Managing and analyzing a high-traffic site is certainly exciting. Your work impacts the experience of millions of people that interact with your site or mobile app on a daily basis. Working with a number of top 1,000 most trafficked sites in US, we have noticed, that there are some common challenges.
Unexplainable traffic spikes from different sources have likely caused a panic in your organization. The issue is often due to sampling. When using a free analytics solution, and not having access to raw data, sampling will greatly impact the quality of your analytical reports. In fact, you might see discrepancies with goals, conversions and sales over the same time period if running reports during different times of day. This is one of the easiest issues to address. An enterprise level analytics solution like Google Analytics Premium will give you access to raw data that you can then analyze without any discrepancies.
High traffic sites are often so popular due to content that they publish. It is often challenging to determine the attributes of top performing vs. low performing content. Proper analysis can help you generate a clear picture of what makes your content a hit with the audience.
If you are running a content site, most likely you rely on advertisements. Whether you are using DoubleClick for Publishers or another platform for serving ads to your visitors, combining this data with your analytics is a challenge. At the end of the day, you want to know what content makes you the most money, and tie your web analytics data with your marketing data.
Running a high traffic site demands attention to all the details. You need the ability to know what is not working and the agility to make the changes on the fly. Real-Time analytics that you can rely on is without a doubt an absolute must. Real-Time analytics for your app can help you determine and fix issues that cause app failures before somebody posts a negative review.
Sometimes looking at the data over the past week or month won't do. You might need to run reports for data over the past 2-3 years. If this is the case, you need access to the data warehouse where all this data is stored.