BI for Executives: Applying Business Intelligence

​In this article, we'll use Onyx Reporting's workshop, "The Strategist's Journey" as a framework for addressing the question "What can BI do for me?"

Last week I received an email from an IT director and senior financial controller asking, "How can I convince the executive suite to invest in a BI project?"  The email went on to hint that there were limited resources available to allocate to new initiatives. 

Ironically, when analysts describe why BI projects fail, they never mention the not-so-insignificant challenge of convincing management to green-light new BI initiatives.  If my experiences are any indicator, the challenge stems from technologists using technologists' language to champion projects to business people who evaluate based on business merit. ​

Overcome the Language Barrier

When the executives ask  "What value can business intelligence bring to my company?" technological benefits or features--applications, infrastructure, and tools--seldom (read never) translate well.

​We Want You to Save Money Too.
The subtext of the email from the IT director implied the opportunity for a critical paradigm shift in the executive suite's understanding of business intelligence.

1) Business Intelligence is not an IT/IS project.
Business Intelligence includes reporting and analytics that support measuring and controlling progress toward achieving strategic initiatives.

The budget roadblock suggests an underlying motivation to increase profitability by cutting costs and expenditures. The only reasonable way forward is to define and create structured and automated ways for monitoring, measuring and controlling costs.  Hint: that's business intelligence.

I recently had a conversation with a senior controller at a national car lot who lamented (and gleefully proclaimed) that her company could not keep track of where money was going, but "I will find it all!" How much easier and more effective would the controller be at their job with a robust analytics system augmenting their efforts?

2) BI develops in stages and looks different for every organization.
To control risk and address budget concerns, organizations frequently instantiate the first BI initiative as a small project then iteratively add features.

• Phase 1: describe the environment with historical reporting and analytics (ex. report year over year balances; net change of G/L accounts; and min/max/avg. transaction volume and size per account)
• Phase 2: diagnose and explore data with drill-down reporting (detail reports into invoices, expenses, and individual transactions)
• Phase 3: create a model that predicts 'normal' account activity
• Phase 4: prescribe action when the automated system flags irregular transactions.

If you're thinking, "That's just fraud detection at my credit card company." Yes, fraud detection is a sophisticated implementation of business intelligence.  
​
The easiest way technologists can overcome the language barrier when communicating the value of BI to an organization is to think like an executive!  ​

The Journey of 1000 Miles Begins

Instead of selling or even striving to define BI, Onyx collaborates with management to define what and how strategic challenges can be quantified and optimized.

• What are your BHAGs (big hairy audacious goals), and how do you measure progress toward achievement?
• What are your organization's core value-added business processes and which factors (either positively or negatively) affect their efficiency?
• What are key forecasting or budgeting exercises across departments?

When working with local clients, we'll break for lunch at my favorite pie shop in Bristol; because the excellent pie coupled with lamentable service offers an illuminating case study on applied BI.  After ordering the first round of appetizers, I'll pose the following:

"Business intelligence facilitates strategic decision-making by regularly measuring key business processes and then automates responses or creates a means for analyzing the data. How could this pie shop improve customer service as guest count scales from hundreds to thousands to millions of customers per day?"

Step 1) Define Good Customer Service

• customer order received within 5 minutes of being seated.
• no more than 5-minute wait during the lunch rush.
• upsell alcohol with every meal.
• offer personal service.

Step 2) Identify data points required to measure or achieve success

• track milestones during a guest's visit (time of arrival, seated, order taken food delivered etc.)
• optimize food production based on historical demand
• identify successful drink and food pairing options
• add 'personal touch' based on a customer profile (favorite drink, birthdays etc).

​Before lunch has arrived, we've has napkin-ed how this mom & pop pie shop can grow into a global franchise with customer service excellence as the secret sauce to their success. 
​
One-liners like "BI solutions measure and answer data-driven problems," sound glib and non-illuminating, so when challenged to answer "What value can business intelligence bring my organization," BI champions may gain more traction by sidestepping the invitation to 'sell BI'  and instead, weave a vision of analytics rooted in delivering tangible business value.

Iterative Business Intelligence:  more fun than classic car restoration

Once executives catch a hint of the ways BI can be applied to addressing strategic challenges, it's like tapping the idea factory of business insights. "We want all the features!"  Because one project cannot deliver 'everything', Onyx recasts BI projects as an ongoing process that iteratively adds depth and breadth to the ever evolving analytics and reporting environment. 

This agile methodology is not unlike the cash-strapped, gearhead street racer who buys a '67 Ford Mustang for local commutes and slowly kits the beast with modern racing parts.  Yes, from day one (or within a few weeks) the solution does its primary job of getting from point A to point B, but each mini-project will swap-in high-performance parts or refurbished upholstery and the ride keeps getting faster and sexier.
​
By adopting an iterative development cycle, executives empower the entire organization to watch, engage and interact with the functional BI solution from day one.  Instead of disappearing into the basement for months at a time, by working in short sprints the project team can have more regular touches with the organization and react to changes in the business landscape or revisions in project requirements as new data realities come to light (see the side note on data realities).

