Reporting and Analytics: Transactional, Analytical, or Blended Approach?

Reporting and Analytics: Transactional, Analytical, or Blended Approach?

Does your company have a Chief Information Officer in its C-suite? If so, you might be a little ahead of the curve. Having a CIO oversee a business’s information and technology functions can mean having a centralized and coordinated influence in its various departments. With the strategic use of data, the CIO can reinforce the company’s overall goals throughout its divisions in ways other C-suite executives can’t, as data and IT systems are essential to every department. With CIOs at the helm, companies are breaking down silos with data-driven decision-making.

How Do CIO’s Manage Data?

Regardless of a company’s size, most of them process and store massive amounts of data. These can be challenging to manage. Unfortunately, many companies don’t have adequate data technology to meet this demand.  

And the demand is high. Studies show that although most businesses collect all the right data, on average, less than half of their structured data is used for decision-making. What’s more, less than 1% of the average company’s unstructured data is ever analyzed. On top of that, some companies are dealing with data breaches and rogue data sets, both of which can halt or negatively impact operations, and their technology is simply not capable of dealing with all of this.

For these reasons, one of the most important decisions the CIO must make is choosing a system for data reporting and analysis. There are many different types of database systems with a plethora of features to choose from. But for simplicity’s sake, we’re going to divide them into three main categories:

  • Transactional,
  • Analytical, or
  • Hybrid Transactional/Analytical Processing.

Transactional Vs. Analytical Databases

What is the difference between a transactional and an analytical database?

Traditionally, many companies, especially those whose revenue depends on sales, have used a transactional database (also known as online transactional processing, or OLTP) for the day to day tasks of running their business. For example, a hotel chain would record a transaction for each guest who checks into a room at each of their locations. The transaction would probably include the time and date of check in, the number of adults and children in the party, which room number they will be staying in, and the price of the room.

At the end of the day, the manager for each hotel could run a report showing how many rooms were occupied that day, how many were vacant, and the total sales.

Running the business VS. managing the business

While transactional databases are most often used for running a business, an analytical database, or OLAP, helps the CEO, CIO, and other C-suite executives manage the business. These systems pull data from many different transactional databases in order to perform complex analytical queries in a timely manner.

So the CIO of the hotel chain can analyze data from room check-ins, food service, concierge service, airport shuttles, housekeeping inventory, water usage, and any other transactional information the company may have. In seconds, the CIO could run an analytical report showing, say, one particular hotel’s busiest time of year, and the customer’s use of room service at various times of day. They can compare costs such as workers’ salaries, food, and kitchen materials to sales.

Using that report, C-suite executives can determine how many cooks and servers should be clocked in for each shift, and how late it would be profitable to keep room service open. They can also be a little more proactive with their use of data. If room service is popular, they could run a promotion offering a free meal with each stay. Or perhaps invest in a creative chef to invent new menu items.

How data-driven decision-making enables growth

An analytical database supports critical reporting needs and enables growth. Going back to our example, the hotel chain could analyze occupancy data in each of its locations to determine where they should expand, perhaps by building a second hotel nearby. Or they can identify the costs and benefits of shutting down a hotel that isn’t performing as well.

When businesses use data to drive their decision-making, the opportunity for growth increases dramatically.

What Can a Blended Approach to Data Management Do?

One size doesn’t fit all

Unfortunately, it is difficult for any one system to support both transactional and analytical data management. A transactional database would not be able to perform a complex analytical query in a reasonable amount of time. This is because the transactional system is designed for data entry rather than data retrieval.

Going back to the hotel example, let’s say the manager for one of the hotels runs his regular nightly report showing how many rooms are occupied and vacant tonight. Looking at the report, he notices something unexpected. Only half of the hotel’s rooms are occupied tonight, and this is usually a busy season for him. He wonders, “Is this just an anomaly? Or is it the start of a trend?”

He tries to run a query in his database showing the occupancy rates for each night over the last month compared to the same time last year, identifying any other outlying data points. While it might be possible for his transactional database to produce this information, it will take a very long time, and it will tie up the system, slowing it down for any employees who may be trying to use it to check in a new guest.

At the same time, analytical databases produce very fast analytical reports, but using it for individual transactions would be tedious and slow. The database is organized to allow businesses to easily find and retrieve data. The architecture of the database literally aligns data in columns and then plucks what’s needed for a particular query, putting it all together into an intelligent report. But it does this by using large batches of data that has been input elsewhere.

If the clerks at our hotel tried inputting check-in and occupancy data directly into the analytical database, it would take a very long time.

Combining transactional and analytical capabilities

In order to thrive in today’s market, though, businesses need transactional and analytical capabilities. Some executives get around this by purchasing both types of systems and using extract transform load (ETL) technology to move their transaction data into their analytical databases. But many companies complain that the ETL process is too slow. It can take five days for data to transfer through ETL, meaning companies can only use their systems to make decisions based on past data.

To solve this problem, many companies use a Database Management System (DBMS) with blended features. With this type of database, businesses can create, update, retrieve, delete, and manage data.  

Blended databases are also known as Hybrid Transactional/Analytical Processing (HTAP). HTAP is a term coined by Gartner, Inc. Gartner says that HTAP “breaks the wall” between transactional and analytical processing so that businesses can make informed decisions in real time. Recent advances in research as well as OLTP and OLAP capabilities make it possible for transactional processing and analytics to operate on the same database.

Advantages and disadvantages of HTAP

One major benefit of this type of DBMS is that it allows businesses to analyze data and inform decision-making in real time. Rather than waiting five days for transactional data to transfer into a separate analytical system, C-suite executives can make decisions on the fly. Often, business moments occur and pass quickly. If the company doesn’t react immediately, that could be a huge missed opportunity.

While HTAP offers great potential for business innovation, it does have some challenges. As it’s a relatively new technology, there are not many people who are well-trained in its use. There also aren’t well-defined best practices.

How can your business take advantage of HTAP?

What if your company doesn’t have a CIO? You might want to consider contacting a headhunter to help you hire an interim executive or interim management team to help you get started.

Another good option is to contract with a management consulting team to manage your DBMS for you. In fact, the CIO is one of the most outsourced positions in the C-Suite.

Tatum has a strong bull pen of talented executives and management consulting professionals with the experience to help your business succeed. Call us today to find out what we can do for you.