Attacking the Business Intelligence Backlog with Analytics Services

Bookmark and Share

Image courtesy of Shutterstock.

Imagine that the travel manager at your organization - we’ll call her Tracy Travel—is looking for a way to leverage data to cut costs and better understand her travel program. She knows that with appropriate analysis she could impact the business, but fears her projects would never make the priority cut with the information technology team.

When it comes to data analysis, companies often have a long project list with each project having it’s own set of specific needs—the net result is a lengthy and growing project back log. Since the late 1980’s Business Intelligence (BI) vendors have been attempting to address this backlog through various “ease of use” enhancements. We’ve seen big improvements with the latest generation of cloud-based BI solutions, yet despite these advances, obstacles remain for those who just want the analysis, like our travel manager Tracy.

First, the data is often hard to understand, especially when it comes from third parties. Second, even if the BI solution makes it easy to apply advanced analytic techniques, applying them correctly is an entirely different matter. Finally, managers like Tracy have their day job to consider, which leaves little time to invest in troubleshooting these obstacles. The experts in IT can address the first two issues, but this still leaves us with the backlog of projects 

Analysis Projects that Work Well with a Cloud-Based Solution

Luckily, with the advent of cloud-based analytical services, providers can now process, handle, and analyze data to meet this unserved need. However, how should you evaluate which analysis projects work well with cloud-based solutions?

A few years ago the International Data Corporation (IDC) introduced a framework for evaluating the relationships between types of decisions and how they can be supported by analysis.

The IDC framework evaluates decisions across four dimensions:

  1. the risk associated with a decision
  2. the degree of collaboration and ad hoc analysis required to make the decision
  3. the quantity of decisions
  4. if the analysis can be automated

Let’s define a spectrum—with strategic decisions on one end and tactical decisions on the other—and evaluate the decision across these four dimensions. Strategic decisions utilize ad hoc analysis; these decisions are typically collaborative in nature and often many people are involved in making them. Collaborative discussions lend themselves to question and answer sessions/analysis. Getting them wrong has significant risk which means you’re willing to test and explore results. As you can imagine, this type of decision-making doesn’t lend itself to automation.

Tactical decisions are often centered on the questions you should be asking about your business. For example, questions around the accounts payable process, overall T&E spend or how to combat fraud. There’s not much risk with getting any individual tactical decision wrong; however, you make many of them in the normal course of business. Tactical decisions that are made on a recurring basis at the front lines of business and often by a single person are perfect for automation. Cloud based analytical services or cloud based analysis can work well here, particularly when there is a need for repeated analysis.

Advantages to Cloud-Based Analytical Services

With cloud-based solutions, third-party companies are able to deliver analysis to organizations and those organizations act on the analysis without the need of a data scientist. Most analytical service providers have advanced extraction and mapping tools needed for any job, because they already spend a large amount of time converting different sets of data from various sources. Additionally, most of these providers assemble databases of third party information to make their analysis robust, competitive, and useful.

Of course, an organization can curate the data itself, but this presents hefty challenges, and is cost prohibitive, particularly for the tactical analysis needs. This is because the data comes in a variety of formats, for instance in a flat file vs. xml file. On top of the challenge of gathering the data, companies seeking to curate data themselves also have to spend time and money converting that data into a suitable format for analysis such as solving discrepancies between the foreign keys and your primary keys. Even if the analysis projects like Tracy’s made the prioritization cut, many companies would end up spending more for analysis on their own than they would with a third-party provider.

In a way, the cloud provider is getting much higher utilization out of the scarce and expensive data scientist resources, as the provider can afford to invest deeply in sophisticated analytics because costs are spread over many clients. Higher utilization also means providers are becoming increasingly proficient with advanced analytic techniques that are most meaningful for specific data sets, which is especially important in today’s world where there’s a shortage of skilled data scientists. A third party provider’s ability to service multiple clients creates more meaningful statistics because they are based on the data of multiple organizations, thus creating a better opportunity for business benchmarking.

For example, we recently compiled a database of addresses for the most popular gentleman’s clubs, which we use often to detect out-of-policy spending for our Insights On Demand customers. Although this list was time consuming and costly to compile, we spread this cost across all of our clients, as they all use and receive the benefit from it. This is why analytical service providers readily invest in various extraction/mapping tools and external data sets—because they will use the tools and data more than once.

Cost savings are the biggest advantage to using an analytical service provider, but not the only one. Again, because of their scale, analytical service providers can invest in user interfaces that present the analysis in an easy-to-understand format. This allows organizations to take advantage of the information gained from the analysis, act on the information, and implement change quickly throughout an organization. Previously, analysis was served up to businesses that then tasked their own analysts with scrutinizing the results to create interpretations, forecasts and courses of action based upon the findings.

The final perks to using a third party analytics provider: as data from multiple customers courses through one platform, algorithms improve as the customer base enlarges, and everyone receives the benefit. Cloud-based providers also leverage third party data sources to better detect fraud, further their own predictive analytics and improve search capabilities.

Examples of Use

Corporate Spending

In corporate spending programs forensic analytical services could be used to monitor all corporate card transactions and flag expenses that are out-of-policy or highly suspicious. Analysis of an accounts payable department could highlight duplicate invoices, or leveraging a vendor address database could uncover fake or fraudulent vendors. Monitoring alerts could notify management of a “problem traveler” who repeatedly abuses the system or is creating fake expenses to put money back into his/her own pocket. Analytical Services also help create action plans and shape behavior; reviewing potential red flags could lead to a change in policy if most travelers seem to be committing the same offense for a seemingly good reason.


Corporate and government compliance is a tricky area that is becoming increasingly difficult to monitor and lends itself well to the analytical services model. With the recent news on Fortune 500 companies, such as Avon and Wal-Mart involved in costly FCPA investigations, it is in any organization’s best interests to create an effective (and cost efficient) monitoring system to combat non-compliant activity. Leveraging analysis to flag these expenses reduces compliance risk and allows management to handle situations in a more efficient method than manual oversight alone.


Lastly, and perhaps the most common application of cloud-based analysis, is benchmarking. After all, without knowing where the average is, how can we know if our own efforts are above or below “normal?” Keeping data in the cloud allows for benchmarking opportunities like never before, giving organizations greater insight into how similar businesses are handling their finances. With benchmarking, companies are able to compare their spending, best practices, strengths and weaknesses to others. In finance, the benefits to benchmarking data related to T&E and P-card programs and within accounts payable are immeasurable.

Benefits Specific to IT

There are a lot of analytics projects at companies, and many won’t make the cut, even though they have a provable return on investment for the company. IT as a whole is often understaffed and overburdened. Even if a subset of those projects lend themselves to the cloud based analytical services model, it is worth investigating as they can put time back in the day of IT and clear up the project backlog.

If cost is a consideration, or has been in the past, the advent of the cloud has produced the ability for third-party analysis companies to scale their services in a more cost effective way than on-premise models of yesterday. Providers can actually deliver a better answer than achievable internally, and do it in less time with fewer headaches. In short, information technology teams need to be taking advantage of all that advanced analytical services have to offer.

About the author

Patrick Taylor is the CEO of Oversight Systems and an authority in the convergence of business analytics, information security and the implementation of technology to boost organizational performance.