We live in a sea* of data, too much really. It takes a rocket scientist* to make sense of it. Maybe not.
*We love poor metaphors that involve the vast size of the ocean and true stereotype of all knowing rocket scientists.
The worst thing that can happen, and the motivation to share this post, is to use the data to make the wrong decision. Here is something we have learned, and it is somewhat offensive, so be ready; what makes a marketer GREAT is also what makes them bad at statistics. Most marketers are BAD at ‘google analytics’ math. The ability to prevent a data grenade will go a long way to not only achieving results, but knowing why.
The ability to ‘see’ the list of possible actions from the ‘sea’ of data….takes a few skills. We’d like to lay out a few of these things below. They are not a super secret, but often they are missed.
Put the Pin Back in Your Analytics
1. Go Slow
Ok. There are too many tools. Start slow. Pretend there is some type of real consequence for getting it wrong. If you are eating at a new restaurant, you don’t order everything and stuff it all in your mouth? This is such a gross example, sorry for the image, but it’s the same thing. Seriously, go slow.
Start here: Learn WHERE people are coming from and WHAT are they doing on your website.
The most valuable thing you can do is to understand your unique visitors, where they came from and what they did. If you have a high volume of unique visitors and they came ‘direct’…meaning they typed in your website address and 80% of them bounced..meaning they didn’t do anything on your website; THAT is very valuable. Something is causing it (or caused it). Getting a lot of ‘organic traffic’ that lands on a certain page, but the pageviews are low…maybe you hit on a good combo of key words….enhance your call to action on that page.
2. Measure Correctly
In a marketing world where you can do anything and everything, that is exactly what happens. We advise our clients to take a breath, and answer this; if you could only measure one thing, what would it be? This typically takes an hour or more; everyone wants (or thinks they want) to measure something different. A sea of choice only creates a data grenade; so much information, there is no clear action…OR WORSE…the wrong action. Oh My. Just measure the critical items; get momentum both in measurement and action.
Start here: Identify what you want to measure; your list will be long. Then group them by ‘why’…then order the metrics most important within each group. Measure #1 for each group.
3. Sample Size
We all have had the boss that has recited the 80/20 rule. Many decisions are made quick and using ‘numbers’. However…if your sample size is small, or your time period is small; you are toast. Burnt toast.
Be sure to understand ‘Statistical Signifcance‘.
We’ve used this graphic more than a few times as we talk to those that may not have the attention span necessary. That is ok though; it is their job; our job is to make great measurements and analytical conclusions. The signal is what you want to hear, the noise is what you don’t want to hear and the critical part is the sample size…a GOOD decision can’t be made without listening to enough people; a risky one can be though. It is your job to prevent this.
Start here: Use terms others understand, Nobody actually liked stats class; even rocket scientists didn’t like it.
Web analytics is really just a unique form of listening to your customers; you are listening to their behaviors; not their words. No bias; just truth. Hard to argue that.