7 keys to a successful business intelligence strategy

BI success requires more than just a strong technology platform. It takes laser focus on processes and personnel — and a business-first approach to gaining insights from data.

Business intelligence (BI) is essential for business growth and competitive advantage, yet reaping benefits from BI requires more than implementing the technology that enables it.

In fact, deploying the technology is the easiest part of any BI initiative, according to Boris Evelson, vice president and principal analyst at Forrester Research. Getting the personnel and processes portions right are much more challenging, he says.

As such, organizations must addresses personnel and processes as key facets of their BI strategy if they want to be successful. Moreover, BI strategies should be broken down even further to address ownership and continual improvement as well.

Following are seven essential components of any successful BI strategy, according to several BI experts.

1. Give business ownership over BI
Organizations that place BI in the hands of business users have greater success rates than those who confine BI within IT, Evelson says. This may mean embedding BI within lines of business or having BI operations report to the chief digital officer or chief customer officer.

“The business must absolutely be in charge,” he adds.

Although the complexities of early BI technologies put IT in charge of many BI programs, today’s tools are more intuitive, allowing them to go straight into the hands of business users who can run the queries that matter to them.
Similarly, the speed at which users need access to data and insights derived from BI has increased dramatically in recent years. Today’s business users often need actionable information in real time and cannot wait for IT to generate reports. As such, IT ownership can be an impediment, rather than enabler, of BI success, Evelson says.

2. Monitor BI use and adjust as necessary
Although the business should own BI initiatives, IT must remain an active partner in monitoring and evaluating use of BI systems.

As Evelson explains: “Rather than putting up roadblocks, monitor what they’re doing, what data sources they’re accessing, what tools they are using and how they are using them, whether business unit A is using BI more than business unit B.”

In this way, he says, the CIO can set thresholds in partnership with business units. For instance, the CIO will know whether a few analysts in marketing have downloaded their own tool and are successfully using it, in which case it may be fine to leave them alone. Likewise, the CIO will notice when that BI application has seen an increasing number of users across business and has thus become an enterprise-grade environment and a mission-critical enterprise app that requires additional discipline and governance.

3. Validate, validate, validate
Organizations may be tempted to quickly spin out lots of BI capabilities, but quality outweighs quantity, says Chris Hagans, vice president of operations for WCI Consulting, a consultancy focused on BI.

“It’s better to have fewer things you trust than have a whole lot of things that are suspect,” he says.

As a result, organizations need a strong validation process that focuses on enabling access to all the data needed to answer queries. It should also prevent problematic data from entering the BI system so that it doesn’t produce faulty insights. In addition, the validation process should be agile enough to respond quickly to requests for new BI functions.

Hagans points to a hypothetical use case in which a BI tool generates reports on net sales figures. If that tool takes in data on sales but doesn’t figure in the number of sold items that are returned, then the end information is no good.
Moreover, Hagans says validation remains critical not just to ensure accuracy but also to head off skeptics.
“It only takes one or two people saying, ‘I don’t trust the data,’ to invalidate a report. That can tank a whole project, and then reports just become worthless,” he says.

4. Focus on business problems first, then on data
Don’t take a build-it-and-they-will-come approach to BI initiatives, Evelson warns. Too many organizations build data repositories, lay BI on top and then expect business users to jump right in and play, he says.

“What works much better is a top-down approach, one that’s about business outcomes. We don’t start with ‘Where’s the data?’ We start with solving a business problem,” he says.

Evelson lays out this example: Marketing spots a customer churn problem and wants to understand why customers are leaving. The organization should focus on delivering the capability to answer marketing’s business question by first deciding what metrics need to be measured, accessing the data needed to calculate those metrics, and then enabling marketing to slice and dice the data.

“We need to identify a clear business problem first and what metrics we want to analyze, and then at the end of that we talk about where to get the data,” Evelson says.

5. Prioritize — and build in processes for improvement
A successful BI strategy anticipates both expansion and improvements, according to BI leaders. As such, organizations should know what business insights they want and which ones are most important so IT can deliver what’s most critical to business users first and work its way through a priority list. Moreover, the BI program should be able to shift as the priorities change.

“It has to evolve with what the users and the people inside the business community need,” Hagans says.
Similarly, the BI strategy should build in processes to advance and improve how the system works. Evelson recommends an iterative approach, so that the BI tool can expand and improve as business units use it and determine where it meets their needs and where it doesn’t.

6. Upskill ‘citizen’ data scientists
In its 2017 “Magic Quadrant for Business Intelligence and Analytics Platforms” report, research firm Gartner says “the number of citizen data scientists will grow five times faster than the number of data scientists” over the next several years.

Cindi Howson, a research vice president at Gartner, says executives already recognize that there aren’t enough data scientists to meet demand; they’re also struggling to hire or identify in their existing ranks the citizen data scientists they’ll need.

“We’re talking about the in-between information analysts. They know the business domain and the questions to ask,” she says, adding that there’s a need for software that’s easier to use so organizations can better enable these workers.
Howson anticipates that software improvements will eventually allow businesspeople to ask and answer their own questions on unmodeled data sets. As that happens, organizations will need the right people to take on this role of citizen data scientist. They’ll need inquisitive workers with analytical skills who like to ask questions, know how to interpret the information they’re getting back, and are comfortable using software to improve business outcomes.

7. Empower staff to tell stories with data
On a similar note, Todd Nash, president and principal of CBIG Consulting, a professional services firm that helps clients leverage their data assets, says he has worked with organizations where workers understand how to use the insights offered by their BI tools to tell stories that help others understand “what the data is trying to say.”

He says these people use the reporting and visualization functions built into BI technologies to develop narratives that help maximize the value of analytics.

“You have the data and the tools to tell the story, and you need people to marry those up,” he says. This approach is not just about having people who produce slick-looking reports; Nash says these users are able to make connections with the data that others might not see, thereby offering new insights that businesses can leverage for gains.

He says executives need to support and enable these workers as they explore those connections and present their insights.

For example, he says workers analyzing store sales figures might see how smaller weather trends — not just big storm events — have subtle impacts on sales. They might want to draw in external weather data to further analyze trends to better understand how the stores can optimize sales given this new insight.

“There’s all kind of internal and external data to take advantage of to get much better insights,” Nash says, adding that successful BI programs allow analysts to move beyond measuring standard key performance indicators.

“There are just lots of different ways to challenge yourself,” he says, “and part of that is challenging every KPI and making sure you’re taking advantage of the information available to you to understand.”

Read original article: Mary Pratt, CIO.com