Ep. 20 – Analytics on Steroids: Best Practices and use cases

In today’s episode, we will discuss the challenges business leaders face in making decisions with outdated and inaccurate information (coming from siloed applications) and how they can improve by building a dashboard from real-time information sources. Click on Subscribe if you want to be informed of new episodes.

Remember I told you in previous episodes (episode 7episode 8, and episode 9) that IT leaders often need to balance their financial and human resources dealing with short-term needs and long-term goals?

Analytics is a perfect example: Analytics is your capability to evaluate and respond to business needs and customer preferences using reliable information.

In this episode, we will talk about how solutions have matured to where you may improve your existing analytics platform without disrupting your business (no Big Bang approach).

On your mark, get set, go!

If you are like me, you know how powerful Excel is; you’ve been tempted many times to support your business decisions by extracting information from a siloed application (CRM, ERP, SCM, Call Center, you name it), and immediately after you faced a couple of challenges.

Major issues building dashboard from Siloed Applications

Data quality is critical if the purpose of building a dashboard is to assist business leaders in making well-informed decisions. The other essential part is having the information needed by the business use case (Information = relevant data)

Typically, these are some of the issues that you initially face:

  1. Lack of accuracy in data extracted: Many records (typically 2% – 10%) may have inaccurate information or lack standardization. Data format, currency issues, field types, typos, and fields with missing information, may be fixed by filtering and improving records using an ETL (Extract, Transform, Load) approach. Right? Well, not precisely. Generally, normalization and cleansing create an 80% IT overhead (costly, time-consuming, and sometimes ineffective). Once we have the data cleansed, we also need to find a way to load new data cleansed as well.
  2. Data Privacy: Sometimes, the data is available on the system, but you don’t have access due to security and compliance policy. You may have access to totals, but you don’t have access to specific transactions.

Hm. May you share an example?

Imagine you want to know total sales per state, Country, Zip Code, or branch, but you don’t have access to customer personal information. Each transaction is needed to create the report. Still, not everybody needs access to unique transactions, as the results will be distributed based on your business security clearance.

This is a report I took from LinkedIn Analytics showing the interaction with readers /subscribers this newsletter had in the last 90 days: 9642 interactions and 129 feedback. Thank you for your preference!

Even though I may see the daily interaction level, I don’t have access to any reader’s name who is reading a particular episode. (That is why your comments are very welcome, to build real-time interactions and to learn from your comments and choices)

Another example?

During the COVID-19 vaccine trial, volunteers learned they would participate in a double-blind trial: neither the volunteer nor the physician at the point of vaccine administration would know if the volunteer was receiving a vaccine or a placebo. The business leader knew everything at the proper security level, but the data was filtered as appropriate at a lower security clearance level.

Can you improve these dashboard limitations?

Yes, a couple of maturity steps. The first step is building an Enterprise Data Warehouse (EDW).

Pros and cons of Building an Enterprise Data Warehouse (EDW)?

An Enterprise Data Warehouse is a substantial improvement, as the EDW can collect and process data sources from multiple siloed applications, but now you have as least two issues.

  1. Not user-friendly for Business Leaders: Sometimes, there are restrictions on the business leader’s capability to create a dashboard. These solutions are typically designed to be used by IT subject matter experts. Sometimes the IT response time to develop new reports is not the best because of…
  2. Lack of consistency with data cleansing: IT is jam-packed with semi-manual processes to filter data at the origin (ETL). In many cases, the time utilized by IT to connect new sources of information, cleanse the data, keep the data cleansed, and create reports, is four times bigger than the time used to analyze the information. Business Intelligence capabilities can reduce this impact.
  3. Suitable for Enterprise Applications, still issues with Live Data streams: Typically, these EDWs are very effective in connecting siloed applications. However, they are ineffective in connecting to a live data stream, such as user feedback from mobile apps or social media channels.

Can you improve these limitations?

Yes, this is what I call Analytics on Steroids.

