If you're looking for help to solve business problems using analytics, you'll find yourself surrounded by hundreds of vendors, technologies and services. All of them claiming to use analytics to help your business succeed. But how do you decide what the right solution is for your business? We're here to help. To reduce risk and enhance success, I have developed six practical ways of finding the right analytics solution.
Implementing Business Analytics
Often, the term "analytics" in business can often feel ambiguous. There are hundreds of companies that claim to "help" by using analytics. However, that's analogous to claims like: "I can help you with math." Sure, but a third-grader doing fractions has very different needs than a university student solving partial differential equations. Both activities constitute math, but specificity matters. Few analytics-driven business providers offer the necessary specificity in their pitch. Analytics can be used to describe a wide array of topics from aggregating databases, data visualization and reporting, to cutting edge advanced statistical modelling.
So, without further ado, here is my list of six ways to find the right analytics tool for your business:
1. The "Sexy" Terminology of Business Analytics
Terminology in the analytics space is continuously evolving. It's hard to read through a business magazine without seeing mention of machine learning, deep learning, AI, VR or other analytics software.
These concepts may be at the forefront of analytics solution providers' marketing at the moment, but they have been in practical application for years. When terms become "buzz words," many people begin to lose sight of the business problem they are trying to solve and instead look to apply whatever the latest tool may be, regardless of whether or not it is the right tool for the job at hand. Do not lose sight of the overall goal or business problem you have to deliver on.
2. Set Your Data Analytics Framework First
Frequently companies will finance large, complex, and expensive solutions or a business intelligence analyst (Ferrari's) to "solve" problems without first considering the infrastructure (e.g. people, structure, targets, etc.) that need to be in place to achieve success.
More often than not, the brand-new Ferrari (expensive but useless solution) will sit collecting dust in the garage. Before buying a Ferrari, we recommend first learning the fundamentals of driving, auditing your vehicular (or, really, analytical) needs, then test driving some cheaper alternatives.
Pilot programs or home-grown solutions are an excellent solution to test an organization's capability to transitioning into a data-led organization.
3. Don't Get Hung Up On Achieving "Perfect" Data Quality
This is most frequently a point of contention with those who spout the adage "garbage in, garbage out." Let me preface the remainder of this section by saying that I agree that data quality management is key. However, there is a vast range of quality between perfect and incomplete.
I've seen leaders so focused on gathering a "perfect" data set that they go months without being able to make a business decision. We recommend utilizing whatever most complete information is available, even if it only gets you 80% of the way there. This approach allows businesses to make actionable decisions and creates an environment that encourages learning about the quirks and inconsistencies in the data that require future adjustment.
4. Avoid Technical Jargon
Put simply, make sure you understand the data analytics process and can describe what you are doing in plain English (so simple even a toddler could understand). For example, a tool that segments customers using hierarchical Bayesian k-means clustering methods can be simplified to "grouping likewise things based on attributes."
Technical jargon is endless and often used to portray the illusion of complexity and sophistication. I often hear vendors respond with "it's complicated," followed by a slew of technical jargon that is near impossible to comprehend. Swallowing pride and admitting a lack of understanding can force the conversation back to an understandable language, ensuring you understand what the tool you're buying actually does. Make sure never to commit to a tool or project which you do not fully understand!
5. Use the Sniff Test
The sniff test, a valuable tool to use in everyday business life, is just as pertinent in the analytics space. Most people would be familiar with the concept of doing a smell check or back of the napkin calculation to ensure the outcome is reasonable.
Nevertheless, I see professionals every day, trusting the output of analytical tools blindly, without the sniff test. Recently, we worked with a client that was using pricing elasticity outputs to quantify increasing prices. They were shocked to see a significant volume impact when the price was adjusted. There was an obvious underlying error in the model not accounting for distribution gains, artificially lowering the elasticity on individual SKUs and costing the company millions. A simple sniff test would have identified the problem and saved millions.
6. Show Me the Money!
When embarking on any new analytics initiative, we recommend remembering the wise words of Jerry Maguire: "Show me the money"!
It's easy to get lost in metrics like clicks, conversions, and market share, etc., but if you don't understand the financial value of those metrics for your organization, they're useless. If a click on an advertisement does not lead to increased interest in your product or convert into revenue, does it make sense to invest in a solution to drive and monitor clicks? We think not. A narrow, one-dimensional viewpoint can be costly so, before investing, have a clear vision of your goals in the investment you're making, and the potential financial return.
ABOUT THE AUTHOR Michael Stanisz is a Partner at Revenue Management Labs. Revenue Management Labs help companies develop and execute practical solutions to maximize long-term revenue and profitability. Connect with Michael at [email protected]