INFRASTRUCTURE, DATA, OR ANALYTICS AS A SERVICE
One way to reduce the burden of building a big data infrastructure and an army of data scientists within a corporation is to use external sources of analytics. The big data marketplace has provided a much- needed avenue for the external sourcing of analytics. A strong business model around monetization has facilitated a rapid rise in cloud-based companies offering services.
Most IT organizations would like to build the big data infrastructure within the firewalls of the organization. However, this has turned out to be a difficult task. Big data analytics requires a sizable big data infrastructure, which is hard to acquire, install, and configure. The personnel closest to the task must have appropriate data engineering and data science backgrounds. Also, there should be adequate tools for analysis. However, that means the organization procuring the tools should know which tools to acquire. The evaluation process and then the subsequent training of the individuals take time and money.
This process is not much different from any other technology that was imported in the past. However, the big data industry is rapidly evolving cloud-based services that reduce these lead times. By offering the analytics infrastructure in the cloud, the need no longer exists to evaluate, acquire, and install hardware and software. At a fairly low cost, the infrastructure can be configured in the cloud. The service provider also shows up with data scientists and data engineers who build a library of analytics and customize the library to individual needs. Often, these services can be configured in days and start producing results.
There are different types of service providers. Classic cloud providers, such as Amazon Web Service (AWS) or Softlayer, offer hardware and software assets in the cloud without any preconfigured analytics. Google analytics, Coremetrics, and Adobe are examples of analytics as a service on web traffic. A number of telcos, cable, and satellite operators are offering data as a service in which they provide data about their customers to their business partners. Cloud-based analytics companies offer social media sentiment analytics. These analytics providers bridge an important gap in breaking the ice for big data analytics. In a survey last year, I found over 30 independent programs across a large telco using a number of external, cloud-based sources. These programs provide a much-needed initial experience in using big data analytics, and as long as they are properly contained, the experience gained is invaluable.
While it is relatively easy to understand infrastructure as a service, data or analytics as a service is harder to organize. In offering data to third parties, the supplier must either aggregate the data across customers or obtain permissions to share the data. While it is easy to obtain permission against a discount, very often customers do not recall giving permission, which is often buried in pages of contract terms. Global brands are cautious in sharing data with others, as a backlash from customers can lead to irreparable damage to the brand and their mainstream business. Often, aggregation of customers is the safest approach to data sharing. Instead of providing raw data in aggregate form, the providers are also attempting analytics using the data.
Most of these are examples of analytics in which the original data was already external to a corporation. A proper analytics requires combining external data with internal data. Can we ship sensitive data to the cloud to be combined with external data by a third-party analytics as a service company? There are serious customer privacy and regulatory issues relating to data sharing across organizational boundaries. However, customer attitudes are rapidly changing with regard to the legitimate correlation of customer data, which benefits the customer. At the time of writing this book, there are two serious market experiments going on. First, the start-up activity using venture capital is experimenting with the subscription-based collection of customer data. These sources are independently collecting customer order, mobility, or other information with the premise of sharing that data at the aggregate level with others. However, many large providers are running pilot programs to share their data either with explicit customer opt-in or blanket opt-in, and with opt-out for those who are not interested in sharing. Analytics as a service will evolve over time as marketers begin to share their queries with each other and devise ways to share their data.