Analytics for Digital Transformation, Best Practices:

There is a recent trend in buying SaaS Solutions to accelerate the creation of Business Dashboards interacting with Enterprise Data Warehouses.

These are some of the critical ingredients of the solution provider leaders in the segment.

  1. Cloud-ready, security compliant: A SaaS solution is a practical choice as many data sources are already in the cloud. There is no critical security exposure increase, and there is an essential reduction of complexity by allowing the solution provider to deal with provisioning, supporting, and maintaining the solution at the current version.
  2. Easy to connect/integrate: The solution provides powerful resources to connect to multiple siloed applications (via approved, certified APIs) and live data feeds (mobile apps, social media channels) to capture user interaction and feedback. This is a crucial ingredient, as one of the essential elements of Digital Transformation is allowing the organization to interact in real-time with customers and partners and learn and adapt from these interactions.
  3. Reduced IT SME workload: The solution provides substantial resources (Artificial Intelligent based ETL and AI-based cleansing mechanisms) to correct data inconsistency in the source and keep the data cleansed in the EDW. This is super important because now IT has reduced its workload supporting data integrity and is wholly focused on helping the business (remember, 80% in cleansing the data and 20% in analyzing the data)
  4. Data Clearance enforced via User Profiles: The solution includes user profiles, so even though data is still under IT control, as security is implemented, business users will have access to information based on their security clearance profile (data fields and data granularity).
  5. Data Privacy compliant / Data Masking: The solution also provides a data mask; the data extracted can be randomized and encapsulated to protect user identity, but it is still relevant for analytics (editing username, or physical address, among others).
  6. Building data subsets to support multiple teams: Existing solutions provide services using a multi-tier architecture: Capturing Data and Loading information> Processing Information > Extracting Information to Data Lakes and Warehouses to address specific use cases with relevant data. Typical use cases are Cost-Saving analysis, Merger and Acquisition evaluations, Marketing Campaign profitability, Biz process optimizations, Customer preferences, and market trends changes.

Remember, Analytics is not about a tool or a technology; it’s about responding to market needs based on quality information. Technology and solutions must address these needs by enabling the business leader to develop effective reporting safely, securely, and with autonomy. Extracting and building subsets of information is essential because it is an effective way to improve response time.

Everybody wins!

Benefits per user personas using last-generation Analytics

  • Business Leaders: Now, they can create Dashboards in no time, using a friendly, self-provisioning tool in the cloud capable of connecting all relevant sources of “information” (data with context, including Enterprise systems and real-time customer interactions)
  • IT Leaders: Building value in real-time, without affecting team members’ workload (work/life balance, anyone?)
  • IT Security: Information compliant with regulations.
  • IT Subject Matter Experts: What is Work/Life balance? Does anyone have a better approach for supporting the business focus on business results instead of cleansing data and dealing with data inconsistencies?
  • Customers: Even though some customers hold comments to congratulate your success in responding to their needs, they will recognize your effort by increasing brand loyalty and long-term profit.

Competitors? Well, not everybody will be happy. Typically, suppose you are the first in your industry segment to improve your Analytics. In that case, your competitors will have difficulty dealing with your innovation and increased maturity level doing business in real-time.

Hmm, I see, any other way to improve Business Analytics in a short time?

Well, not really; this is a perfect example of using available choices to respond to short-term needs.

You don’t need to update all your siloed applications to build practical Analytics. 

Still, the improvement of having an Analytics engine with your current IT platform is mature and compelling enough, so you may decide and make a low-risk decision. As you deal with upgrading your legacy systems, your Dashboard will be enriched with new information coming from AI, ML, and IoT functionalities captured by your upgraded Enterprise Systems.

Good enough?

Section Bar

What are your thoughts on the subjects raised in this edition of the Digital Acceleration Newsletter?

Share them in the comments below, and if you have ideas about other topics you’d like to see covered in this newsletter, feel free to add those suggestions.