Everything about Web and Network Monitoring
Everything about Web and Network Monitoring
To keep competitive in today’s environment business leaders and stakeholders need to develop a clear predictive analytics strategy. And while it’s easy to assume that Big Data and PA are only for big businesses with well established budgets, nothing can be further from the truth. The path is wide open and resources are readily available for small business leaders to jumpstart their PA strategy today. The oft-quoted phrase, “A journey of a thousand miles begins with a single step,” is appropriate here because getting aligned with Big Data can seem daunting for some. But with clear planning and deliberate strategizing, a predictive analytics can be a big win for your small business in 2014.
What is Your Business Challenge?
The driving agenda for any predictive analytics project must be the business problem. This has to be front and center because the success of the project will be evaluated based on how well it provides actionable results and insights for the rest of the business. As one resource well states, “keep the business problem at the center of everything you do.” Don’t let PA become just an aspect of IT but ensure that it offers value throughout the entire business.
Look across all lines of your business to see where the painpoints are. What are your highest expenditures? What steps need to be taken to improve the bottom line? Ask your key stakeholders to join you in brainstorming the top areas where you’d like to cut costs within the organization. For example, advertising is a prime area where the return on the investment is notoriously low. How much do you spend on advertising and mailings? How about predicting the likelihood of which of your customers will respond to your targeted advertising? If you can predict these questions with a fair degree of certainty, the cost savings will follow.
Applying web analytics to your targeted web traffic can save you money as well. Look at your site traffic and KPIs and review major trends over time. Cost savings may be staring you right in the face!
Also, consider retail analytics as a solution to understand your customer buying patterns and the most appropriate merchandise to offer them. Understanding customer visits to your store and high peak periods can help businesses offer more customized shopping experiences.
Some basic questions you may wish to ask are: How can we cut advertising costs by 30% in the next 2 years? How can we increase our customer base by 25% in the next year? Who is our most profitable long term consumer and how we should target them?
Understanding Your Data
Once you’ve identified a business problem to address in your organization and have gotten buy-in from the key stakeholders, the real fun starts. Understanding your data environment is the first step in the data preparation stage. Some basic things you’ll need to explore are the variety and types of data available to address your business question and support your predictive model? You will also need to know if there’s enough data and enough history and granularity in the data to correlate with your analysis. Also, you’ll need to know where your data is stored and who oversees it? Who are the subject matter experts that manage the data sets? Invite them to become partners in your predictive analytics project.
Once the data sets for analysis are identified, you’ll want to understand the data characteristics such as possible values, data types, data formats, the time period of the data, and how the data was collected. You’ll also want to consult metadata sources such as system documentation, the data dictionary, database listings, and any available data processing personnel.
There are growing numbers of companies on the market that specialize in helping businesses to analyze their existing data, or else to generate new data, in order to derive a 360 degree view of their customer interactions and behavior.
In the next segment of this series we’ll take a closer look at the data preparation stage, which is often the most time-consuming part of predictive analytics.
