Data Analytics: You don’t have to be big to do Big Data

Some leaders of small- and medium-sized businesses (SMBs) may think that they don’t need advanced analytics: they are doing just fine, thank you, with reports and basic dashboards.  Besides, they don’t have the resources to leverage data, or don’t need analytics.

From my experience, they should reconsider:  reports only look in the rearview mirror.  Modern, forward-looking tools and techniques are more accessible, powerful and affordable than ever.  Businesses are to the point that they can’t afford NOT to do data analytics – especially because their competitors probably are.

If you’re not sure how to get more out of your data, and then how to reap the benefits, this series of blogs may be helpful to you.

Data Process GraphicContrary to the hype about “data science” producing “new insights” from “advanced analytics,” technology and data-driven discovery are the investment and the means; they are not the end results.  The real work begins when the business must change in some way to take advantage of the new discovery.

Viewed in its entirety, the effective use of analytics has five components:

  1. Data that are unique and valuable;
  2. Statistical and other technical analysis;
  3. A clear decision and strategy;
  4. Implementation of new practices or processes; and
  5. Measurement of the impact.

As the analytic program grows, it will need to adopt good practices for data management to ensure the results are repeatable and reliable, and can expand over time.

What data.   The first choice is where to focus: externally on the market, or internally on your operations.  Starting with internal data has at least two advantages:  a) your data are unique to your company and thus can be used for competitive advantage; and b) you understand your data better than you are likely to understand an external data set.  The most common challenges are to manage it properly and ensure its quality.  Initially, these tasks can and should be done without expensive data warehouses or a massive data infrastructure; it is more important to understand the potential value of data before investing in a field of data dreams.

Start with the end in mind: the goal or pain point you want to pursue.  Not sure you have any data that helps your goal?

  • If you want to expand your product line and/or increase sales, your financial system contains customer data that might reveal market segments for closer targeting.
  • If you want to hire better people or reduce attrition, your payroll system may contain hiring, promotion and termination data that indicate the profiles of successful employees.

If you want to relieve pain points in your operations, you’ll need data about your processes, such as their efficiency, quality and actual capacity (both in-use and total).  Many SMBs don’t measure their processes, but systematic collection and analysis of process data often identifies out-sized opportunities, especially after you’ve looked at everything else.  Start with the basics, which can be measured non-disruptively by monitoring inputs (resource consumption), timing (applied time vs elapsed time), and outputs (quantity and quality).

What to do first and second. An “explore first, manage second” approach allows new data users to identify the potential value of becoming data driven, which is essential to determining how much to invest in data science, data warehousing and data governance.  Good exploration is multi-disciplinary, integrating knowledge of the business model, the organization, statistics and data.  It also requires good habits:  if you want to trust the results, you must be able to reproduce them.  Here’s where data management, data governance and documentation come in.  I’ll expound on these topics in a future blog; for now just start with a basic data catalog (what data you have and where you got it) and log the steps you took to combine and analyze the data.

Tools.  Let’s face it: businesses run on Excel.  Excel is an excellent tool for basic analysis, but it is not designed for combining data from multiple sources or for advanced statistical analysis or visualization.  You will need additional tools for those purposes, all of which should run on a laptop as well have server-based versions that support sharing across the business.  You’ll also need a relational database that is relatively inexpensive and easy to manage, such as SQL Server.

Next: Taking action.  Note that your analysis can lead you to new insights, but only acting on them will produce new results.  Any new action will require the business to do something new or differently; you will ask people to change.  Change is good but change is hard.  Therefore, initial actions should be incremental, starting with a pilot to manage the risk and expense, and to measure the costs and benefits.  Extending the new approach to the entire business requires organizational development – not just “change management” – in the form of training, coaching, teamwork and sometimes re-organization.   All of this must be done within the context of the organization’s culture, because as W. Edwards Deming said, “culture eats strategy for lunch.”

In this aspect, SMBs have a distinct advantage over large companies: their smaller size reduces the magnitude and complexity of implementing the change.  Smaller companies typically have cultures based on personal relationships and trust – “we’re like a family here” – making it easier for everyone to be engaged and supportive from the start.  That said, adopting new ways of doing business, even when justified by compelling data, is typically the greatest challenge of any analytics program.

After the pilot and subsequent full implementation, the SMB must measure the impact, typically in the form of improved (or new) key performance indicators, enabling the calculation of the return on investment.  The metrics also serve as data for further improvement, in a virtuous cycle of data-driven performance.

Key take-aways:

  • You already have the most valuable asset that you need: your company’s data
  • Tools and techniques for data analytics are more accessible, capable and affordable than ever
  • Your analytic process must be repeatable to be reliable and trustable
  • The biggest challenge begins AFTER the analytic discovery: changing the business to improve performance

Future blogs will explore analytical techniques and tools, practices for data management, change leadership, and the evolution and maturity of analytics programs: reality vs. hype, and how to achieve real and lasting impact.

Tags: , , , , , , , , , , , , , , ,

Leave a comment

You must be logged in to post a comment.