The Fairway Technologies Blog

blog

Getting Started with Business Intelligence

In BI, database, busines intelligence No Comments

Every executive, manager, or department head wants to understand how their decisions will improve their team’s performance, increase overall efficiency, and ultimately boost the company’s bottom line. While we can’t always predict what’s going to happen, we can use modern tooling to base our decisions around known facts—data—that the business has uncovered.

What is Business Intelligence?

Business intelligence (BI) is the use of analytics tooling to provide prescriptive and predictive insights, and answer overarching, high-level questions by aggregating data from within and across the different verticals of a business.

Phew. Let’s break that down a bit. To start, let’s look at two high-level questions:

  • Is the company healthy?
  • Are we making the right decision?

Before answering the first question, we need to define what makes a company healthy.  Some common indicators that people typically use are revenue, stock price, or other sales-driven metrics. While these are undoubtedly important, there are underlying health indicators that need to be considered and included as well.

Imagine a medical device company that just announced groundbreaking sales for last quarter—beating analyst estimates and sending their stock price soaring. Pretty healthy, right? Not quite. That market valuation doesn’t account for a number of things that could be going on under the hood: attrition in the R&D department or throughput bottlenecks caused by warehouse machine failure. The health of our medical device company relies on all aspects: operations, human resources, research, and sales. Each of these business units has differing information they rely on, whether they are reports emailed out daily or weekly status updates, but how do they all relate to one another? This is where BI fits in.

For the second question, let’s pivot and look at the world of online marketing: conversion and adoption rates, SEO, focus groups, A/B and early access testing. In many cases, the decisions made by a marketing team can change the brand or identity of the company. These decisions aren’t based on gut reaction, instinct, or premonition; they’re based on data. Using the example of rebranding an entire company is a bit extreme, but, given the opportunity, wouldn’t you want to make a decision knowing that it was the right one? Knock, knock—it’s business intelligence.

So…It’s a Fancy Phrase for Reporting and Dashboards?

Not exactly. Solid BI platforms will include reporting and dashboards to make it easy to digest the insights they create, but, when implemented properly, BI encourages individuals to adopt a different mindset than what was offered by reporting and dashboards from days past.

We’ve all experienced “reporting-by-Excel”—the weekly status reports sent out in a mass email, bound together by a single Excel workbook. No, you’re not alone.  Everybody, from Fortune 500 companies to small mom-and-pop shops, has been down that road. There has been a wave of products over the past decade that has helped with the sanitization, automation, and visualization of these reports, but at the end of the day, most products still offer the same insights that the Excel workbook does.

The problem with this structure, and manual reporting in general, is that it’s error prone, unscalable, and results in data that is “pushed” whether you need it or not, rather than “pulled” when you do need it. Data that is published through a rigid reporting structure doesn’t answer questions that have yet to be asked. BI provides the ability to pull data when you need it, enabling you to answer the subsequent questions that arise from your reports and dashboards.

 Where Do I Get Started?

1. Know the Space
Do your research on what each product offers, how extensible it is, and whether it fits your budget. It’s worth spending time understanding whether or not the product will fit all of your needs because migrating between different products can be a headache. Some key areas that differ between the leading products are: 
  • Supported Data Sources - SQL Server, Oracle DB, static files
  • Report/Dashboard Sharing - across the organization or between individuals
  • Types of Visuals Available - static out-of-the-box vs. custom built
  • Level of Data Transformation – ETL (Extract, Transform, Load) capabilities or just formatting/style
  • Customization - white-labeled products, APIs, embedding
  • Security and Deployment - cloud-first, on-premise, hybrid, user and group-based security

Microsoft Power BI, Tableau, and Qlik are leaders in this space, but there are dozens of other solutions that may fit your needs.

2. Define Your Boundaries & Identify the Owners
Similar to how an org chart shows the breakdown of business units and hierarchy between people, it’s important to understand and identify the various data silos that are spread throughout the organization. This exercise makes it easier to see how a BI reporting infrastructure should be laid out, and determines which individual, or group of individuals, will be held responsible for the data or reports that are created.  For each data silo, answer these questions and keep stakeholders involved in each step of the process (data discovery, report creation, implementation, and rollout):

  • Who owns the data that you’re working with?
    (i.e. Who should you notify when data needs to be fixed or updated?)
  • Who is the subject matter expert or individual who can explain the business rules that need to be applied?
  •  Who is responsible for maintaining the BI infrastructure for this data?
  • (i.e. Who is in charge of updating the key metric definitions when they change?)
  • Who is responsible for the insights that come from this data?
    (i.e. Who will this data impact?)

3. Practice Iterative Development
Treat BI infrastructure rollout like software development; it should be iterative.  Gather requirements, deliver often, gather feedback, adjust delivery, and repeat. BI infrastructure is not something that’s created once—it’s a constantly evolving analytics assistant that will grow and adapt just as much as the business itself.

4. Focus on Adoption
Previous styles of reporting fell into two camps: user-driven and IT-driven. User-driven reports were chaotic, messy, and there was no way of telling what version had the truth. IT-driven reporting was equally as cumbersome, as it required users to submit requests to IT to dredge up the data and stick it in a report in a way that made sense to IT, not the business. 

The BI space has tried to solve this problem by making it easy for users to create their own reports, with very little involvement from IT. This doesn’t mean that IT shouldn’t be involved (see the “Common Misconceptions” section below).

You’re changing the way people work, and rather than trying to overcome the inevitable resistance to change, get them involved early in the process. Identify people who understand the data or domain knowledge and use their expertise to drive the report and dashboard construction. They’ll be excited (rather than pushing back against something new), the reports will make sense to the business, and IT will get to have their say in any necessary governance.

 Common Misconceptions

1. You need a data warehouse. Not yet.
Start small. Let each business unit answer their own questions by building reports and dashboards specific to the KPIs they rely on. It’s okay to take the data that’s in a silo and start working with it. If you start by trying to join everything together, you’ll quickly be overwhelmed.

2. You need machine learning. Nope.
Trust me—there are unanswered questions floating around your office that can be answered with the data you already have. Once you’ve got your foundation laid, then you can start worrying about predicting the future

3. You need to be a data analyst to work with the tools. No, but …
Some tools do require a level of skill, but these skills can be learned. Most of the industry is catering toward the non-developer user, making it easier and easier to dive through data

4. Business intelligence can be owned completely by the business (not IT). Don’t believe the hype!
It’s a common notion that “BI breaks the mold” and isn’t driven by IT. While Microsoft, Tableau, and other Business Intelligence players would like to make you think this, it’s still important to have IT involved in the conversation.  Licensing, provisioning, and infrastructure needs will likely fall on IT’s plate, and you want to have them involved in the governance aspect.


5.The ROI isn’t there. Well …
The ROI is there and can be measured in terms of efficiency and confidence.  The short time needed to spin up a base-level BI infrastructure empowers leaders to make decisions more efficiently, and enables insights that may be too cumbersome or complicated to compile manually on your existing reporting infrastructure. Similarly, the ability to easily adapt and adjust your BI infrastructure in response to the ever-changing needs of the business will save countless hours and dollars in the end.

 

Summary

Business intelligence can be transformative to the way an organization operates, but it also can be intimidating to get started. There are intermediate steps and milestones that will, not only help you build a solid foundation, but also help adoption of this new tooling grow organically across the organization. One of the biggest mistakes is implementing a “game-changer,” and not guiding end users’ understanding on how or why the tools and processes are being implemented.

Now, it’s time to take the first step. I hope that this gives you the information you need to start adopting a business intelligence, and more importantly, a data-driven mindset.

New Call-to-action
New Call-to-action