Lessons from Sports Analytics for Small Businesses
Moneyball is a great book, and from what I've heard, a good movie, as well. It details the analytical tools, known as SABERMetrics, that the Oakland Athletics used in the early aughts to achieve success on a limited budget. There are quite a few lessons that can be drawn from the type of data-drive decisions that were made, and I want to look through some here. Not all are strictly analytics related, but I'm gonna talk about them anyway, because it's my blog. So there.
Here are three important lessons to be learned from Data-Driven Sports Strategy:
1) It's important to weigh the benefits of several smaller, less profitable options versus one bigger, more expensive one:
Think here of sports contracts, especially in a salary cap situation. Would you rather have one five star player and nine mediocre ones? (let's assume a ten player roster for the sake of this discussion). Or, would you rather have five good players and five mediocre ones? Or, would you rather have ten average players? The same logic can be applied to clients. It might be better to have many smaller clients versus one high profile client. Then again, it might not. It depends on how you rate the options. Let's look at this another way to further illustrate my point. We'll give each player a star ranking, and see how that affects things.
Option 1) One Five Star Player Plus Nine One Star Players
Option 2) Five Four Star Players Plus Five One Star Players
Option 3) Ten Three Star Players
Now, it's possible that a circumstance arises that means you value the players I've classified as One Star at Two. Maybe they're hidden gems, or maybe they're trainable (in business, think about a customer who starts out only buying one product but eventually buys several more as the equivalent). In that case, Option 2) looks better.
It all depends on your circumstances, and it's always something to consider. It may be that the huge player is worth Ten Stars, and the others all worth much less. This situation means that Option 1) looks better (although still not ideal).
Of course, there are other factors to consider
2) Risk management is a non-zero factor:
Let's say that your star player goes down with a season ending injury, or retires, or whatever analogy you want to use for loosing a big contract. If that is very likely, you're ahead to manage your risk somewhat. However, if you tend to have longer term contracts or your retention rates are good, you're more likely to be fine. Tying it back to sports, think about the difference between golf and ice hockey in injury rates. If you're a golf style business, having one huge contract could be profitable. But, if you're a hockey style business, having depth and flexibility could be key.
3) Acquisition cost is key:
Let's go back to our first example, and add a bit more to it. Instead of looking at a salary cap and fixed values, let's look at variable values in a sport like baseball with no salary cap (again, assuming that the roster of players is ten and costs are relatively uniform). We're going to try to provide the best value for money (benefit/cost). Even sports teams must be ultimately profitable, and a great team could be extremely expensive. I'll lay out the three alternatives again, and we can add variables to go more in-depth:
Now, let's assume that a Ten Star Player costs us $20, a Five Star Player costs us $10, a Four Star Player costs us $8, a Three Star Player costs us $6, a Two Star Player costs us $4, and a One Star Player costs us $2. (If you want, you can add a mental million after each number, I'm just saving some zeros.) This leads us to the following:
1) (1X5)+(9X1)=14 / (1X10)+(9X2)=18 | 14/18=.77
2) (5X4)+(5X1)=25 / (5X8)+(5X2)=50 | 25/50=.50
3) 10X3=30 / 10X6=60 | 30/60=.50
1) (1X5)+(9X2)=23 / (1X10)+(9X4)=30 | 23/30=.77
2) (5X4)+(5X2)=30 / (5X8)+(5X4)=60 | 30/60=.50
3) 10X3=30 / 10X6=60 | 30/60=.50
1) (1X10)+(9X1)=19 / (1X20)+(9X2)=38 | 19/38=.50
2) (5X2)+(5x1)=15 / (5X4)+(5X2)= 30 | 15/30=.50
3) 10x2=20 / 10X4=40 | 20/40=.50
Let's look at what happens if the costs are $30, $15, $12, $9, $6, and $3, respectively:
1) (1X5)+(9X1)=14 / (1X15)+(9X3)=42 | 14/42=.33
2) (5X4)+(5X1)=25 / (5X12)+(5X3)=75 | 25/75=.33
3) 10X3=30 / 10X90=90 | 30/90=.33
1) (1X5)+(9X2)=23 / (1X15)+(9X6)=69 | 23/69=.33
2) (5X4)+(5X2)=30 / (5X12)+(5X6)=90 | 30/90=.33
3) 10X3=30 / 10X9=90 | 30/90=.33
1) (1X10)+(9X1)=19 / (1X30)+(9X3)=57 | 19/57=.33
2) (5X2)+(5x1)=15 / (5X6)+(5X3)= 45 | 15/45=.33
3) 10x2=20 / 10X6=60 | 20/60=.33
4) Synergy is key:
I'm not going to return to our first example for this one, I'm gonna keep it a bit more simple (mostly because I know I'll be tempted to make a messy table out of everything if I'm have both synergy and cost of acquisition to combine. Let's assume that we are the hiring manager. Instead of our team representing clients, we're looking at our team as our staff. Is it better to have a team of ten with high individual skill and no cohesion? Or a team of low individual skill and low cohesion? Or a team of moderate individual skill and moderate cohesion? Let's take a look at a quick example: (I'll add skill points to each team member's indivual score and multiply it by ten, since the teams are uniform for this example.)
What about if the industry (or sport) is more team focussed and less individual skill intensive (football vs a relay race)? I'll flip the weights and look:
It's important to figure out how each of these metrics applies to your business. Keep in mind, these examples are quite basic, and I did that intentionally.