---
title: "Run, Hustle. Eight Ways Data Can Fool You"
description: "In today's business world, not knowing data analysis is outdated. Many leaders demand data for decisions, making data analysis crucial and data analysts one of the top ten promising careers. However, data can be misleading. For example, during the US-Spain war, the Navy's death rate was 9 per 1,000, while New York City's was 16 per 1,000, but these figures are not comparable due to different demographics. This article outlines eight common ways data is used to mislead, from fabricating data to using charts to distort reality."
author: "黄成明"
publisher: "New Distribution"
email: "zhaobo258@gmail.com"
telephone: "+8615854817671"
published: "2015-06-07"
language: "en"
canonical: "https://xinjignxiao.com/en/articles/run-hustle-eight-ways-data-can-fool-you-432a797a/"
markdown: "https://xinjignxiao.com/en/articles/run-hustle-eight-ways-data-can-fool-you-432a797a.md"
original_source: "https://mp.weixin.qq.com/s/MZRnxYj5hEF_YxjEAJpEXg"
translation: "https://xinjignxiao.com/zh/articles/%E5%A5%94%E8%B7%91%E5%90%A7-%E5%BF%BD%E6%82%A0-%E6%95%B0%E6%8D%AE%E5%BF%BD%E6%82%A0%E5%85%AB%E5%BC%8F-432a797a.md"
attribution: "New Distribution — https://xinjignxiao.com/en/articles/run-hustle-eight-ways-data-can-fool-you-432a797a/"
usage_policy: "https://xinjignxiao.com/ai-policy.txt"
---

# Run, Hustle. Eight Ways Data Can Fool You

> In today's business world, not knowing data analysis is outdated. Many leaders demand data for decisions, making data analysis crucial and data analysts one of the top ten promising careers. However, data can be misleading. For example, during the US-Spain war, the Navy's death rate was 9 per 1,000, while New York City's was 16 per 1,000, but these figures are not comparable due to different demographics. This article outlines eight common ways data is used to mislead, from fabricating data to using charts to distort reality.

Now, if you're in sales or marketing and don't understand data analysis, you're behind the times. Many business leaders say, "Show me the data; how can I make decisions without it?" Data analysis plays a vital role in modern business management, and data analysts are among the top ten most promising careers for the next decade.

Let's look at a case of using data to fool people:

> During the US-Spain War, the death rate in the US Navy was 9 per 1,000, while at the same time, the death rate for New York City residents was 16 per 1,000. Navy recruiters used these figures to argue that joining the military was safer. Do you think this conclusion is correct? Of course not. The two figures are not comparable. Soldiers are young and strong, while the resident death rate includes the elderly, sick, and disabled, who have higher mortality rates. The correct comparison would be between Navy personnel and NYC residents of the same age group.

Indeed, 9‰ and 16‰ are not comparable.

Business managers detest "fake" data. The reasons are obvious: "fake data" leads to wasted resources, poor decisions, and missed opportunities. Let's summarize the common types of "problematic data" to help you develop a discerning eye. Note that "problematic data" doesn't necessarily mean "fake" data; sometimes the data is real, but the conclusion is false. Here are common ways data is used to mislead:

**1. Fabricating "fake" data to fool customers or consumers**

Forgive me for using the verb "fabricate."

This is common. For some people or organizations, data integrity is meaningless; they make up whatever data they need. They are the "fabrication committee." In such cases, always ask why and demand the data source. Remember: "No data (source), no truth." For example, newspaper circulation figures are always a mystery. I don't know the answer, but I know:

1. Media-reported circulation is often their highest ever, and they conveniently drop the word "highest."
2. Some newspapers, to boost circulation, would haul newspapers from the printer straight to the dump. This was blatant and shameless fraud, later banned.

Check this sentence for errors: Sales rep Xiao Qiang has 24 clients. In April, the non-duplicate purchase rate was 78% (Note: non-duplicate purchase rate = clients with orders / total clients). The answer is wrong because 78% is impossible to calculate from these numbers.

**2. Cherry-picking (Targeted Sampling)**

This is a hidden and deceptive tactic. What is cherry-picking? It's when you assume a conclusion first, then select the most favorable group for a survey or study, and finally claim the result is universal. For example, to make average wages look high, survey office buildings; to make them look low, go to labor markets. This is a trick, but many people love it!

