Analytika: The Way of Analytics

Analytics All-Around

This is the first part of a series (called Analytika) of articles related to analytics in the world. Whether you’re an MBA grad, a new or experienced Product Manager, or a veteran Executive — in today’s world of business, non-profit or any form of leadership, data-driven decision making can be an added benefit to your organization. Lets look now at a few examples in our world.

Fashion Analytics

Zara uses an adaptive supply chain (management approach) process of replacing experts predictions with data analytics. The process goes roughly like that explained below:

• Zara HQ only allows its stores to order two weeks worth of inventory rather than for a season.
• HQ collects weekly sales data from their stores and analyses its items’ sales against supply.
• In addition, the analysis measures performance metrics of different features with supply and number of sales.
• For example, certain colors sell more than others during particular time periods (or) certain attributes of a particular item are better than other attributes at particular time periods (or) seasons etc.
• Zara HQ then uses these data insights to provide inventory to its stores making sure, they keep up with trends based on customers’ demands and hence, renew their products fortnightly.

Baseball Analytics

The way of the analytica are also used in sports like baseball. Former baseball player Billy Beane used analytics to take Oakland Athletics to the pro team it is today. He discovered that scoring runs were highly correlated with certain analytics.

For example, he found that a team with high on-base percentages (a measure of how often a batter reaches a base), is a team that more likely scored more runs and as a result were more likely to win the game(s). When implemented, the most immediate result was that the team had more walk outs than strike outs.

Hospitality Analytics

Carnegie Mellon researchers used a large data set to build a data model that included finding the correlation between quality of photos and booking history and whether the correlation makes a significant difference in consumer demand.

In the process of research they also determined the types of photos that would be the most effective in attracting customers. Some results included the way the bedroom was portrayed. The better the look of the bedroom, the higher probability a batch would be booked. Some examples of attributes that helped included:

1. Having windows in your bedroom;
2. Having drawn out (or no) curtains;
3. Taking photos from the corner of the room that includes both the bed and windows;
4. A bed with pillows, bed sheets, that were white or ivory in color;
5. Having photos or art in the room adds an extra touch to the look.

HiPPOs

With data all around, how can organizations use it during decision making — especially when society has taught management and leadership based on the traditional MBA-way which valued HiPPO mentality.

HiPPOs are a term described to categorize the Highest Paid Persons in Organizations. These tend to come from the world of the MBA as the traditional MBA valued gut decision, intuition and tradition over data analytics.

An independent study by QualPro found that the HiPPO based decisions make no effect or hurt performance 75% of the time (even worse when decisions were made at random). The best way to avoid HiPPO decisions is to have data ready when engaging in discussions that may lead to HiPPO decision making.

© Pexels

Data-Driven Decision Making

Let’s get into the meat of what it means to use data in decision-making. Decision making using and driven by data can be categorized into three domains. These are:

Data Analytics

• Finding patterns to draw conclusions and derive insights

Data Science

• A combination of statistics, math, programming and problem solving to find patterns along with cleansing, preparing and aligning the data.
• These tend to be techniques used when trying to extract insights and information from data.

Big Data

• This is analytics and science applied to large data sets and are typically unstructured data.
• There is no defined measure of when data becomes Big Data.

Analytics Types

When making decisions, the decision makers must identify the types of data analytics available to them. The types of analytics used in the domain of data-driven decision making can be categorized into three large classes:

Descriptive analytics

• This answers the question, what has happened?
• It condenses historical or real time data into smaller meaningful nuggets of information.
• For example, in internet marketing analytics, descriptive analytics are used to summarize large numbers of search, display and social media advertising campaigns into smaller sets of metrics that show the average click-through rate; conversion rate and; return on investment of these advertising channels.
• The main objectives of descriptive analytics is to find out the reasons for success and failures in the past.
• The vast majority of big data analytics used by organisations falls under the category of descriptive analytics.

Predictive analytics

• This type uses data from the past to predict what may happen in the future.
• It is based on having a solid understanding of the past.
• For example, predicting the likelihood that a new prospective customer will respond to a promotional email campaign. We analyse historical data from prior email campaigns where prior customers responded and did not respond to those email campaigns. A data model is built based on this past data and helps assess the probability that the new prospective customer will respond to a future campaign.
• Most organizations employ predictive analytics after hey have gathered enough descriptive data analytics.

Prescriptive analytics

• Once descriptive and predict have been “mastered” by the organization, prescriptive analytics is used using the models of the past and the forecasts of the future.
• This method is all about optimization through algorithms to determine the sets of actions needed to achieve the most desirable objective, given the forecasts/predictions of what is likely to happen.
• For example, based on the prior two examples, above. Once we have descriptive data from marketing campaigns and predictive data models from a myriad of marketing campaigns, prescriptive analytics models can be used to determine which types of campaigns should be sent out to which type of prospects to maximize sales and stay within the marketing budget.

As organizations evolve and gain experience, they move from descriptive to predictive analytics and finally prescriptive to make data-driven decisions that are informed decisions and actions for a myriad of functions.

At Vikasa Studios, we work with all types of analytics. Our approach is based on our client’s needs, as part of our Think Tank service.

To Be Or Not To Be

Although, it is current a fad to say words like “Data Analytics” or “Big Data Analytics”, when making decisions, analytics can be used in some instances while not in others.

To Be

Instances in which analytics can be used:

• Sales and Customers Relationships
• Supply Chain and Operations
• Human Resources
• Finance and Accounting

Not To Be

Instances in which analytics is not recommended:

• Trying something new and innovative where no prior history exists
• If there is a rare event — analytics cannot be used to predict outliers or anomalous events.
• Confirming something that the confirmer already (preexisting) believes (Confirmation Bias)

Analytics has been taking the world by storm and will continue to do so, in the present and future. There is no doubt that it will not go away anytime soon and as such, it helps knowing the basics when wanting to make decisions based on them. We hope you enjoyed this brief introductory article and ask you to subscribe to our publication to find out more related to analytics in decision-making and other areas, we here at Vikasa Studios delve into on a daily basis.