Information Collection Techniques for Industry Research

In the fast‑changing Internet industry, the ability to quickly learn about new domains and build insights is an essential skill for professionals.

One effective way to acquire such knowledge is through industry research. In this session, teacher Xiao Teng invited Tencent’s Lai Xubo, who has over seven years of consulting experience, to share information‑collection techniques for industry research.

Note: All images are internal Tencent CSIG materials and may not be reproduced without permission.

Lai’s presentation is divided into three parts: the general workflow of industry research, basic information‑gathering techniques, and the pitfalls of deep interviews.

Industry research studies an industry as a whole, examining interactions among companies, the industry’s own development, cross‑industry links, and geographic distribution.

The fundamental purpose is to improve business stakeholders’ macro‑level judgment of the industry, ultimately driving profit; in short, the core questions revolve around money.

The basic framework is illustrated in the diagram on the left, with the right‑hand side showing eight written questions that operationalise those core issues.

Having a solid framework is essential, but the research also depends on the information required within that framework.

When a new research topic is assigned, Lai recommends six steps for information collection.

The first step is to read news to gain a basic perception of the industry, which prepares you for the second step of deciphering technical terms in research reports. After these two steps, you can conduct a benchmark‑company study, producing a preliminary industry report that later gets refined through deep interviews, synthesis, and polishing for clarity and accuracy.

Information collection can be broken into four steps: define requirements, choose channels, improve efficiency, and build habits.

Define Collection Requirements

Two points must be clarified: what problem are we trying to solve, and what information is needed to solve it. Problems should be broken down, prioritised, and communicated regularly with stakeholders.

For simple, point‑specific needs, data can be delivered quickly; for broader, line‑type needs, comprehensive collection and analysis are required, extending the delivery timeline.

When encountering a topic that requires definition of a term, the first step is to clarify the term itself.

Choose Collection Channels

Information is classified into three categories by difficulty: public, semi‑public, and non‑public.

Public information includes national statistics, listed‑company annual reports, and industry reports.

Semi‑public information consists of two types: information deliberately confined to small circles, and information lacking dissemination channels.

Non‑public information is only accessible to specific individuals; it is valuable but hard to obtain, typically gathered through personal networks or expert consultations.

Collecting Public Information

Public sources are the most commonly used. Lai recommends authoritative channels such as official statistical bureaus, reputable consulting firm reports (read the original English version when possible), and cross‑validation of repeatedly cited data.

Annual reports of listed companies are crucial; they can be sourced from US, Hong Kong, A‑share markets, and other primary markets.

Collecting Semi‑Public Information

For information deliberately kept within a small circle, the best approach is to become a member of that circle. For information lacking channels, field investigations can create new dissemination pathways.

Collecting Non‑Public Information

Non‑public data is hard to obtain but highly valuable. Two main strategies are building personal networks (“thick‑skinned” outreach) and seeking expert assistance through specialist firms such as GLG (global) and BCC (domestic).

Improving Information‑Collection Efficiency

After defining requirements and channels, Lai suggests a four‑step efficiency process: determine keywords, focus searches, organise importance, and refine information.

Step 1 – Determine Keywords: generate 10‑20 keyword variations (with/without quotes) for each core problem.

Step 2 – Focus Searches: allocate a dedicated time block (half‑day or full‑day) to search those keywords, collecting documents into a single file; switch keywords after ten result pages.

Step 3 – Importance Sorting: group collected materials, marking each as important, general, or irrelevant.

Step 4 – Information Refinement: read important materials deeply, extract needed facts, and discover new keywords for further search.

Company Annual‑Report Reading Tips

Annual reports are vital public sources. Lai advises searching for keywords like “Salar” to capture both singular and plural forms of “salary”.

Developing Good Search Habits

Lai emphasizes three habits: daily accumulation, willingness to share, and focusing on key industries.

When information collected is uncertain or unavailable, deep interviews become necessary.

Interview Subject Selection

Choose interviewees based on availability: start with frontline staff, then middle‑level, and finally senior leaders; adapt when the pool is limited, and consider the interviewee’s past experience if current role lacks relevant information.

Interview Guide Preparation

The guide’s purpose is to organise the focus of information collection, not to script every question.

Content should include: unresolved questions from prior collection, logical back‑casting of potential gaps, and background/character traits of the interviewee to tailor conversation topics.

The guide should cover over 90 % of needed questions, with remaining gaps filled during the interview.

Deep‑Interview Tips

For “cold” interviewees, close‑ended questions and concrete scenarios help elicit answers; if they remain unresponsive, replace them with more engaging subjects.

Article ends here. For the latest articles on cloud computing, micro‑services, and architecture, click the original link or scan the QR code.

Recommended Reading:

Evolution of High‑Performance Computing (HPC) Technologies and Industry Trends

Understanding Storage Performance Bottlenecks

DNS Load Balancing in Distributed Multi‑Active Data Centers

NVMe‑over‑Fabrics Architecture Overview

Warm Reminder: Scan the QR code to follow the public account and click the original link for more cloud‑computing, micro‑service, and architecture resources.