When learn­ing to sky­dive, there is a cer­tain amount of train­ing that has to be done on the ground. Any­one can jump in tan­dem (attached to some­one who has done all the train­ing), but becom­ing a licensed sky­diver means gain­ing some expe­ri­ence. That expe­ri­ence is par­tially book work and a lot of prac­tice. Being an effec­tive ana­lyst requires a lot of the same tenac­ity and hard work, which is good because I wasn’t quite ready to aban­don this anal­ogy just yet.


When begin­ning any sort of process, be it sky­div­ing, pre­dic­tive fore­cast­ing, or even lion tam­ing, it is imper­a­tive to under­stand what you are get­ting into. Famil­iar­iz­ing your­self with the pred­ica­tive ana­lyt­ics matu­rity model in part is not enough. In some ways, learn­ing just a por­tion of the model or only imple­ment­ing it par­tially, is a lot like sky­div­ing with only some of the har­ness being secured. Who does that? When the chute opens (or you gain some exec­u­tive trac­tion), you’re most likely going to fall to your death or, even worse, not have answers for your boss.

In a sky­div­ing ground school sit­u­a­tion, you could be learn­ing how to strap on your har­ness, how to pack your chute, and per­haps even the lim­its of the human body. With ana­lyt­ics ground school you are learn­ing the first steps in apply­ing the tools of the trade. Becom­ing famil­iar with your tools and know­ing the basics are the keys to con­tin­ued suc­cess in analytics.

I think back to Tim Tebow (I can use more than one anal­ogy, it’s my blog) com­ing into the NFL. Tim’s passer rat­ing (his rat­ing based on accu­racy and pro­duc­tion as a passer—hard num­bers, mind you) sur­passed Pey­ton Man­ning, Eli Man­ning, and Tim Couch, all three of which are pre­mier quar­ter­backs in the NFL. So if Tim Tebow had great suc­cess at Florida, why was he largely inef­fec­tive once he entered into the pro­fes­sional league?

As it turns out, throw­ing the foot­ball side arm ver­sus throw­ing over the shoul­der takes some­thing like 0.6 sec­onds more. That six-tenths of a sec­ond was enough to limit his effec­tive­ness in the NFL. While he could get away with hav­ing a slower deliv­ery in col­lege, the level of tal­ent in the NFL is such that a quar­ter­back (or any posi­tion) must be at the height of effi­ciency. For him to be more effi­cient, Tebow had to revisit the basics and relearn some fun­da­men­tals in order to com­pete at an elite level. It’s not just Mr. Tebow either; all ath­letes prac­tice the basics over and over again. Why? Because they use them dur­ing every game. The same holds true for ana­lysts as well; know­ing the basics and prac­tic­ing them over and over again is the key to con­tin­ued suc­cess. The next ques­tion on your mind might be, “Where do I practice?”


It is dif­fi­cult to gain any sort of exec­u­tive buy-in by try­ing to affect your entire busi­ness all at once. The key really comes in with iden­ti­fy­ing smaller, more self-contained parts of your busi­ness and per­form­ing the basics in a prac­tice mode. How does this help you align your orga­ni­za­tion? Fur­ther­more, if you can­not com­mu­ni­cate the out­put to the exec­u­tives in your com­pany, how can you get them to under­stand the value of an ana­lyt­ics pro­gram? The answer is, you can’t. The solu­tion to over­com­ing the ini­tial hur­dles is fairly simple.

  • Iden­tify your spheres of influ­ence. What processes can you affect with­out those processes impact­ing any­thing else? Typ­i­cally, there are small micro-chasms within any orga­ni­za­tion that can be tweaked a lit­tle here and there. Once you have iden­ti­fied these processes within your orga­ni­za­tion, move on to the next step.
  • Take an aggre­gate look at the list of likely can­di­dates. Which processes are you most likely going to see an out­put from, and which are going to be harder to show for­ward progress and pro­duc­tiv­ity? You need to iden­tify the processes that will net mea­sur­able value. It’s not enough to stream­line some­thing and then set back and let your new process go to work. The end result of not hav­ing a mea­sur­able out­come is look­ing like you haven’t done any­thing at all. Hav­ing mea­sur­able met­rics (before and after your adjust­ments) means show­ing value.
  • Have an exe­cu­tion plan in place. A slide pre­sen­ta­tion might look good in the board room, but unless you can exe­cute a plan to solve a prob­lem or improve the process, all you have really done is cause your­self grief.
  • Exe­cute, exe­cute, execute.

Gather Your Tools

Once you’ve iden­ti­fied the process you will be work­ing with, it’s a good idea to gather and become com­fort­able with the tools you will be using. Adobe Ana­lyt­ics Stan­dard, for exam­ple, has anom­aly detec­tion built right into the soft­ware. This is a great tool to use in the beginning.

Here’s an exam­ple of where to start. I might see that vis­its to lead form page drop off. Let’s say there used to be 1,000 leads a day, but the leads have dropped to nearly zero overnight. While this would nor­mally be some­thing that would crop up on weekly ana­lyt­ics report or dash­board, solv­ing the prob­lem quickly means not los­ing trac­tion. So if the ana­lyst detects the anom­aly more quickly and iden­ti­fies the cause, the com­pany wins. Obvi­ously, you can show what would have hap­pened had you not inter­ceded and fixed a bro­ken or not well-executed process.

Think about where you can affect small and mean­ing­ful change within your orga­ni­za­tion, and begin using the tools and processes that will help you suc­ceed. Prac­tice your fun­da­men­tals and never take them for granted. In my next post ‚we will out­line some of the ways you can add to what you’re already doing and begin to add the lay­ers that will cre­ate a for­mi­da­ble and prof­itable pre­dic­tive ana­lyt­ics plan.