With big data becoming more prevalent, even in smaller businesses thanks to big data services, the IT department of many companies becomes the central hub for data analysis. This means IT managers have to understand the types of data that will be most helpful to different departments and will need to familiarize themselves with basic statistics principles to ensure that the data provided is valuable and will help improve the success of each department. When it comes to the marketing department, there are certain types of data that can be particularly helpful and some common pitfalls to avoid.
Look for Flow not Hits
Counting the number of hits on a website may be easy, but hits are ultimately useless unless visitors are actually going where you want them to go. Thus, looking at visitor flow will be much more beneficial for the marketing department. Visitor flow keeps track of where a visitor enters your website, tracks which pages they visited after that and ultimately, which page they were on when they left. This type of data allows marketers to experiment with headlines, links and call to actions, so they can get that visitor on the order form and then to click the “buy” button.
Conversion Rate Roadblocks
The conversion rate is a big factor marketer’s use to measure the success of a particular product or campaign, but sometimes the conversion rate is affected by other factors, such as poor layout or high shipping costs. It’s important to look for and identify these roadblocks not only so they can be removed but also so a successful campaign isn’t abandoned because the conversion rate is low. Looking into where a sale is abandoned can be particularly helpful. Abandoning the sales page could indicate a problem with another of factors, such as not giving enough information, pictures or having text that is difficult to read. Shopping cart abandonment could indicate a problem with the purchasing process or a problem with the price.
Successful Traffic Sources
Marketing departments spend a lot of time building up traffic sources to web pages via advertisements, social media, blogs and PR placements. The effectiveness of these sources is hard to measure, however, if IT doesn’t keep track of which ones actually get the consumer to click through to the website. In addition, looking through the data for external traffic sources is also incredibly valuable information, such as other websites or search queries that send people to your site.
Some Pitfalls to Avoid
- Don’t assume an action’s intent. It can be easy to start assuming a user’s intent if they leave a certain page or don’t complete a purchase, but there are generally multiple reasons for doing those things. Instead of assuming, always narrow down the possibilities as much as you can with additional data, customer surveys and experimentation.
- Don’t Forget Statistical Significance. Big data will help to address this issue by providing much larger data sets to analyze, but sometimes when a query is narrowed down the sample size is too small to be statistically significant. In addition, differences between one traffic source’s success and another’s should only be different if the difference is significant as well.
- Don’t Back Up Conclusions with Data. If you have already decided what your conclusion about a particular question will be, it is easy to find and manipulate data to support your thesis. The problem with this is the data is now providing no real value, and you could be going in the opposite direction that you should be. Instead, be open to any conclusion that data presents on its own.
Hopefully these tips help with your data analysis as you move further into big data, whether through a cloud database or an internal system.
