Metrics for their own sake (analytics thought of the day)

Journalism at its finest!

Sometimes the most observable metrics are the most meaningless. Stock prices happen to be both highly visible and often misunderstood. The price of a stock can tell you precious little about how a company is performing.  In fact, over time you may notice that the price of a stock has gone from $100 to $75. You might be inclined to think it’s lost considerable value while in reality it might have split, then improved 50%!

The headline also makes the reader think that the value Priceline is even remotely near the value of Google or Apple. Not all companies have the same # of outstanding shares.

Google: 325 million shares

Apple: 932 million shares

Priceline: 49 million shares

Quickly multiply that share count by $1000 per share and you’ll see what I’m saying!

Further, the companies in question may or may not pay a dividend which represents a recurring cash value per share. The companies’ profit or earnings per share relative to share price (P/E ratio) may also be different depending on their growth prospects – which may impact your risk as a shareholder. Additionally, the company may have TONS of cash on their balance sheet, which in a way offsets the price of the stock because cash is cash. Apple recently announced a dividend on common shares because they were generating (and not investing) tons of cash!

What does this have to do with analytics you might ask?

Anytime you have 2 or more metrics you’re free to create a ratio. Add in a few dimensions such as date for trending, category or subcategory if you’re dealing with retail data, or even traffic sources – now you have a metric stew. You struggle to find relationships between metrics and your bottom line…you search for answers in the data but the data itself cannot reflect on its origins and tell you WHY it is the way it is, it’s simply a signal, which in many ways is a shadow of the true processes that drive your bottom line metrics (sales, leads, etc…).

While I’m wearing a neon shirt, my shadow won’t give that away. This is really a metaphor for that which you cannot observe. Sometimes analytics is like that. On the other hand sometimes the data is dead obvious, eg – problem was found, change was made, metric moved up! Without data quality neither will ever happen and the shadow you see will be an object utterly different from the one in the mind’s eye.

A story about measurement…

An economist held up a jar of beans to his class, polling the students as to how many beans were in the jar. In reality there were 800. Most under-predicted (400-600) while one made a dumb guess and said ’6,000′ just to be controversial.  The average ‘guestimate’ was precisely correct due to the presence of the outlier…

What’s worse? Saying on average student predictions converge to the true mean, or getting rid of the outlier? The moral of the story is just to be careful about precisely WHAT is going into your metrics. Most metrics are generally useful but don’t take them for granted…they may evolve in unexpected ways.

Over 500 words – must have been a good post!

Jeff