Just like the classic car restoration project, the pie shop's BI solution will grow ever robust in bite-sized sprints.
​
Iterating across Depth
In the first iteration of BI, project strategists simply capture, 'what's going on in the organization, and what key factors are affecting the activity?'  The next phases integrate statistics to facilitate automated decision making.  

While many organizations may not require predictive or prescriptive analytics, as data volume scales, to prevent data overload and mitigate human error additional sophistication is inevitably required.  Imagine if a person had to review every credit card or bank transaction for fraud.  While dealing with the false positive fraud alerts can be a nuisance, banks that don't leverage automated systems are exposed!  

Which risks has your company exposed itself to by forgoing systematic control mechanisms?

​Iterating across Breadth
The savvy project manager prevents scope creep by pruning (or recasting) features that fail to align with executive goals.

During pie-shop discussions when questions invariably arise around capturing data to maximize wait-staff efficiency, the strategist redirects. Yes, we could expand the solution to optimize wait-staff scheduling; however, that feature, while interesting from a cost reduction standpoint, does not actually address the project's stated objective--improving customer service.

Keep perspective and solve the right problems! During future strategizing sessions, the executive team may adopt new initiatives like cost savings, but the PM must retain support from the executive suite, by consistently delivering results that align with current strategic goals.


​Side Note on Data Realities: a technologist's solution to challenging business data.Question:  
Do I need a separate reporting environment from my existing data sources (point of sales, customer loyalty app, CRM software, website, and purchasing systems) ?

Answer 1:  Yes, structured and sanitized BI environments promise consistency in reporting and analytics.  The primary pitfall of reporting against transactional data is that each department could calculate the same measure (ex. profit margin) a myriad of different ways.  Alternatively, an analyst may touch multiple systems to create a complete view of a customer as they travel down the sales pipeline.  
Because BI environments create controlled, sanitized and agreed upon ways of combining data from multiple systems, users always report 'the single version of the truth.'

Answer 2: Structured reporting environments allow data analysts to overcome shortcomings in data realities by applying consistent business logic.  Consider the pie shop's point of sales system which doesn't capture the number of guests served.   A clever data analyst may recommend substituting the count of entrees on a receipt; however, in reality, guests frequently forgo mains in favor of discounts on appetizers, desserts, and drinks during happy hour.  

BI environments accommodate for the gap between reality and reporting requirements.


BI as an Asset
In Coursera's Business Metrics for Data-Driven Companies, Daniel Egger includes data mining a customer loyalty app, in his 20-Item Checklist for Data-Driven Companies.
  
Arguably the data from a 20-capacity pie shop may be statistically insignificant; however, there are other advantages to implementing a loyalty app.  Armed with customer data, wait-staff can support the business initiative of improving customer service by offering personalized interactions like using the guest's name, recommending historically successful beverage pairings, or 'remembering' birthdays.  Additionally, the app may improve patronage from young millennials (consider the demographics of app-driven Uber). 
​
Although Amazon, Starbucks, and Netflix already leverage and integrate data analytics into their product offering, executives may conclude, "We don't need a solution THAT complex!" A smart BI champion will wholeheartedly (dis)agree! 

The answer is to iterate.  Remember, the first question to answer is "what are (or will be) your distinguishing competitive business processes?"  When data drives and supports the differentiation, the definition of 'reasonable feature request' rapidly expands.

Maintain Perspective

Although software is often the most tangible piece of the BI puzzle, when it's time to discuss data delivery, we have to retain perspective.  Remember that executives and line workers will have different interactions with the same data.

A Vision of Your Company on Souped up Analytics

The most successful software vendors are not salespeople; instead, they are crafters of visions. 

What could you do with the data?  Formatted the way you wanted it.  With interactive charts and graphs.  And animated integrations with mapping software.  And accessibility via phone, tablet or on a plane. 
​
By sticking to the extensive and impressive features list, sales representatives avoid addressing technologist issues like data wrangling and clean up.  Nor do vendors get into specifics about what metrics your company should analyze (consider the data realities side note).

​Data interactions take different shapes depending on your daily workflow.  While some users will gravitate toward the robust desktop applications, others will prefer simple stripped-back notifications integrated into the point of sale system, kitchen displays, or mobile devices.  If you don't have a requirements document, shopping for software can be like walking doe-eyed into a car dealership looking for a family van and leaving with a tricked out two-seater.  Maintain perspective.  

This framework enables executives, strategists, and technologists in assessing how business intelligence can be applied to the organization to support strategic initiatives.

1. Use conversations between strategists and the executive suite to outline strategic objectives and core competitive processes.
2. Define measures and controls to track the achievement of those objectives.
3. Because project implies a time-bound effort, Onyx intentionally advocates the vocabulary of 'process' to imply BI that is an iterative effort that evolves as the organization achieves old goals and adopts new ones.
4. Always deliver value by keeping the current  core objectives as well as end user requirements in sight.

The question is never: "what data problems can I solve with BI?" because the glib answer is "anything."  Instead, we reframe the conversation around  "what do you want to achieve?"  From there solving the 'how' is a piece of cake (ahem ... pie).
This article was co-written by Jae Wilson, a lead business intelligence consultant at Onyx Reporting, and Joel Conarton from executive and management consulting firm Catalystis. 

For more articles like this delivered straight to your inbox, please sign up for our bi-monthly mailing list.