The Hotel Problem

After a presentation on web analytics in one of my classes where I was discussing the different types of unique visitors a website can have. A particular part of the presentation seemed to receive a particularly good response, it was when I broke down a common problem for beginners to analytics, known as the Hotel Problem. I decided to show this problem here. This is taken from it's posting on Wikipedia.

Say you have a hotel that offers 2 different rooms. We will look at who uses which room over a 3 day period and how many unique 'visitors' you had.

• The first day, room A is rented by John and room B is rented by Mark. Thus 2 Unique visitors for the day.
• The second day, room A is rented by John and room B is rented by Jane, again 2 unique visitors for the day.
• The third day, room A is rented by Jane and room B is rented by Mark. For the last time, 2 unique visitors came to the hotel that day.

So in summary, each day we had 2 unique visitors, or to look at it another way, each room had 2 unique visitors.

Now for the easy question: How many unique visitors came to the hotel?

To help answer the question, here is a table showing what just happened:

Do you see the problem?

So how many did we have? 3? 4? 6? The numbers and logic support each of these answers.

Actually each of these numbers are correct, it just depends on what metric is being used.

• Based on days, there were 6 unique visitors, or 2 unique users each day
• Based on rooms, there were 4 unique visitors, or 2 unique users in each room
• Based on individual people, there were only 3 different visitors total over the 3 day period.

It is important to be remembered that unique visitors can be interpreted different ways. Depending on the business, one particular metric may be the best.

Expectations

Web Analytics is becoming more and more critical in today's business. Many companies have recently made a large push into this industry including

• Omniture
• Google
• Yahoo
• Microsoft

I am excited about this course (Isys 590R) because of all the new things I will learn this semester. I have experience setting up analytics software on different websites that I have worked on, but I am excited this semester about better learning how to interpret this information. I have been reading about some of the 'rookie mistakes' that are made by beginners like me when trying to interpret data. A couple common problems are the hotel problem and new visitors + repeat visitors being unequal to the total visitors. The reasons behind each of these mistakes are explained in the wiki article.

I expect this semester to learn more technical methods of tracking visitors over webpages. Having only worked with Google Analytics so far, very basic tracking software, I expect to learn about more complex and intensive tools that can be used to better know just what a consumer is looking for.

My last expectation involves this semester's competition. I began participating in the Omniture Web Analytics competition last semester, but my team had to drop out when one member had a pregnancy scare and early birth of his child. This year, and using methods learned in this class, I expect to not only 'participate' but also to 'compete' in the contest.