SaaS systems can make data accumulation and analysis very easy, instantly generating aesthetic reports while holding remote operation capabilities. But this is easier said than done and requires a clear strategy and abundant technical resources.
Employees can bypass the conventional technology strategies by simply subscribing to these products. This can lead of duplicative efforts, lacking oversight. For instance, in the case of cloud storage & file sharing, there is a good chance that if you surveyed your employees you would find a variety of products in use, e.g. Google Drive, OneDrive, Dropbox, etc. While the trend of using digital tools may be appreciated, operations are at risk of becoming redundant and expensive as variety of subscriptions costs are hard to keep track of. Furthermore, the company’s data becomes sprawled, while also exposing it to security threats.
These are significant problems, however there’s one that’s even bigger: the SaaS apps’ ability to generate massive heaps of data. It was stated earlier that companies make use of dozens of apps but let’s take into account the data each app generates. Every login, every modification, notification, warning, etc. generates data. Multiply this with the hundreds or thousands of employees present and the information becomes a mess.
Remember, garbage in means garbage out. The quality of data is fundamental if a company wants to make use of the tools that are part of Industry 4.0. Incomplete user input, broken third-party extensions and poor data policies can all constitute for pollutants for datasets. As a consequence, they waste valuable computing resources.
A study revealed that companies often spend 50 – 80 percent of their time cleaning datasets. This is outrageous, since the resources tasked with these jobs are highly qualified and hired for analysis rather than fixing typos. A strategic approach to data management can solve many of these problems. As with any other business decision, a comprehensive plan devised by the leadership and supported by all functions is essential.
Setting up a cross-functional data management team should be the first step. The team should incorporate technology specialists, not just from the IT department, but also people who use the services, so that their opinion can be accounted for. Once the team has been assembled, the members’ efforts can be focused on three major priorities:
The business processes must be studied in detail. Data generation points should be identified, all the while having two major goals in mind:
A common set of SaaS tools must be identified that will allow teams to work efficiently. You should be on the lookout for tools that are already in use within the work space, making the implementation process easier. The core apps should then be used as part of mini roll-out strategies, so their widespread use can be promoted. If necessary, the workforce must be adequately trained and motivated. Also, once a decision has been made, employees should be instructed to follow it, e.g. no one should be using Google Drive, if DropBox has been chosen.
Once the new processes have been designed, consider how the SaaS applications can support overall data analysis goals. Do you now have the ability to mine data to get an insight into the employees’ work habits? Can API connections be built between services to smooth the flow of data?
If the company has implemented an enterprise-level business platform, then merging data within a single system can be a viable choice. With that being said, you should be very strict with the choice of data you wish to retain, because bogging down the system with unnecessary data is the last thing you want to do.
Now, you should engage the entire team into brainstorming ways through which the data can be effectively utilized within day-to-day operations. Creating dashboards and making them a part of team meetings is a good strategy to work with as they allow employees’ KPIs to be tracked with ease while offering insight into missed targets. The employees should also be made aware of all these strategies, as once they know that their inputs matter, they will care about data quality.
Everyone in the organization must understand the importance of maintaining the integrity of data and take part in it as an obligation.
Communication plays a major role in user adoption. Within every team, the manager should play an active role in convincing and motivating members. The outcomes of the collected data must also be shared on a regular basis so that employees know the worth of their actions. Similarly, the dangers of problematic data should also be stated clearly, and if necessary disciplinary actions should be tied with it.
Data management should be a key aspect of the company's training programs in order to ensure employee participation is maximal. From time to time, the concerned department can initiate team meetings, workshops, webinars and lunches so that employees get all necessary guidance and coaching.
If user adoption and process alignment are the key goals of your data initiative, then effective maintenance and governance will perform the role of a “how-to” manual. At all times, the data management team should work towards developing an updated roadmap for preserving data quality. The document should be accessible by team-leads, so that they can always have explicit information at-hand whenever they run through a difficulty. Doing so would also ensure that queries don’t pile up on the data governance team, and consistency is maintained.
The guide should be clear, concise and to-the-point. It should incorporate roles for data security, hygiene and strategy, providing comprehensive instructions as to how users should interact with the data. A data security model should also be included in the guide, ensuring that teams have access only to the data they are concerned with.
All of this may sound like a full-time job, and yes, it is. All these issues can be tackled by trained individuals that go by many names: Data Scientist, Business Analyst, Data Officer, and so on. You should never refrain from hiring the required technical resources, as the cost of poor data management can be much higher compared to the person’s salary.