Six Common Problems with Data (Data-Driven Decisions)

Last week we talked about the importance of making data driven decisions in business. But there is something you need to watch out for. There can often be problems with the data. I will give you six examples in this video.

Twitter: https://twitter.com/FinanceWolves

TikTok: https://www.tiktok.com/@wolvesandfinance

Facebook: https://www.facebook.com/wolvesandfinance

Instagram: https://www.instagram.com/wolvesandfinance

VIDEO SUMMARY

All the problems we will discuss today can be described in one simple concept. Decisions are data plus judgement. We are talking about data-driven decisions, but you never want to make decisions on data alone. You always need to apply your own judgement.

Let’s go through some examples:

1. Bias. People make this mistake all the time. People will often have an agenda, and then find data to support their opinion. These people use data to shut down an argument. They will say, “The data says this, so you have to do what I say. End of discussion.” The issue is that often times the data can be interpreted in different ways. You have to be careful your bias is not influencing the data. Our goal is to get the best decisions, so ideally the data should be influencing your decision, not the other way around.
2. Errors. A lot of time, data just has errors. Or you might accidently have pulled the wrong data set from the wrong time period, or the wrong department. That is why you should never blindly follow the data. You always want to be double checking for errors. It is common for accountants to perform a “sniff” test. You pull the data, perform an analysis, arrive at a conclusion, but then you need to take a step back and ask yourself, “Does your conclusion ‘smell’ right?” If it does not seem right, or does not make sense, you should go back and check your data.
3. Statistics can be misleading. It is easy to generate statistics that match an agenda rather than reality. But wrong statistics will not lead to the best decision. One common example of this is for toothpaste ads that all seem to claim they are recommended by 9 out of 10 dentists. How many dentists were surveyed? Was the survey performed fairly or were the dentists pre-selected so they would select the preferred answer? It is easy to create misleading statistics, especially when you are dealing with percentages, so you need to be cautious you do not draw incorrect conclusions.
4. History does not necessarily predict the future. Just because you have historical data, does not mean you can use that to predict the future. For instance, just because you had a certain level of revenue last year, does not mean you will have the same level of revenue next year. Of course, history is useful in understanding the future, but you need to be careful drawing conclusions. The obvious example of this is how every single economic recession seems to take everyone by surprise. Just because the stock market is going up, does not mean it is going to keep going up.
5. Data may not capture all relevant issues. Many decisions are complex, and data sets may be too narrowly defined to support complex decisions where there are multiple stakeholders. For instance, if you are making decisions regarding employee compensation, bonuses, vacation, or training, you can look at salary and turnover data fairly easily, but does that really capture employee morale. Are your employees happy? Will your salary data tell you about happiness? I would argue that is an incomplete data set and may not lead to the right decision.
6. Short-term vs long-term issues. Many times, the best decisions in the short-term are different than the best decisions for the long-term. This has long been an issue with developing incentive packages for company management. Stock price and bonuses tend to incentivize short-term thinking that will actually harm the company in the long-term. This might require different data sets.

These are six common problems with data, but they all show the importance of the original concept: “Decisions are data plus judgement.” This is the mindset you need when you approach decisions. Data is a useful tool, but it is not the decision. You always need to step in with judgement.

There are often computer scientists that are building artificial intelligence machines. And these scientists claim computers will automate all decisions in the future. I always laugh at this, because people are forgetting what makes a decision. We need people to step in and use their judgement. You never want to blindly make decisions based on the data.

Let me use a simple example. Imagine you are running a manufacturing plant. This plant is making products and generating some toxic waste. You have to make a decision. Do you pay the cost to properly dispose of the toxic waste? That is expensive. Or do we simply dump the waste into the nearby river? That is cheaper. No one lives on the river, so the only things impacted is the wildlife, and wildlife cannot sue the company for damages. No one will likely find out about the toxic waste dumping. What do you do? The data is telling you that dumping toxic waste will make you the most money.

You might think this is obvious, that of course you would not want to dump toxic waste in the river. But that is an ethical judgement, that actually goes against the data that shows the drop in your profits. You need to step in and use your judgement to do the right thing.

Leave a comment down below letting me know what you think!

If you find these videos helpful, please subscribe to my YouTube channel.

Neither Zach De Gregorio or Wolves and Finance shall be liable for any damages related to information in this video. It is recommended you contact a CPA in your area for business advice.