The business I work for is the most interesting aspect of my pointless existence. The numbers are a direct indicator of performance. These numbers can be approached as having a “high score”, where the highest score is the best. In order to have the high score, you need to mark certain the numbers are reliable. When you see your high score in Super Mario Brothers, there is no doubt Mario’s coin collecting is being counted accurately by the Japanese programmers. If only business were that simple.
Sam does a monthly report each month. I’m watching last month closely as we had spent $7,500 to re-design our homepage and my instincts tell me this will boost our conversion rate. Simple truth, if I had to choose between Jennifer Love Hewitt in Heartbreakers (the maid costume not the super tight pink dress which I also like a lot) or increasing our conversion rate .5%, I’d choose JLH. That conversion rate increase though is ahead of Angela Kinsey from the Office.
I’d been monitoring the conversion rate for the past 2 weeks and it was off the map since the installation of the home page. 4%. 5%. 8%. Moon Landing Incoming! The only issue was that it didn’t feel like we had increased our daily activity more than 2x. The orders felt about the same. You’d think you would notice if you normally ship 30 packages that you shipped 50 packages which wasn’t the case. So yesterday Sam and I discover that the site is double counting the conversions! This threw his report #’s off a bit, which is easily correctable, but it proves a very simple point, #’s don’t lie… unless you have bad data.
You Can’t Handle the Truth
Which brings me to Covid #’s. I read this in the WSJ today.
Eight months into the pandemic, U.S. testing data remains inconsistent. Some state agencies report the number of samples tested while others report the number of people tested, muddying comparisons of state positivity rates.
For example, individuals who have tested positive and get follow-up tests to see if they have cleared the infection might be reported as more than one positive case. Still, others who are tested frequently to make sure they aren’t infected and repeatedly test negative might inflate the number of negative results a state reports.
I’m not a numbers geek. I understand them but I’m not going to sift through millions of test subject data and come out with a conclusion that makes a bit of difference. I’m 100% certain though that having data that you’re mixing together without the same means of reference is stupid. If they still haven’t put together a system for consistent data quality after 9 months, why would you trust any of their #’s? They couldn’t even decide on how to get 50 states to track the virus accurately, how could you think that hospitals, that are given incentive to report Covid deaths, are doing that correctly AND that we’re counting them correctly?
I’ll conclude that I’m not a Covid denier. The virus is dangerous to the vulnerable. I’ve been in contact with hundreds of people at this point in the past 9 months and I’ve either never gotten it or have never felt it. Compare this to Howard Stern who hasn’t left his house in the past 9 months and you have two different approaches to the virus. My main concern is always the dying and the older you are, the better chance this disease will kill you. There is less than a 1% death rate if you are between 50-59. The elderly, who I’m sure already know this, should protect themselves. For the rest of us, hopefully we look back at Covid and wonder why China’s #’s are so much different than ours. I don’t even know where Moldova is.
Remember, we base most of our opinions on the numbers and the numbers do not always equal the truth. Ask those awesome pollsters who call up 12 people and predict an election. Yet people will fight you to the death because of a number discrepancy. It’s happening right now in our election and that #60 on that list is inexplicable if this virus is as evil as people claim.