BI at an Inflection Point (Part 2 of 2)
Macro trends illustrate how BI is headed for an inflection point. What does this mean for BI professionals?
By Charles Caldwell, senior director, solutions engineering, Logi Analytics
In Part 1 of this discussion, I examined some of the larger trends driving us to a BI inflection point, including
- The information economy is disrupting the way we do business
- Society has become more data savvy and we increasingly believe we should be able to measure and analyze just about anything
- Everyone has computing power and is connected, which means everyone is now a potential audience for information and analytics
- The growing need for highly social collaboration and knowledge networks that support knowledge emerging, rather than fitting into predefined taxonomies
In this article I'll explore what these trends mean for BI. How do these trends put BI in a position to change?
BI to date has largely focused on:
- People inside the four walls of the enterprise
- C-level executives, analysts, and (to some degree) managers
- Providing a single app that unifies all the information in an organization
- Supporting experts in making decisions and pushing those decisions out to an enterprise
- Presenting a single version of the truth
That focus has served us well through the turn of the last century, and solved many of problems of the day. However, that success has, in fact, contributed significantly to the macro trends above. The current tools are optimized for a set of problems and a context that is increasingly low value and soon will no longer exist. The tools, architectures and capabilities need to evolve.
Here are some of the high-level implications.
BI, Meet UI/UX
As an industry, we are data people. We extract, transform, load, model, optimize, cleanse, enrich, and fret over data. The user interface is usually a dashboard, a PivotTable, or a tabular report. That's all you need, right?
If your target audience is 10 to 20 percent of an enterprise and largely consists of executives, managers, and analysts, then yes, that's all you need. You train your users to work with those interfaces, and they do enough data crunching for the investment to pay off. However, when considerable value exists in providing analytics in the context of everyday decisions front-line workers make, the user interface (UI) and user experience (UX) matter.
This is the big lesson from the consumerization of analytics. The future of BI is on your iPhone, not one monolithic application that answers all questions. Apps are focused on accomplishing one task and doing it extremely well. They are very often data driven, support decision making, and don't sacrifice usability because they need to do everything. Book a flight, pick a restaurant, find a place to live, find a ride across town; apps enable their users to do exactly what they need to do with no half measures or usability short-cuts. The average smartphone user works with 27 apps per month; users love apps and are highly productive. Pervasive BI looks less like a dashboard or pivot-table of all data and a lot more like targeted apps focused on taking action. And, to get there, UI/UX trumps the data model.
Analytics for Mere Mortals
If data discovery has taught us anything, it has taught us that more people need more capabilities than we originally thought. It isn't just "self-service reporting" or "drill-down." They need to be able to acquire data, enrich data, reshape data, perform discovery tasks, perform data quality tasks, publish reports, collaborate, and perform advanced analytics -- and they need it now.
Visual analytics did a great job of getting from prepped data set to cool dashboard, but what about all the data wrangling? You still have to call IT for that. What about advanced analytics? You still have to call a data scientist for that. What about monitoring sanctioned data sets and publishing those to other users? Call the BI competency center!
I'm not trying to get rid of IT, BI professionals, or data scientists, but as analytics becomes more pervasive, these types of capabilities must be provided in packages that enable business users to accomplish a baseline set of capabilities without help from IT. That baseline can't be "export to csv, import into data discovery tool." Business users need to mash up cloud application data with on-premise data, prepare it for analysis, and share it with other users. They need methods for applying advanced modeling techniques that guide them and help them understand the implications of the output and when they should call a data scientist. Without these capabilities, pervasive BI is not possible.
Analytics Goes Social
BI tools follow a very traditional model of thinking: predefine the single version of knowledge and distribute it. Dimensions and facts define the bounds of your thinking. Folders, with one report per folder, limit your ability to search and share. Even data discovery follows a model in which an expert develops a report and publishes it for consumption.
So far, the concept of social BI involves being able to comment on reports based on predefined data and views filed in those predefined folders, maybe with a little desktop sharing added in so that two or three experts can develop the report before they publish it. That model is fine for some things but simply doesn't keep up when the business landscape is fluid.
Social concepts are powerful for facilitating dynamic, agile collaboration and knowledge creation. Don't wait to assemble a dashboard; publish everything into a stream constantly so the entire group can access it in real time. Use concepts such as voting, favorites, or likes to determine the relevance of the content. Enable comments and mentions to facilitate conversations. Use tagging to ensure that as new ideas emerge, they don't get lost in an old folder structure. If BI is going to support dynamic decision making in fluid business environments, it has to enable teams to dynamically form, collaborate on problems, and be able to redefine what matters to the business.
BI Outside the Four Walls
BI and analytic applications have a huge impact when targeted to customers, vendors, field employees. and others "outside the four walls." BI apps are becoming revenue generators in and of themselves in some cases. In many business models, value creation is so dependent on entities outside your own organization that analytics targeted to those entities is critical to operations.
BI architectures have not been friendly to broad-scale, external-facing deployments. The architectures are built to lock customers in from the data tier up through the presentation tier, and typically include heavy legacy client-server components. The temporary "fix" has been "mobile" components or Web front-ends plastered on top of those architectures.
External-facing BI requires a completely new architecture that supports modern Web- and mobile-application architecture patterns. A data tier is great, but let me bring my own as well. The user interface must be flexible to meet targeted use cases, including embedding analytic components into other Web applications, and they must have scaling models that IT operations can work with and the CFO will pay for.
I, Data-Centric No More
The data problem is an important one that we continue to struggle with, but BI has been so data centric that the tools and architectures the industry has provided are showing their limitations. As enterprises demand tools to meet emerging challenges, I see the BI industry heading into an inflection point.
The "data is the only problem" approach we've taken in the past must give way to recognize the importance of UI/UX, increasingly enabling the mere-mortal business user, promoting vibrant collaboration and social capabilities, and reaching everyone with targeted analytic capabilities that fit their needs. These observations are shaping the way we work and develop our offerings, and all involved in the BI industry should consider these trends moving into 2015.
Charles Caldwell is the director of solutions engineering and principal solutions architect for Logi Analytics and has a decade of experience in data warehousing and BI. He has built data warehouses and reporting systems for Fortune 500 organizations such as Unilever and American Express, as well as enterprises in the pharmaceutical, manufacturing, financial services, and public sectors. He completed his MBA at George Washington with a focus on the decision sciences and has spoken at industry conferences on topics including advanced analytics and agile BI. You can contact the author at [email protected].