Ryan Byrd's Tech Ramblings

Business Intelligence? More like Business Misinformation…

As CFO for a Salt Lake City eCommerce company (annual revenue 17 million), I am frequently asked about how we make and lose money, where we can save dollars and how we can increase profitability. Stakeholders want timely, actionable business intelligence (BI).

Based on my experience, stunningly enough, most companies (from large to small) have highly anemic BI systems. But BI is one of the most important business tools; effective biz-intel gives instant feedback to management about how the company is performing. Traditionally, we rely heavily on financial statements like the cash flow statement and the balance sheet and ad hoc reporting to provide business performance insight. These traditional methods are rendered less effective because of the time delay of financial statements and one-off report preparation. Financial statements are completed typically 1-2 weeks after a month is over. In many industries, that information is too stale for good decisions, and the feedback is too late for agile AB testing of business initiatives. In many cases, untimely intelligence is worse than no intelligence at all.

To better understand business reporting, I’ve broken it into three sections:

1. Historic: Historical reporting is financial statements that are prepared by the accounting department. These are typically used for external entities like banking and investors.
2. Current: Current reporting is much more effective for day to day business management. This is BI reports and managerial accounting.
3. Future: Future reporting is projections and forecasting. This is for planning, acquisitions, and obtaining debt and equity financing.

Each of these categories is crucial for business success, but the #2 (current) is the one most neglected and yet most important for day-to-day business operations. Fast and accurate current reporting is a proven competitive advantage. Effective current reporting makes businesses more adaptable and nimble in the ever changing business environment. Many companies have initiative to create business intelligence, but most have been unsuccessful, wasting large sums of money. A quick brainstorm netted these 5 major obstacles to effective BI:

1. PLANNING: BI needs to be a well thought-out plan that doesn’t necessarily require hundreds of thousands of dollars purchasing software and servers. It always seems like companies are just throwing money at the situation and still cannot create effective BI.
2. KNOWLEDGE: BI is created by technical types. Techies are the ones that know how to create advanced queries and display the custom data, but techies are not usually trained on creating usable and intuitive reports. This was the biggest issue with my company. Techies created BI reports but they were very misleading (and often inaccurate). So in order for BI to be effective you need the fusion of technical know-how combined with skill in financial and business reporting.
3. ACCESS: Real-time access to the underlining data is extremely important to creating effective BI. The I.T. department wants to protect its data for stability, and the finance department wants to protect its reports for privacy. Of course the data is confidential, but there are ways to summarize data and not provide all the details.
4. DISPLAY: It is very important to display the data in an effective format. If all you display is a line graph of daily sales for the last month, that really doesn’t say much. Sales should be compared to the past – recent past and distant past – to give a comparison.
5. SPEED: When creating a single dashboard that summarizes millions even billions of database records, speed becomes a major obstacle. On my next blog post I’ll talk about how I overcame report slowness.

STAY TUNED:

In the next installment, “Zero to DIY BI in 30 days”, I’ll explain how I surmounted these obstacles, and how we went from bad data to effective BI in 30 days. We’ll talk about integrations, real-time data, data warehousing, KPIs, and iterative development.

Matt Budd, CFO
With contributions by Ryan Byrd, CTO

picture credits: Confusion by Adi Ron (2005)

Because tech companies seem to have a high rate of turnover, you’ll soon likely be in the position of leaving or interviewing an employee who is leaving. Either way, it’s important to find answers to some important questions such as:

• If the CEO left unexpectedly today and you were put in charge, what are the first things you would change?
• What could have changed six months ago that would have prevented you from looking for a new job?
• If you weren’t looking, what factors tipped the scale when an opportunity came up?
• Who do you think is next to resign and why?
• If one person leaving the firm would cause you to think twice about leaving, who would that person be?
• Why didn’t you leave us sooner than now?
• How did your manager communicate your responsibilities? DO you think he or she was fair and reasonable?
• Describe any areas of conflict that have affected either your performance or morale, or that you believe affected other employees

src: Inc. Magazine, April 2006

• Does the business idea lend itself to venture capital investment?
  1. is it the next big thing?

  2. can it produce 5x-20x return on an investment of 5 million in 3-5 years?
  3. does it have some sort of infringement protection (network effect, patent, big head start)?
• Are there tiered service levels (e.g. free, basic, pro)?
• How many paying subscribers at what monthly rate do you predict in 1,6,12,18 months? (how did you arrive at your prediction?)
• Are there possibilities for affiliate relationships?
• What variable and fixed costs will you incur?
• How will you advertise and drive traffic to your site?
• How will you handle customer support?
• Is there any inventory to manage?
• How much staff is required? Where will you find the best people?

Start with a SWOT analysis

Then, find answers to these questions:

• Are there competitors?
• Does the service satisfy an important need?
• Is it a one time/infrequent purchase or would it be on a subscription model?
• What is the estimated profit margin?
• Does it depend on a lot of people resources to run it?
• How easily would it scale?
• How easy would it be for competitors to copy?
• Does it lend itself to patent/copyright protection?
• What is the market size?
• Are there substitute services?
• Does the service lend itself to the “network effect?”