research - investigate whether something is truly factual, what the real situation is, what impact it might have. Solving one or two core issues that affect company value should be sufficient.

Q3:

When we're doing research projects, for example, if you studied chemistry but need to deal with physics or other technical fields, how should we expand our capabilities during the research process?

Answer:

That's a great question.

For instance, I graduated with a chemistry degree, so when I started working on research about power batteries and battery materials, it was relatively easy to get started. Research in physics, chemistry, or engineering-related fields definitely requires some STEM background and foundation.

But not having that background doesn't mean you can't do research - it's really about the return on investment. Simply put, if you have a business background, researching food and beverage companies or textile companies might be easier to start with, compared to researching semiconductors or computer technology which could be more challenging. You need to know where your circle of competence lies.

But if you really want to research semiconductors despite having no prior knowledge, it's not impossible. You can read relevant books, find experts to learn from, or learn from friends. When you know your boundaries, you'll have a better sense of the return on investment.

On the flip side, if you're researching a highly complex project that few others are studying, and you manage to understand it deeply while others haven't, it might take more time but the barriers to entry are higher. If you're researching something relatively simpler that's easy to start with, but you want to go really deep and see things others don't see, that might become more challenging. So knowing your circle of competence and which industries/businesses interest you more is particularly important.

If you can gain someone's trust and get reliable information from them, that might be more valuable than meeting many people but only getting superficial information.

Q4:

I have two questions.

First, you mentioned earlier about understanding an interviewee's position before interviewing them. For example, when researching chain stores and talking to store owners or distributors, you might find they have negative feelings toward the company, like if the company is restructuring distribution territories which might eat into their market. Based on your research experience, how do you assess a distributor's position, especially when you don't know them well?

Second, as you mentioned, when first meeting distributors, we might not fully understand their position or them personally, so the reliability of information from short-term research might not be high. When we have limited time, energy, and resources while researching a company, which approach do you think is better - broadly researching multiple stores or distributors to cross-verify information, or spending more time with one distributor to gain their trust?

Answer:

Regarding the first question, understanding your research subject's position beforehand is really important, like in the case of distribution channels you mentioned.

When channels are being consolidated and companies are increasing control over distribution, these distributors usually face a lot of pressure from the company, so you might hear many negative responses. When you know they're in this situation, you need to be aware of the context when listening to their feedback or information - which parts are influenced by their position. Having this awareness is generally enough.

For example, when a company starts channel reform, trying to get an objective view from distributors can be tough because their interests are affected - they're unlikely to say positive things. Even if they fully trust you and are telling what they believe is true, you still need to consider that there will be bias based on their position - what we often call "where they're sitting."

For the second question, about whether to dig deep with one person or cast a wider net and listen to many, I think this is a trade-off. Since everyone has limited time and energy, if you can gain the real trust of one or two research subjects, the information they provide might be more valuable than what you'd get from a broad approach, like having eight meetings arranged by sell-side analysts in one day. Of course, gaining someone's trust is like being a journalist - there's no one-size-fits-all technique that works with everyone.

Alternatively, after meeting many people, say eight, you might find one or two who genuinely trust you - this varies case by case. But if I had to choose between the two methods, gaining one person's trust and getting reliable information might be more valuable than meeting many people but only getting surface-level information.

Especially from a long-term perspective, someone who trusts you can introduce you to others who trust them, helping you truly expand your research network.

Q4 Follow-up: So we should understand this from the perspective of the interviewee's interests and position, right?

Answer:

Yes. Take the investor relations department as an example - they typically try to paint a positive picture of the company because they're responsible for the company's market value. Even if they tell you about poor quarterly performance, they'll usually add that a rebound is coming soon and say these issues are already reflected in the stock price.

Investor relations typically tries to make the company look good, while competitors usually try to make it look bad - that's just human nature. From another angle, your own distributors will say one thing, while competitors' distributors will say something different. Their stories usually don't match up. So it's important to recognize that people speak from their own positions. They're not trying to deceive you - they just have their own perspective, and they might not even realize it. But when we receive information, we need to keep this in mind.

Q5:

Regarding organizing research information - when we have diverse information from research, how can we effectively organize it and what tools can we use?

Answer:

This isn't an easy question to answer.

As I mentioned earlier, if you're researching market-related topics, you can organize the collected information around a specific question you're trying to answer.

However, when doing secondary market research, especially when dealing with information from company management or internal sources, you need to be mindful of confidentiality.

For example, I've seen some sell-side analysts sharing detailed research reports openly in chat groups or posting them online using tools like Youdao Notes. Personally, I feel this isn't very respectful to the research subjects - you should at least get their consent first, and there might be compliance issues to consider.

So when doing research, be thoughtful about when to take notes or record conversations. If your subject seems uncomfortable with your methods, be extra careful. For particularly important or core issues, consider their privacy needs and be careful not to let sensitive information leak - this helps build trust.

Q5 Follow-up: Could you elaborate on distinguishing between opinions and facts? When dealing with complex information, what's your process and what tools do you use? Do you write research reports directly, or do you use journals or other formats?

Answer:

I usually start with notes since they're part of my knowledge acquisition process.

If you keep logs, you can track your research on a company or industry over years, following specific questions you're trying to answer.

