Tipping Siloes: Integrating Data and Streamlining Processes for Efficiency
In the commercial world, data fragmentation often occurs as a result of M&A activities: there is always a rush to bring the newly acquired business units into the financial system and to ensure that the new employees are paid out that system. Data migrations related to transactional systems, however, are frequently not performed, as these tend to be more difficult and involve cultural changes.
When I worked in the media industry, our company acquired 18 business units in 2006-7. Often, employees from the acquired units were migrated to the new payroll systems within days of the acquisition and AP/GL operations were integrated within a few months. Customer-facing transactional systems, however, were typically not integrated into the systems used by other business units. They became data siloes, with any corporate reporting done through manual spreadsheets and journal entries. This model of integrating acquisitions is not uncommon in many companies.
In the public sector at local level, departments and elected offices are frequently very independent. Either these departments have their own IT groups which are not coordinated with enterprise IT or IT responds to a specific departmental request for a solution and IT delivers it. In all of these cases the result is the same: data is isolated, fragmented or “siloed.” For example, departmental siloes resulted in the deployment of four distinct electronic document management systems (EDMS) within the county where I work; each of the departments had requested an EDMS solution at different times with different business requirements and no one insisted on standardization.
In both the private and public sectors, the cost of data siloes is high: there are both direct costs, such as the need to support multiple EDMS systems, and large opportunity costs due to data not being used effectively, the lack of standardized report and data structures, or the inability to understand the business. So why it is that data siloes are allowed to perpetuate and impede our ability to leverage ‘big data’, when there are few technical reasons that prevent the siloes from being knocked down?
In short, the answer is ‘culture.’ In the case of business unit acquisitions, integrating transactional systems would necessitate the change of workflows and configurations in the transactional system(s). The effects of such changes could be enormous: it might require re-thinking product pricing models or how customer interactions should take place. Historically, we have shied away from change management, even if the benefits of change are substantial and the costs of maintaining legacy practices are high. In the worst cases, nonstandard systems result in poor corporate reporting with a flurry of spreadsheets and hours of manual effort to report financial results.
In the public sector, there is no immunity from the reticence to confront cultural change. We, too, are fearful of changing business processes. When it comes to integrating data, we are conscious of governmental transparency, yet we are hesitant to share data across departments or between agencies. There are exceptions. Data about criminal justice cases can be shared, following specific compliance rules, between the federal, state and local law enforcement agencies. But, for the most part, public sector siloes are stout and the effort to tip them over and make data accessible has been limited.
To leverage enterprise data more fully, what is needed is a concerted effort to confront cultural change and to secure small ‘wins’ demonstrating the tremendous power that using data smartly can offer. A few examples from my experience will illustrate the potential of leveraging data.
At a business unit within a large media company, there was a sales director responsible for subscriptions who realized that his marketing and sales operation was suffering. His strategy was to increase his marketing spend because that strategy had worked in the past. Only now it was not working and there was a demand to improve the bottom-line. Realizing that only a datadriven argument would persuade the director to abandon his timehonored strategy, the corporate IT department worked with him to build a data warehouse and data marts to analyze demographic and historical subscription data. We then taught the sales director how to use the query tool and encouraged him to slice and dice his data, looking for ZIP codes and demographic clusters that seemed to have atypical sales and marketing costs. Within a couple of weeks he found the answer. There was one ZIP code in which a contract sales force was selling new short-term subscriptions and the customers were never remitting their payments, resulting in a product being given away gratis and a sales commission being paid. The total cost to the business unit: $750,000 annually. Once this practice was stopped and that subscription offer was eliminated, the hemorrhaging of cash also stopped. By merging customer and demographic data, the director was able to use data toisolate the problem and create an action plan to address the issue, in lieu of just spending more marketing dollars. In a recent public sector example, the combination of aid recipient data and Department of Motor Vehicle data has allowed our county to realize significant cost savings while at the same time delivering aid benefits to needy citizens more quickly.
With the implementation of some inexpensive technology and a single, concise workflow that relied on “mashed up” data from the State-wide aid system and the DMV, IT was able to deliver a solution. Now, when an aid applicant comes to the County, a reception clerk can swipe a driver’s license, querying DMV records to confirm the applicant’s residence. Concurrently, a second query goes against the State’s aid recipient database to determine if the aid applicant has received aid previously in the State or if he/she is a new applicant. The supporting documents submitted by the applicant are then scanned into the applicant’s record and the entire package is forwarded to a case worker for review. Following the review, an interview with the applicant is scheduled and aid is either approved or denied.
The new process can now take as little as 30 minutes from the time of application through the scheduling of the interview, down from five to ten days, as a result of combining all the data in a digital workflow. Previously, documents were photocopied and scanned into the EDMS system; residency was determined by the documents, which required manual review; and the determination of previous aid also required a manual look-up in the State system of record. With the new process, nearly 10 FTEs are saved annually and aid applicants can be interviewed hours after applying.
Challenging the change-resisting retort of “we’ve always done it that way” allowed us to develop a single, generic digital workflow instead of 70 different workflows for each unique type of case. In this instance, a small, yet significant, change, facilitated the quick review of cases saving time, effort and providing better service.
Why haven’t we pursued changes to use technology effectively and to serve customers and citizens better? It’s all about people and our inherent discomfort with change. Combining data and implementing optimized processes is scary; change is more easily avoided than embraced. Changing that culture by presenting the realm of the possible and achieving incremental successes, which demonstrate the efficacy of the changes, will move us forward to better service for our customers and citizens at lower cost— which should be the desire of every CEO and public servant to offer and every customer or citizen to receive.