Money-Making Analytics (Part 2): Evaluating the Technology and Tools
Now that you have finalized your analytics strategy and key stakeholders, you must choose the best technical solution for your company. Learn how to establish the right technical elements to enable the existing company culture, and not fight it.
- By Shikha Verma
- February 9, 2016
In My American Journey, Colin Powell writes, “Experts often possess more data than judgement.” I love that quote because it highlights the importance of judgement or logic over the vastness of data. Often times, we run into people full of facts and opinions but lacking common sense and judgement. To me, the analogy of data and judgement is equally applicable to big data tools and analytics. Just as good data is useless without good judgement, big data tools are useless without good analytics.
In the previous article in this series, we discussed the approach and questions to ask to create the specifications for an enterprise’s money-making analytics project. In this article, we will evaluate the technical landscape and company culture to determine the right technology set for deploying the analytics tools.
Evaluation of the technical landscape has four essential parts.
Part 1: Evaluate the data landscape
Your evaluation of the data landscape requires that you understand and map each potential source of data you wish to analyze. Some sources might be internal, some might be external. Some might contain structured or unstructured data, email messages, or images. Some are legacy sources, others have never been tapped before. At the end of your evaluation, create a detailed picture and catalog of all potential sources and what they share as well as what data integration you’ll need.
Part 2: Evaluate the technology landscape
Your technology landscape evaluation consists of evaluating the transactional systems, data integration tools and data visualization or Analytic tools that exist within the company. Wherever possible, leverage existing toolsets and skillsets to gain speed of delivery of the money making analytics. Select tools that favor agile development, as we will be iterating through multiple versions of the money-making analytics. Don’t fall for waterfall style development for these analytics. Timing is critical for money-making analytics and nobody will know the right questions or design for any of these. Iterate, iterate, iterate!
Part 3: Evaluate the quality of the data
If this is the first time action-oriented money-making analytics is being implemented in a company, chances are that the data quality is highly suspect. Thoroughly evaluate the data quality of key dimensional attributes such a customer name and customer data rollup hierarchies, product name and product data rollup hierarchies, and vendor/partner name and vendor data rollup hierarchies. Data rollup hierarchies are extremely important; they are used to organize the data. For example, a customer data hierarchy could be based on customer geography because that’s how most companies organize the sales force -- so it would vital to have customer ZIP code, state, and sales region data clean in addition to having a clean individual customer name.
Fixing data quality issues long-term requires a great deal of rigor and discipline, so think of short-term and long-term approaches. Because timing is critically important, you will likely not have weeks or months to implement data governance if it doesn’t exist. You will have to determine quick ways to improve the quality of key data elements in addition to pursuing the right long-term path.
For example, you may have to perform a one-time cleanse of customer name and hierarchies and determine a way to periodically cleanse the data manually until the systems and processes are corrected to prevent bad data in the first place. This step usually takes the most effort to convince and persuade leaders and users across your organization because data quality can be a difficult concept to understand. It’s even harder for many people to accept a companywide data governance program. It will take time, so keep short-term approaches handy and be ready for some rework.
Part 4: Evaluate the Company Culture
Once you have a good idea of the technical landscape, it’s time to dive into technology usage patterns and also work to thoroughly understand the technical skill set of the primary user base. Understanding the existing company culture about tools and technologies is essential when you choose the best toolset for your money-making analytics.
Timing is also critical. Work with (not against) the company’s culture to create the biggest impact with the least resistance to technology adoption. It’s not about the newest technology or shiniest gadgets. Stay laser-focused on the money-making potential of your analytics project..
I had the opportunity to work with a technically laid back company where spreadsheets were the data analysis tool of choice. Spreadsheets were so prevalent that even senior C-level leaders wouldn’t accept or rely on any numbers that weren’t presented in spreadsheets. After we figured out the money-making analytics this company needed, we made a conscious choice that spreadsheets would be one of the primary modes of distributing the results of our analytics.
Each set of analytics performed incorporated a spreadsheet version that could be sorted or filtered using familiar spreadsheet-like functionality. The analytics workflow that knit all the analytics together helped drive the needed action and adoption from the analytics users. For example, if the analytics tool was focused on helping a supply chain group decide when to ship the product in order to meet the customer’s arrival window, the workflow in the analytics tool would guide them towards the right answer to make their decision-making process easier and more intuitive.
On the opposite end of the spectrum, I also had the opportunity to work with a technically advanced company where spreadsheets were frowned upon and the CFO would walk out of a meeting if the analytics was provided in a spreadsheet. Obviously, the technology choices we helped this company make were very different. We chose a technology that was connected to the data warehouse and would show live numbers through widgets on the users’ dashboards. In such a company, choosing a low-tech option would have been detrimental to the usage and adoption of the money-making analytics.
This step is a tough one for most technologists because we want to use the latest and greatest technical advancements so the company is prepared for the future. Although that is certainly the right long-term goal, keep your focus on speed, ease of adoption, and the right culture fit for your first set of money-making analytics. Once you gain momentum, it will be must easier to move to the more advanced technologies if they work for the company.
With your strategy finalized and technology determined, we’ll look next at how to build your money-making analytics that help users take action.