Market research firms and some government agencies are masters of this. For instance, a region announced it would cut housing prices by a certain percentage within six months. After six months, they claimed success, but the public didn't feel any price drop. Why? They played a numbers game: the initial sample was urban average prices, but after six months, they included suburban prices in the average.

Most market research firms are keen on cherry-picking. Many business owners instruct research firms to sample according to their desired conclusions, then use that data for advertising and PR to deceive consumers. Some companies' survey data is real (i.e., sufficient sample size and no targeted selection), but the conclusion is false. Companies can also cherry-pick conclusions. For example (hypothetical data for illustration, don't take it seriously): a toothpaste ad claims using their brand reduces cavities by 23%. This data comes from a survey. It sounds appealing because you think "reduce" means fewer cavities. But behind it could be: 23% had fewer cavities, 40% had no change, and 37% had more cavities (though unlikely).

Look at this picture and you'll understand.

**3. Tian Ji's Horse Racing**

You've probably heard of Tian Ji's horse racing strategy. Misleading with this tactic is common. Example: At the end of 2010, a well-known B2C site held a "National Shopping Spree" event. After it ended, someone tweeted: "Based on transaction data, the average daily transaction volume during the four-day promotion far exceeded the combined daily average sales of Gome, Suning, and Bailian in 2009." This statement is technically correct, but the comparison is invalid: comparing your promotional peak with others' regular daily sales is meaningless. It's like Liu Xiang winning a Paralympic gold—different categories.

Another set of data: From Dec 20 to Dec 26, 2010, the weekly box office for "If You Are the One 2" and "Let the Bullets Fly" were 240 million and 210 million yuan respectively (Note: "If You Are the One 2" released Dec 22, "Let the Bullets Fly" Dec 16). Can we conclude that "If You Are the One 2" outperformed "Let the Bullets Fly"? From a pure data perspective, these figures are not comparable because Dec 20-26 was the first week for "If You Are the One 2" but the second week for "Let the Bullets Fly." Typically, blockbusters peak in their first week. If we look at first-week data: "Let the Bullets Fly" earned 290 million in its first 4 days (average 70 million/day), while "If You Are the One 2" earned 240 million in its first 5 days (average about 50 million/day). So "Let the Bullets Fly" actually did better!

Tian Ji's horse racing is essentially selecting data to fit a conclusion. Data comparability is something we must always watch. Sometimes seemingly reasonable comparisons are actually unreasonable.

**4. Systematic Errors in Data Analysis**

Data analysis can be affected by human factors or systematic errors. Example: HR wants to survey employees' opinions of the new general manager. Options: strongly like, like, neutral, dislike, strongly dislike. Anonymous voting. Results: strongly like 25%, like 40%, neutral 20%, dislike 10%, strongly dislike 5%. Since it's anonymous, you might think the data is reliable (assuming no sycophancy).

My answer: not necessarily. Many employees might not have voted. They might not have known about the survey or were too busy. Also, those who abstained might have wanted to vote "dislike" but chose not to express their true feelings. Think of abstentions in the UN General Assembly. Additionally, if the options were reordered (strongly dislike, dislike, neutral, like, strongly like), the results might differ even with the same voters!

**5. Seeing is Believing? Chart Deception**

If the above hasn't fooled you, here's data plus charts for double deception. You think a picture is proof. Look at these two charts—which do you prefer?

Notice the difference? Both charts use the same data: market share grew from 23.5% in 2005 to 24.8% in 2010, an increase of only 1.3%. The first chart looks like a dramatic rise, while the second shows no highlight! Some companies even hide the Y-axis values, claiming it's confidential. If so, you're completely fooled.

For some, these charts serve different purposes: the first can be shown to consumers to exaggerate market share growth, while the second can be shown to the board when asking for more investment, arguing that growth is minimal and needs funding. When you need to scold subordinates, the second chart works too. Of course, this is a joke. Excel usually auto-adjusts the Y-axis, but it can be manually adjusted. If adjusted too much (like in this example), it must be clearly labeled; otherwise, it's misleading.

**6. Pre-set Conclusions**

This is easy to understand: you have a conclusion first, then find data to support it. For example, this expert's analysis to match the government's proposed retirement age of 65 is quite a stretch:

Another example of pre-set conclusions: This image went viral on Weibo years ago, titled "I finally discovered the secret of 1-9."