Regarding facts - what we think are facts aren't always permanent facts, and many things are dynamic. It often takes long-term verification to determine what's truly factual. The best approach in reports is maintaining good intellectual honesty - clearly mark what you've verified as fact versus what still needs verification. You might use red lines or crosses to indicate what hasn't been double-checked or cross-verified yet.

Time tends to resolve many questions when researching a company or industry. Previous statements and claims get verified naturally. If someone's past statements consistently prove true in annual meetings and research notes, they earn higher credibility. But if someone frequently makes unfulfilled promises, you learn to be more skeptical and do more cross-checking.

The purpose of field research isn't just to do it - it's to verify information you already have, test preliminary views, or validate what people consider to be facts.

Q6:

When conducting field research in a relatively short course timeframe, researchers' preconceptions can easily influence their question design, communication, and conclusions. They might look for evidence that confirms their existing beliefs. This effect might decrease with longer research periods, but do you have suggestions for avoiding this in short timeframes? Or should we question whether field research is even appropriate for short courses?

Answer:

In research, we don't do field research just for the sake of doing it. If you spot potential issues that could affect a company's core value or long-term prospects, you can use various research methods - whether it's technical research, document analysis, or patent verification. Field research isn't always necessary.

Field research isn't about checking boxes - it's about verifying information you already have, testing hypotheses, or validating what people consider facts. For example, if a home appliance company has previously inflated sales through channel stuffing, field research can be very helpful in verifying this, though it might work for a year or two but isn't sustainable long-term.

Honestly, I don't consider meetings with top management as field research in the strict sense. Information from channels, competitors, and customers often differs from what the company says, which makes field research necessary.

A few years ago, Professor Dong Baozhen, who taught us, actually stood outside warehouses counting delivery trucks and talked with packaging workers about daily output. This kind of research often reveals situations quite different from common perceptions.

In China's capital markets, where there's a focus on shorter-term gains, field research can help you see things others might miss - that's its real value.

For time management, we can focus narrowly on specific aspects and dig deeper than others. Sometimes understanding one aspect deeply can be more valuable than having broad, shallow knowledge.

Q7:

You mentioned that sometimes it takes multiple attempts and long-term relationships to get satisfactory answers about crucial aspects of a company's value. Could you provide a specific example?

Answer:

Going back to the sales channel example I mentioned earlier. I once wanted to verify whether a company was inflating revenue and profits through channel stuffing, so I researched their distributors.

With help from friends, family, and sell-side analysts, I contacted several distributors, but initially only got fragmented information, with each person's situation being different. After meeting two or three times in the first year, they started trusting me more because they saw I wasn't using their information for stock trading, but genuinely trying to understand the situation.

One distributor particularly came to trust me because he saw I wasn't just seeking short-term gains but doing serious research. He found it interesting and became very helpful. He even offered to connect me with his friends in different regions (since distribution situations vary by area). During annual distributor meetings, which might have hundreds or thousands of distributors, many people get to know each other. Once he trusted me, he could recommend local contacts wherever I wanted to research. He listed six or seven people, and from each of these contacts, I could connect with several others. By connecting these dots into lines and then surfaces, I gained a clearer picture of the overall distribution channel - maybe not better than the sales company itself, but definitely better than sell-side analysts.

Even though we started with just one person's introductions, their full trust meant their contacts also trusted us, which greatly helped our research. While we can't be friends with everyone, having one or two people who truly trust you is invaluable and can lead to significant progress.

This also answers the earlier question about time management. Even within a course's timeframe or two to three months of spare time researching a company, we can narrow our focus - like studying just the channel situation and potential channel stuffing, since this affects long-term company value.

We might only have a few hours per week for research, but full-time sell-side or buy-side professionals, while working full-time, have to cover all aspects of a company and often dozens of companies in an industry, so their depth in any one area might be less than yours.

I hope this course experience helps everyone gain something valuable by going deep in specific areas. Think of it like this: all analysts have a certain circle of knowledge about a company - some circles bigger, some smaller, with maybe 90% overlap. But you can add a small peak to that circle - your deep research area.

To use an imperfect analogy, it's like getting a PhD in one specific aspect - going deeper than others in one area might be more valuable than having the same broad knowledge everyone else has. That's why reports don't necessarily need to give valuations or buy/sell recommendations - thoroughly researching one specific aspect is already quite impressive.

The longer you research, the more efficient you become. Like building a house, you need to lay the foundation before building the second floor.

Q8:

Setting aside report writing, how do we distinguish true from false information when researching a company long-term amid complex information? How can we improve research efficiency and focus on value points? Based on my experience, analysts face huge amounts of information, and when researching large, established corporate groups with complex organizational structures and business lines, how can we work more efficiently and avoid wasting effort?

Answer:

First, most students taking this course submit reports because they want to apply Professor Jiang and Professor Chang's value investing principles in practice this semester - that's just one part of research. But for long-term research, once you've deeply studied a company for about a year or two, you develop a sense for distinguishing real from false information, whether it's media releases or sell-side reports.

Most importantly, when you reach a certain depth or duration of research, you develop common sense, and time helps verify many claims through cross-verification. For example, after studying a company for a while, you can quickly identify which social media accounts speak for the company, which represent distributors, what media tends to focus on, and which accounts are just creating hype.

Generally, research efficiency improves with time, but personally, I can't easily distinguish effective from ineffective information when first looking at a company. It takes time - like building a house, you need to lay the foundation before moving to the second floor.