Ask an Expert: Understanding the Predictive Analytics Landscape with Beyondsoft

By Colin Gibson

Since being established in 2004, Beyondsoft Consulting Inc has been helping enterprise companies resolve IT business challenges in many areas, including data and analytics, translation services, QA and testing, cloud computing, digital platforms, and consulting and strategy.

One of their core competencies in data and analytics is predictive analytics – using actionable insights to help customers drive better business outcomes. Today, companies are using data to understand their customers’ buying behaviours, manage resource allocation, and determine which lines of business to invest in. Predictive analytics takes the guesswork out of business decisions and helps streamline operations.

Today, Dr. Rob Newman, Director Solutions and IP Development at Beyondsoft, is shedding some light on the current, as well as future, landscape of predictive analytics.

Can you provide examples of the kinds of business advantages that predictive analytics enables?

Companies that leverage predictive analytics reap a goldmine of insights that position them to accelerate success and outperform competitors. Predictive analytics enables businesses to boost productivity, drive revenue, generate quality leads, reduce risk and optimize products, among countless other unique business objectives.

Predictive analytics can be used to build recommendation engines for eCommerce companies to increase consumption – Amazon and Netflix are the obvious leaders in this area – but so are companies that provide other services for their clients. Airbnb is an excellent case study because their software engineering team use predictive analytics and machine learning to recommend how to price a property to maximize the probability of it being rented on any given day. To learn more about Airbnb’s predictive analytics strategy, I’d highly recommend checking out their informative posts on their Medium website.

Other companies use predictive analytics for marketing campaigns. Misdirected marketing communications can damage a brand identity, wasting time and money, while highly tailored campaigns can yield a much higher Return On Investment (ROI). Machine Learning can clearly identify and target the people who are most likely to be receptive to marketing materials, building brand trust and generating a fantastic ROI.

Finally, we’ve used machine learning to help customers design online training systems to optimize course structure, identify the best instructor for a particular course, and target prime demographic(s) to market the course to for the purpose of maximizing completion rates.

Prior to the advent of predictive analytics, how were your clients using data to make decisions?

Many of our customers were using their institutional knowledge, subsets of historical data, and intuition to make decisions for their future business processes. While this method isn’t inherently wrong, predictive analytics can simplify and streamline these decisions and deliver far more accurate results. Predictive analytics also allows you to change input parameters to model different outcomes. You can generate insights that remove unconscious human bias, which enables our customers to objectively assess the best path forward to achieve their business goals.

What are some of the barriers to entry that your clients experience in terms of implementing predictive analytics? What do you use to mitigate or resolve these issues?

The business benefits of using predictive analytics far outweigh potential drawbacks. With that said, we find that there are two common challenges customers must overcome when considering implementing predictive analytics.

  1. Unrealistic expectations: Right now, as detailed in the Gartner Hype Cycle for 2017, machine learning (known as “predictive analytics” when applied to data analytics) is coming down from the “Peak of Inflated Expectations.” Machine learning is perceived by many business leaders and executives to be a silver bullet that will solve their current or future business problems. Companies often believe that everyone else is already using predictive analytics and that by implementing predictive analytics they will be thought of as cutting edge. Occasionally, companies also mistakenly believe that their data is of high enough quality to be used to train machine learning models.
    For us, this means that we have to set realistic expectations from the start. We need to have candid discussions with customers around their data provenance, data governance, and data integrity.
  2. Data security: Customers often wonder if their data is secure in a cloud-based, software-as-a-service model, despite the majority of reports stating how much more secure data is in the cloud than in typical, on-premises environments. Educating business owners on the current state of data security is one of the first things we do with new clients.

We work to be transparent by communicating with and involving customers in every step of the process. From the start, we take the time to help dispel any misperceptions and set realistic expectations around predictive data as well as data security. We help them to understand the importance of the quality and quantity of their data.

We also encourage our customers to start their predictive analytics experience with a short-term discovery and proof-of-concept (POC) engagement. Our data architects and data scientists spend time at the customer site working with stakeholders to understand the business questions they are facing. This involves reviewing their current success criteria and KPIs, researching their data quality and quantity, and determining what business insights we can confidently generate from a custom predictive analytics solution.

If there are foundational data integrity issues, the engagement may shift from predictive analytics to a data-cleansing discussion. Some companies may not be ready for predictive analytics. This may be due to their current business processes or a lack of data integrity and/or data relationships across multiple data sources. In such instances, our Data & Analytics team can step in to help connect data sources to organize, analyze, and present data in a usable form, which enables customers to discuss, monitor and make the right decisions at the right time.

Ultimately, we want customers to get the best return on their investment. And sometimes that means fixing data problems before rolling out a full-scale predictive analytics solution. And sometimes that means working on a smaller-scale project with a quality data source and clear business objectives.

Traditionally, Beyondsoft’s customer base is enterprise-focussed. Are you seeing SMBs beginning to use predictive analytics as well?

Approximately 80% of SMBs are using some form of analytics, ranging from ad-hoc reporting, to simple Excel spreadsheets, to comprehensive data warehouses, but only 30% are taking advantage of predictive analytics. Our Machine Learning-as-a-Service is built specifically as a cost-effective way for SMBs to achieve the same level of predictive insight into their data that enterprises commonly have already.

Why should more SMBs consider implementing some form of predictive analytics?

There are many reasons. Predictive analytics provide SMBs with a competitive advantage by allowing them to make business decisions proactively, using predictive data, rather than reactively, using historical data. Whether building an online recommendation engine, rapid targeting for marketing, anomaly detection for security associated with credit card transactions, or automated workflows for service desks, predictive analytics can process more data faster, generate better quality results, and modify and repeat analytics tasks more efficiently than a team of human analysts.

One attractive option for SMBs looking to invest in predictive analytics without breaking the bank is BeyondLearning. BeyondLearning by Beyondsoft is a comprehensive, cost-effective, subscription-based Machine Learning-as-a-Service (MLaaS). Our consultants help customers understand the value associated with predictive insights. We work alongside customers to build and deploy an end-to-end predictive analytics solution that positions them to use data to its fullest potential.

To learn more about Beyondsoft and BeyondLearning, visit their website: