Data Due Diligence
A funny thing happens when your web traffic starts to travel up and to the right … you tend to overlook the outliers. Sure, you’re aware that those tiny spikes can offset your overall raw dataset by a small margin, or even a relatively significant one. But what about 30% of traffic? What if almost one in three uniques is screwing up your whole data? That’s what happened after postponing data due diligence.
In effect, what I thought to be exponential growth was just normal growth, after removing every single spike to smooth out the overall data. My new data is more accurate, less exciting, but it’s also more motivation to try harder and see if there are any improvement I can make to drive it up higher, faster. If you’re analyzing your data through Google Analytics, I implore you to clean it all up by removing data from yourself, your friends (maybe even your town if it’s fewer than 80,000 people) and any cities that have spiked only once a year (I’m looking at you Ashburn, VA and Boardman, OR). Set up those special segments in GA and dive a little deeper into your data. Mahalo.