The "secret" is that 1 has one angle, 2 has two angles, 3 has three angles, and so on. This is a classic case of pre-set conclusions. I can barely accept 4 having four angles, but how does 7 have seven? And the most absurd is 9; the author worked hard to prove 9 has nine angles.

In reality, many business leaders unconsciously fall into the trap of pre-set conclusions (let alone conscious ones). For example, a leader tells a subordinate: "Xiao Wang, do you think this month's poor sales are due to low repeat purchase rates?" That's a mild pre-set conclusion; the subordinate will investigate repeat rates as directed.

A severe case: At year-end, the boss says: "Xiao Wang, analyze whether we can achieve our 1 billion target next year." Is that a hint? It's an explicit instruction. Rest assured, the "savvy" Xiao Wang will find ways to justify the 1 billion target, but the poor salespeople on the ground will suffer.

**7. Various Rates That Are Hard to Calculate**

It's said the top three cities for divorce rates are Beijing 39%, Shanghai 38%, and Shenzhen 36% (data from news media). Upon checking, the formula is: divorce rate = divorces / total marriages. At first glance, no problem. Is the 2010 divorce rate simply 2010 divorces divided by 2010 marriages? Wrong! It's not apples to apples. People who divorced in 2010 are not the same as those who married in 2010. This calculation is not only unreliable but also misleading. Most media divorce rates are calculated this way.

How should we calculate divorce rates? We can modify the formula: The divorce rate for people married in 2000, as of 2010, = (number of people married in 2000 and divorced in 2010) / (total marriages in 2000). Incidentally, if we calculate the divorce rate for the 2000 marriage cohort each year, we could analyze whether the "seven-year itch" exists.

Many retail companies calculate return rates monthly, which is similar to divorce rates; they need to be categorized to be accurate.

**8. The "If...Then..." Hustle**

This method is common in startups or scam companies. The typical phrase is: "If every Chinese person gave me one cent, I'd be a billionaire." Sound familiar?

An evolved version: China has 1.3 billion people, our target market is 30% of them. If 20% of those buy our product at 100 yuan each, our sales would be nearly 10 billion. So the market is huge, and we have a great opportunity!

Another: During Spring Festival, WeChat bound 200 million personal bank cards in just two days. If 30% of users send 100 yuan red packets, that's 6 billion yuan in flow. If payment is delayed by one day, with private lending monthly interest at 2%, the daily return is conservatively 4.2 million yuan. If 30% of users don't withdraw cash, their accounts could hold 1.8 billion yuan in cash, with no interest cost (via @data view).

If you pay attention, many WeChat business (micro-commerce) people use this hustle.

There are many data tricks, especially in China, a place full of knockoffs and scammers. So, keep your eyes wide open. May you soon develop a discerning eye.

Reply with the following keywords to search for related professional articles: Sales Supervisor, Second-tier Management, Regional Manager, Distributor Management, New Channels, City Manager, Competition, 2015, Manufacturer-Dealer Game, Product Stagnation, Terminal Visit Management, Route Management, Deep Distribution, Internal Management, Sales Skills, Profit Improvement, Recruitment, Distribution, Daily Management, Team Motivation, Trade Promotion, Sales Misconceptions, New Product Launch, Township Market, New Product Pricing, Sales Target Achievement, Closing, Market Visit Inspection, Baijiu, Beer, Sales Increase, Agency Products, Channel Conflict, KA, Terminal Merchandising, New Market, Market Operation, Learning, Book Recommendations, Inventory Management, New Salespeople, Consumer Promotion, Execution, Old Products, Expired Product Handling, Model Market, Franchise Recruitment, New Media, Distributor Development, Performance Appraisal, Assessment, Annual Planning, Shopping Guide, Morning Meeting, Display, Transformation, Inventory Pressure, Holidays, Distributor Cost Control, Channel Operation, Marketing Theory and Laws, Brand Truth, Order Conference, Team Management, Training, Debriefing, Debriefing Report.


---

## Copyright and AI use

This article is sourced from New Distribution. Search, quotation, summarization, and model training are permitted, but every use must credit New Distribution and retain the canonical source URL.

Contact: zhaobo258@gmail.com · +86 158 5481 7671