Self service, business led BI is big and getting bigger. Gartner predicts that market will be worth billions and the number of vendors offering these products will constantly grow and expand. Gartner’s report correctly points out that we are in state of more data from more sources; where enterprises want more users enabled to explore, search and discover insights. At the same time, enterprises have more data needs to be converted into business insights and strategic action.
However, this is easier said than done. Making BI easy, intuitive, self-service and ubiquitous in the enterprise is a complex problem. This complexity needs to be hidden away from users such that users are able to achieve their objectives without getting lost in the products. As Alan Cooper, describes in his book “The Inmates are running the Asylum”, modern BI products need to be user friendly and designed to work in the way the average user expects them to work.
This is where the modern suite of products struggles. These tools
- Fail to fully understand and address the requirements of the entire analytics lifecycle in a big enterprise. Too often, the focus of the product is on the sales demo or the creation of a dashboard and not on enabling the entire organization to be data driven.
- Fail to fully understand and address the significant drawbacks of having a completely ungoverned, unfettered access to data and a creation and publishing interface. Too many users creating content at the same time makes it harder to publish, access and find the good content and sideline the bad, low quality content.
- Fail to enable seamless real time or near real time action and reaction based on analysis of data leaving analysis and insights gathering dust in dashboards. The focus on dashboards and alerts takes the focus away from converting the insight into action and a change in how the enterprise and its system/processes/products act and behave.
- Fail to enable real collaboration between users of varying needs, skill sets, experiences and domain knowledge.
Without understanding and addressing the above issues, these products hit one or more of the following adoption roadblocks:
- Products can easily delivery simple use but struggle with complex use cases; where complexity is in either the variety of data, the analysis or the required speed of the analysis.
- Complex use cases require an expert level understanding of the product AND the data/analytics restricting the use of the product to a smaller group of advanced/power users.
- Unfettered access to creation/preparation of data, insights (as data sets or dashboards) leading to clutter, noise and reduced quality of service for all users. In addition, fragmented use of the product across the enterprise leads to analytics and insights in silos that are not reusable and often duplicate with the resulting insights being sub optimal due to missing context, data or expertise.
Gartner’s report touches upon the concept of bimodal IT where IT needs to enable two modes of delivery simultaneously focused on stability and agility and measures the BI vendors on enabling agility in enterprises. This is where the definition of a visionary BI product needs to be extended.
An analytically driven, agile enterprise is characterized by the following three traits
- A minimal time to insight
- A minimal time to action and reaction
- A governed access to creation and consumption of data, insights, alerts and actions.
The above are the traits shared by the most successful digital native enterprises such as Google and Facebook and this is what separates an analytically driven enterprise from the rest of the pack.
Enabling an enterprise to achieve these desired traits requires that BI vendors stop optimizing for the first run or the first year scenario. BI vendors need to understand and plan to enable thousands of users over multiple years, over hundreds of data sources to create, search, discover and consume content and subsequently act and react.
These vendors also need to prevent isolation and silos that get created through a “Shadow IT” approach to BI. BI vendors also need to be careful in how they paint and portray IT organizations and build capabilities and features in their products that leverage and enable IT to govern, manage and improve the experience of the self service users.
So what should BI vendors focus on i.e. what should be the criteria for visionary BI products and what should Gartner look for determining the real visionaries?
Self-service hypothesizing, experimentation and rudimentary analysis
Business users have the context, knowledge and ideas about their business domain. A good self service, business led BI product needs to enable these users to leverage their knowledge to hypothesize and experiment. By broadening the mouth of the funnel to generate more ideas by more users, the enterprise increases the chance of uncovering good ideas that can be properly groomed and converted into analytically sound insights.
Enable finding an answer if it already exists i.e. one less dashboard is always better in the long run
Self-service BI needs to enable business users to focus more on utilizing an insight in their workflow rather than creating the insight. This means that if a high quality insight already exists (because it had been created by someone else earlier) then the business user should be easily able to find it. If one does not exist but a similar one exists whose methodology is reusable, the business user should be able to find it and utilize it appropriately. A good self-service BI product should not end up making the process of discovering good analytical content a needle in the haystack exercise.
Don’t try to change the business users’ day job
Business users already have a day job and the problems that they typically deal with are complex with many moving parts. It is important that products either enable complex analytical scenarios that require a mix of intelligent profiling, self service data prep and analytics in their entirety or enable a strong collaborative model where business users can collaborate with expert/power users to satisfy their requirements.
Focus self-service enablement efforts on the long tail
If a typical enterprises’ analytical use case are charted according to complexity, it ends up as a long tail as the relative complexity reduces. Self service, business led analytics is ideal for dealing with this long tail and quickly identifying the analytical use cases that show great potential and implement these scenarios as high quality, production deployed, analytical pipelines that are high performing and integrated into business workflows.
Focus on standardization and governance of access, creation and discovery
BI vendors need to focus on how users of varying skills, experience, analytical mind set and comfort with the product are able to search, discover, access, create, consume, collaborate on and integrate into production workflows analytics and insights. At the same time while enabling many users to access the systems concurrently and create and consume content, vendors need to ensure that all access is governed and monitored for the highest levels of service and best user experience. Unsuitable workloads should be correctly deployed or isolated and segregated to ensure the best possible service levels for all users.
In addition, BI vendors need to provide APIs that expose a layer of metadata, which stores business logic and asset definitions in a non-proprietary format enabling a bidirectional flow of metadata across many different deployments of diverse BI products managed by central or shadow IT. This enables the removal of data silos in an organization and the storage of analytical IP in a common, consistent format enabling reuse and with a lifespan that extends the life of any single BI product.
Lastly, the products need to provide a platform and tooling that is adapted to the nature of queries in an enterprise. By standardizing on a small set of tools/products, enterprises can redirect users to the appropriate tool that is best suited for their purposes and analytical use cases. This prevents users from wasting time on tools that don’t and can’t deliver on their needs and reduces the over time to success by matching the user’s needs with the appropriately configured product or tool.
Reduce the time to action
Last of all, visionary BI products need to measure how well they reduce the time to action in an enterprise as evidenced by their ability to (at the same time) integrate many different sources of data, integrate with many different business workflow applications through both push and pull mechanisms, ability for data and analytical prep coupled with discovery of existing data/analytics and the ability to execute data prep, analytics and actions on streaming data. Reducing time to action is and should always be the barometer for success for any BI product; visionary or not.
This article is published as part of the IDG Contributor Network. Want to Join?