The technology world is constantly evolving. Streaming services have revamped the music and movie industries, helping to usher in a new era of on-demand content. Mobile apps such as Uber and Lyft have transformed how we travel, while others—such as Airbnb—have changed how we rent accommodation.
This trend is called Anything as a Service (XaaS) and can include products, tools and technologies. As more and more companies move to the cloud, they start providing increasingly sophisticated IaaS, PaaS and SaaS solutions, drastically reducing maintenance efforts. As a result, new concepts such as Machine Learning as a Service, Security as a Service, even Games as a Service and Data as a Service are emerging.
What is Data as a Service?
Data as a service, or DaaS, refers to exchanging machine-readable data for something of value between two or more organisations, according to Wikipedia. However, this definition is not the only one and what constitutes 'DaaS' varies depending on interpretation.
Some interpretations do not restrict the data to being machine-readable only. Others omit the value exchange part to incorporate organisations that provide access to data for other reasons rather than profit. Many interpretations refer to cloud computing.
However, the general idea behind all these varying interpretations is that one party provides data to another.
Classifying information services by type
There is not always a clear distinction between the terminology used to describe multitude of data services on the web. For this discussion, let us consider a split similar to the DIKW model.
Data as a Service
It is ground zero - raw data that still requires the acquisition and a great deal of effort to use (typically, cleansing, enrichment, transformation). It is not consumer-friendly. Sometimes it's big data that cannot be used without a specific set of technologies that can deal with its sheer volume or velocity—for example, satellite photos and vehicle traffic.
Information as a Service
One level above DaaS. The information provided could be used as is and may or may not require additional data acquisition or processing work.
One method of distributing this information is via websites. In this case, data resides solely on the DaaS provider side. A website offering business information about companies, like Crunchbase, could be considered an Information as a Service platform.
Analytics as a Service
The next level includes visuals such as maps, gauges and pie charts, and means of self-service data analysis, including aggregation, filtering and transformation among others. Data acquisition is not required.
Insight as a Service
The top level is about providing predictions, prescriptions, what-if analysis and other insights, more often than not powered by machine learning and deep domain knowledge.
It's worth mentioning that vendors often offer a package deal which combines the services mentioned above. For example, the European gas market analysis platform appygas provides a map of gas flows, planned route maintenance information (falling under Analytics as a Service), and a route calculator for what-if analysis (falling under Insight as a Service).
The problem with the above categorisation is that the two last concepts are often used to describe software-based offerings. To avoid complications, we will use DaaS as a generic term that covers all these concepts, yet keep in mind the variety of different service forms.
Unlocking the value behind the DaaS platform
Companies may turn to DaaS for several reasons, including if acquiring the data is their only option or if it is easier than creating it themselves.
The pitfalls of data collection & management on your own
Before moving to the advantages of a DaaS, let's look deeper into the problems with collecting or managing data on your own to see why the DaaS market appeared.
- Increased data cost. The cost of gathering, creating and processing data internally with development resources and infrastructure is more expensive than one might initially think. The price is not flat or transparent due to the many hidden or unexpected expenses involved. Let's say we need to extract a few datasets from a website—not the most developer-friendly data source. The data-acquiring process may be complex, including multiple interactions with the website that may be required to display the desired data. Maintaining a web scraping solution is even more challenging and expensive due to occasional changes on the scraped website. Moreover, after obtaining the information, you need to clean, transform and store it in a way suitable for analysis. Then you'll need to design business intelligence dashboards. Now imagine that instead of doing all this work yourself, you could simply receive ready-to-use dashboards and reports.
- Longer time to deliver insight. Not only does the example above cost more to do on your own, but it also takes considerably longer to implement.
- Data governance and quality. Maintaining data quality requires effort. Typically, the best solution to this is a data stewardship process. It relies on dedicated data stewards who spend significant time ensuring data platform assets are clean, up-to-date, available, documented, secured, and conform to the company's data standards.
- Ensuring compliance and enforcing security is pivotal. Storing sensitive data is an integral part of the data management strategy and should not be treated lightly, which of course, comes at a price.
- Documentation, ease of use and availability of business data are just some of the many things that need to be considered.
- And finally, as mentioned above, there are cases when collecting data on your own is not feasible due to the enormous investment needed. Weather data (forecasts or historical) or satellite images are prime examples. Buying them is easy, and while not cheap, there are multiple options to choose from, but making satellite images requires, well, a satellite.
Benefits of the DaaS platform
The advantages of DaaS below relate to the issues previously mentioned:
- Reduced costs, both direct (data acquisition, data storage, data integration, etc.) and indirect (maintaining data quality etc.)
- Reduced time needed for users to start getting the value out of data (sometimes it's close to zero and boils down to buying a license and logging into a web portal)
- Easier to enable data-driven decision-making and build a data culture if the information is accessible via a click or two
- Simplified compliance (no need to care about the physical location of the data assets, retention periods, and so on)
- Better and more transparent budgeting
DaaS platforms typically remove the pains of data management listed. Although they can still be prone to problems like data quality and availability, in most cases, you have a strict SLA defined by a service provider that will help protect you from these risks.
Related Insights
