<img height="1" width="1" style="display:none;" alt="" src="https://px.ads.linkedin.com/collect/?pid=1905948&amp;fmt=gif">

Background Banner Video

Lorem ipsum dolor sit amet, consectetur adipiscing elit


Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua.

Enabling Product Strategy with Real-time Analytics Part 2 of 2

This blog will discuss how analytics should be implemented efficiently. The two topics covered in this document are: The technologies and techniques required and how to visualize the data in an easy to understand dashboard.

If you haven't read the previous blog which discussed the reasons as to why analytics should be integrated into a company for a better understanding of the customers and the products, we recommend you do so first.


Technologies and Techniques to Enable Real-Time Analytics

Once the goal and procedure for getting real time analytics set up, it's time to determine the tools and technologies required to enable that.

One of the best methods to set up analytics is that different teams of the company collaborate with each other. Product managers, Data Analysts, Web Developers and Data Engineers should work together to determine what metrics are important for the business, what are the critical features that aid with capturing and visualizing data and how those would be implemented. This collaboration would save time and reduce potential future errors.


One of the easiest and reliable web analytics tools is Google Analytics.  GA can be used in collaboration with other tools to collect, analyze and visualize data easily. Google Data Studio can help in visualizing data and is integratabtle with GA. The relevant homegrown data sources of the company should also be connected to the web analytics’ data sources. The reason behind this is that there should be a “Source of Truth” while analyzing the data for the product.


Visualizing data


Keep it simple

The first and foremost thing to keep in mind is to create an easy to understand and a simple dashboard. Non technical people who might have a lesser knowledge of raw data should not have to waste a lot of time understanding the visualizations. For example, in some cases it would be better to show the total number rather than showing the number in percentage and vice versa. Percentages are used when there is an emphasis on the  impact of the metric.  For example, if the source or the medium of incoming users to the website has to be compared, then showing percentages helps in visualizing how one source compares to the other sources.

There is also no need to make the visualizations futuristic or add an extra layer on those. 2d graphs work as good and are as informative as the 3d bar graphs. The color scheme of the dashboard should match the company’s theme and that gives it a more personalized feel. Therefore, easier but analytically correct visualizations go a long way.

Choose representative visualizations

The charts should be representative of the data and the dashboards should not be “filled” just to make a dashboard. The visualizations chosen should be the ones that are the most relevant and help understand the business goals. The charts should also be ordered to add a bit more information. Adding unnecessary charts is just directly proportional to the time interpreting the dashboards. Stakeholders might not have a lot of time to spend on understanding the charts, so those should be clear and to the point.

Choose the right chart

The charts should be chosen according to the audience, the project and purpose of the dashboard.The most used chart is the bar graph and many analysts use just that to represent the data. But that does not mean that bar graphs represent the data more accurately. It might even be misleading and hard to understand. For example, to show changes over time, a line chart would be a better choice as compared to bar charts. Showing interactive maps instead of tables with locations and numbers is better. To represent percentages and impact of various things, Pie charts will work the best. When working with text data and Natural Language Processing (NLP), word clouds can be of great help.


Don’t “lie” with graphs

The graphs shown should not misrepresent data in the dashboards. Analysts should not intentionally or unintentionally mislead the stakeholders. One of the common mistakes done by many analysts is not starting the Y-axis on 0. This leads to people thinking the impact and difference is way greater than it actually happens to be. Below example shows it perfectly: