Too Much Data in Reports
Published: 02/19/2010 02:26 a.m.
Ever used a program that does reports and when you go to select a report you get 100 options? How about a dashboard that features 50+ very important reports? I think there is a better way.
The Lemonade StandLets say I am opening a lemonade stand. I use ice, lemons, sugar, water, and cups. I sell large cups for $2 and small cups for $1. At those prices I make a $0.30 profit on the small and $0.60 on the large (simple, I know). We'll also assume I make a fixed amount of lemonade every morning and throw away whats left at the end of the day. I want to design a report or a dashboard that shows me actionable data.
Lets start with some reports I could make:
- Amount of sugar used today, this week, this month
- Amount of sugar left over today, this week, this month
- Amount of lemons, cups, etc. used today, etc.
- Amounts left over, etc.
- Large cups sold today, week, etc.
- Small cups sold, etc.
- Total cups sold, etc.
- Lemonade left over today, etc.
- Sell out time today, week, etc.
- Averages of any of these per day, week
- Average lemonade sold+left over per day of week over last 50 days
- Small to large cup ratio over last 20 days
- Profit totals per week, last 10 weeks
- Holidays in the next 10 days.
Small to large is to balance the prices and profit margins. If I'm selling much more of one size, then I raise the price. Supply and demand. I can also tell which size of cups to buy.
Profit totals let me track if I am making money or losing it. I can see how well my business is growing.
The holidays lets me know when I can expect a drastic change. I'll do my best to make more lemonade on those days.
Looking at these four simple reports daily can help me to make decisions about pricing, purchasing raw materials, and projecting growth or bankruptcy. Are they the absolute only reports I need? Probably not. But they will give me a great picture of my business and give me feedback I can use to make decisions.
The Other ReportsWhile there are a hundred other possible reports, I don't need to see those. They aren't actionable. Each one of the four selected reports tells me something I either need to change or keep the same. I don't answer every possible question, just the really important ones.
I know this is a simple example in a vacuum, but the point is illustrated clearly enough. Keep reports simple and actionable. Decide what will give you information that can be used to make a decision.
Here are some of the pitfalls of reporting. Making reports that are:
- Overly complicated
- Too simple
- Interesting but not actionable
- Too hard to distinguish change
- Too small a sample size to see results
- Too long or detailed
- Not providing feedback
Baseball AnalogyImagine looking at stats from 3 innings of a baseball game. There is very little if any usefulness from these stats. The only thing I can think of that you might use to make a decision is pitches thrown, which could indicate fatigue. Other than that, nothing is useful to make a decision.
Now take an entire game. You may have some insight about a player who got a couple of hits or a closer with 2 Ks in an inning. But the sample size is too small. You may get overly ambitious or depressed because of a small bit of data.
Let's go the other way and look at 4 years of stats for a player or team. Can you make a decision for the next season from these stats? Probably not a great decision. Most of the data might be too normalized or affected by other sources. Baseball is streaky, and teams tend to change some year over year. Over 4 years, you'd have a hard time what to expect in the next year. And some people use that type of data to predict the next four years. Ridiculous.
The ideal stage of data is probably around half a season, or a rolling 60 games or so. You get a big enough sample size while still taking into account current conditions. And you don't need every stat under the sun to make a good decision.
Less Reporting = Less DataIt's really tough coming up with a good, small set of reports. Most people have a fear of missing something, or they want to look at things to be able to know them. There is a desire to be able to extract every detail of possible data, but most of that data doesn't really tell you anything. A great report must be able to do the following things:
- Give you actionable data
- Quickly provide feedback from changes
- Use the right size of data
- Be visually easy to understand
The last part of a good report is that it helps you achieve your goal. Our goal with the lemonade stand is to maximize profit. The four reports we use help us do that both in the short-term and in the long-term.