This post is a reply to the post called “Big Data – What is Means for The Digital Analyst” by Stephane Hamel which provides an excellent overview of big data as a trend or hype and the key technologies involved. Stephane starts his post by mentioning that when he “Googled “Big Data” he got 19,600,000 results… Where there was virtually nothing about big data two years ago there is now unprecedented hype”. The digital marketing community, formerly called web analytics community, seems to be struggling with the key concepts behind big data and implications for digital marketing analytics. At the end of his post Stephane poses the question:
what should a digital marketing analyst do to bridge the knowledge gap between “digital marketing” and “big data”?
Having made the switch from web analytics to business intelligence I have some thoughts on this area.
Data scientist VS digital marketing analyst
First I would like to clarify some confusion regarding the different roles involved. A person working with big data in my opinion should not be confused with a “digital marketing” or “web analyst”. The proper role should be “data scientist”. What differentiates a data scientist from a web analytics or digital marketing expert can be illustrated through the following quote:
“A data scientist possesses a combination of analytic, machine learning, data mining and statistical skills as well as experience with algorithms and coding.”
One of the conclusions of the post of Stephan is that the current base of “web analytics” is generally not sophisticated enough to really leverage Big Data”. He also mentions the primary focus of digital analysts should not be on the lower-layers of infrastructure and tools development. I fully agree with this. To understand why, I think it is useful to look at a concrete example of how big data technology has successfully been applied within an organisation. This helps to get a better understand of the role of big data technology and concepts in the ecosystem as described in the earlier part of the article from Stephane Hamel.
Big Data Technology and LinkedIn
A great example of a successful company that uses big data technology is LinkedIn. If you log into your LinkedIn account you will see a lot of personalized content that is relevant for you, or people from your industry. LinkedIn provides recommendations of people you might know, relevant trending news from internet, jobs that are potentially interesting and more. To implement these services LinkedIn makes heavy use of big data technology. The latest LinkedIn IPad application as shown in the screenshot below is an impressive example of how “big data” technologies are used to create an innovative and superior user experience. Using the app almost feels like a fully personalized digital newspaper with relevant information.
screenshot: new LinkedIn Ipad App
To understand the role of big data in the applications above we find a lot of clues on the LinkedIn site. LinkedIn is an active contributor to the open source community and has a lot of interesting articles and webinars. To provide a brief summary:
- LinkedIn collects, stores and analyses massive amounts of data “from breaking news to breaking servers, real-world events requiring petabytes of data to be collected, analysed and acted on as those events happen.”
- To calculate something like “People you might know” LinkedIn uses Hadoop to “perform graph processing on a massive scale, such as scoring over a hundred billion relationships a day to compute.”
- To make all this “offline” data available to their live site, LinkedIn has developed a multi-terabyte scale data pipeline from Hadoop to an online serving layer called “Project Voldemort”, an in house developed NoSQL distributed database which they have open sourced.
Now looking at the examples above it becomes clear that there are big differences between the technologies being used by big data scientists and digital marketing analysts. In the case of LinkedIn they use big data technology for massive distributed and parallel computing for storing, analysing and finding the most relevant information for each user, before making the information available in a user friendly way for system developers to work with. Technologies as Apache Hadoop and Map Reduce, provide the basic framework that allows LinkedIn to distribute processing of large data sets across clusters of computers, where a single server would instantly run out of memory and disk space. Natural language processing, data mining and machine learning are techniques used to algorithmically determine which content is most relevant for each user.
Areas of opportunity for digital marketing analysts
Now let’s go back to the original question:
“How can we bridge the gap between big data and digital marketing analytics?”
At the end of his post Stephane identifies three areas of opportunity interesting to explore. All three opportunities can in my opinion be categorised as sub-areas of a new field called “data visualisation” within the area of business intelligence, which will be discussed in the next paragraph. Taken from the original article the identified opportunities are:
- Processing: Mastering the proper tools for efficient analysis under different conditions (different data sets, varied business environments, etc.)
- Natural Language Processing: Developing expertise in unstructured data analysis such as social media, call center logs and emails.
- Visualisation: There is a clear opportunity for digital analysts to develop an expertise in areas of dashboarding and more broadly, data visualization techniques (not to be confused with the marketing frenzy of “infographics”).
Data Visualisation Tools: QlikView and Tableau
Based on my professional experience the areas above are excellently covered by a relatively new and innovative area within the business intelligence field called “data visualisation”. Data visualisation products are easier to learn, much more user-friendly and have a higher focus on the end-user then traditional BI products. Within the business intelligence field there are two data visualisation providers mentioned in the “Gartner – Magic Business Intelligence Quadrant 2012″, with QlikView (QlikTech) as market leader and Tableau as challenger (see image below). Data visualisation products provide people the possibility to load large data sets from different sources in memory. Once loaded you can analyse, slice and dice the data, create beautiful visualisations and build live interactive dashboards which you can easily publish and share with other people in the organisation through an online portal. QlikView and Tableau require much less technical competency then traditional BI tools. The most important requirements are having an analytical mind-set and good understanding of your business.
source: Gartner Research Report 2012
To give a full description of how these tools work would go beyond the purpose of this article. However, to see the final results I have added two screenshots below taken from random dashboard examples from the QlikView and Tableau websites. To see real live dashboard examples I recommend going to the QlikView demo portal or Tableau public community section. Here you can view and download many different dashboards examples and click around to experience the products from an end-user perspective. The demo portals are on-demand implementations of their server products which work in a similar fashion behind a company firewall.
Link to live Tableau dashboard
Link to live QlikView dashboard
True power of data visualisation tools
Some final words on data visualisation products. Based on my experience the true power of these product do not come from their powerful analytical capabilities, but from the opportunity to help organisations become more data driven by:
- providing near real time data to monitor business processes
- completely automate manual reporting
- provide a single version of the truth (no more Excel data marts)
- promote data democracy by allowing people to develop and share insights, reports and dashboards with other people in the organisation through the server portal
- provide superior analytical possibilities through in memory technology and drill down possibilities
To summarize this post I believe the differences between the roles of big data scientist and digital marketing analyst are too big to gap. They should be treated as separate domains each with a unique place in the organisation. However, if you are a digital marketing analyst aspiring to work with big data analytics, or you would like to take your analytical career to the next level, I would recommend learning to work with BI/data visualisation tools QlikView and/or Tableau. The products are still relatively unknown within the field of digital marketing analytics and they have the potential to provide you or your organisation with a competitive advantage.
I hope the post was useful and I am looking forward to hear other people’s thoughts on the areas above…