The Great Resignation: Reducing the Risk of Data Loss

We’re excited to bring Transform 2022 back in person on July 19 and pretty much July 20-28. Join AI and data leaders for insightful conversations and exciting networking opportunities. Register today!

The world of work is changing drastically. People have embraced telecommuting and hybrid work, job openings and salaries are approaching record highs, and layoffs are at record levels. Navigating this volatility can be challenging for companies of all sizes. We all know that high employee turnover can negatively affect the continuity of service to customers and slow down operations. However, the departure of staff increasingly brings with it an even greater danger: data loss.

Data loss can mean that vital information is physically lost: for example, if it was incorrectly stored on a deceased staff member’s personal device; lose knowledge of how to access, collect or use data; or lose consciousness that the data really exists. It can cover everything from passwords to customer information and marketing databases, to source codes, developer documentation and other critical parts of business intelligence. The portability and abundance of data coupled with the decentralization of teams via remote working means that the risk of catastrophic data loss increases with each team member who leaves.

Damage may not be limited to loss of information only; there’s room for both business functionality and brand reputation to take a serious hit. Not to mention that data protection laws are being strengthened around the world, creating a very real risk of legal trouble.

Protecting your business from these issues basically means a company reassessing its relationship with data. The first step is to recognize that vulnerabilities exist.

Legacy architecture and data management platforms

Many readers will be shocked to learn that many of the world’s largest companies are using extremely outdated or rudimentary data management solutions. We’ve often encountered organizations that use little more than an Excel spreadsheet to collect and manage some of their most sensitive data. These problems are not limited to large companies; a considerable part of the start-ups also put data management low on the agenda. It’s usually thought of as something to figure out later – “when we’re a little bigger.” However, the reality is that your data management infrastructure is the rock on which your business is built and loss is a very real risk. Complex analytics and information sharing are also supported by a flexible, robust and open data architecture. It must continue to evolve in line with the scale and needs of your business and the advancement of data science techniques.

Taking a piecemeal, or “one and done” approach to planning and investing in data infrastructure, is driving companies loose. It will ultimately be the number one reason why your business could be exposed to so much risk if and when employees leave. That’s why it’s critical to continuously monitor your data infrastructure. The priority should be to ensure that the management and maintenance of your infrastructure is a shared responsibility of the entire company. This means robust procedures must be in place for how information is documented, cleaned, protected and analyzed. By dividing responsibilities across departments and adopting an “always in review” mentality, you can quickly identify where there are gaps in your infrastructure or where systems or policies need to be reviewed. After all, what may work for your development team may be completely inappropriate for your marketing team. This neatly brings me to a major problem that I need to avoid: silos.

Data silos create single points of failure

Information flow can be one of the most pervasive problems a company – large or small – can face. Insights, data and actions stuck in the brains of laptops or systems from one department can be invaluable when shared with another department, but many different factors can get in the way: culture, policy, technical infrastructure and a lack of skills or education (more on that later).

The Great Resignation exacerbates current problems because locked-in data and insights are at greater risk of being lost. All it takes is for one key team member to leave and not accurately document or share critical information.

An immediate way to address this problem is to instruct your managers, IT and HR departments to work together to create a “data exit interview”. To do this, you must first:

Let existing team members self-monitor all information, access, documentation, and other insight they know, have access to, or are in charge of. Combine this information with a company-wide audit performed by your IT team to identify gaps either where responsibility for information is unknown or where information exists that was previously undocumented. Ask your HR and relevant managers to create custom questionnaires for each team member to answer basic questions to answer before they leave about the location, status and access to the data they may have been responsible for or have stored on their devices .

This questionnaire should be taken from an employee well in advance of the last day at the company so that any unforeseen complications can be addressed and your IT team can take a step-by-step approach in handing over controls and access. In that regard, it is critical that access to all data and systems containing corporate data is completely removed.

It’s worth remembering that no data exit interview will be completely foolproof simply because there are cases where people don’t know what they know, or at least don’t think it’s worth sharing. Think about it the same way you would maintain institutional knowledge: questions should be probing and investigative, and all information should be captured and documented in a way that it can be easily captured and shared by your teams. After all, an insight that might not mean much to your HR rep could mean something important to one of your developers.

Break down those (data) walls

Data exit interviews are a quick solution for the short term, but long term improvement and risk mitigation can only be achieved by breaking down all the data silos in your company.

As mentioned above, culture, technology and procedures all play a role in establishing data walls within your company and all three areas need to be addressed. Going into a full silo removal strategy would require a separate, long article – so I’ll quickly summarize some of the key steps that apply to most businesses:

Audit your technology to determine the best tech stack – as we discussed in regards to data management – and extend your audit to focus on all your systems and how they interact with each other. API-driven, flexible, cloud-based solutions that scale with your needs and communicate with each other are often the best choice. Don’t go straight for a mainstream monolithic stack that might lock you in more than you need or won’t go along with your needs. Also, be very careful building your own solution or trying to bend existing systems. Control your data – with the help of outside experts or your own data team, start tracking down where all your information is, the systems used to collect it, the people responsible for managing and analyzing it, how it is updated and cleaned and crucially, what it is used for (or not). Review your policies – how appropriate are your data management, ethics and usage practices? The key is to ask how you can make data a part of your company-wide decision-making policy. Everything from key strategic decisions to the day-to-day actions of each team member.

The final step goes hand in hand with arguably one of the most important changes you can make to protect and improve your data knowledge: education.

Further education, education and training are the answers to the data question

In many companies, using and understanding data is the responsibility of only a handful of people. Not only does this increase the risk of data loss when people leave, it’s also the biggest barrier to a business becoming truly data-driven. Insights have a lower value, employees are not empowered to make their own decisions and innovation is limited to a select number of ‘power users’. It can create bottlenecks and, if senior leaders are not properly trained, lead to poor company-wide decision-making.

Addressing this problem does not mean creating an entire team of data scientists. Training and further education is a broad spectrum and must be tailored to individual team members and departments. The first step is to help everyone understand the basics of data analysis and statistics so that they can interrogate their own data and properly explore insights. Then it’s about giving your team the tools they need to maximize the role data plays in their careers. Technology, managers and procedures must all enable and support this process. Training people once and then expecting them to become data-driven right away is not going to work. It requires a change of mindset where every action or decision, if possible, must be supported by data. Making it a regular part of your business, such as incorporating data skills into the way you assess pay and promotion, can embed this into your company culture.

Take it easy, but recognize the need to act

If you’ve come this far, you’ve probably realized that the risk of data loss is a symptom of much bigger problems facing businesses. The more exposed your business is to losing vital information upon departure, the more likely your approach to data is not fit for purpose. But don’t despair. Recognizing the risks and committing to addressing them is a great first step, and the rewards you’ll reap go far beyond protecting information. The key is to plan your approach and test your projects. Don’t walk into all your data infrastructure and break out or send your entire team on intensive data training. Start with what you can solve every now and then, then experiment with your approach and carefully assess the ROI. It’s much easier to take on bigger challenges if you know what works and what doesn’t for your business and you’ve already seen the benefits of smaller initiatives. Becoming data-driven is not so much a journey with a fixed goal, but rather a change of mindset supported by constant evolution and innovation.

Natalie Cramp is the CEO of Profusion.

DataDecision makers

Welcome to the VentureBeat Community!

DataDecisionMakers is where experts, including the technical people who do data work, can share data-related insights and innovation.

If you want to read about the very latest ideas and up-to-date information, best practices and the future of data and data technology, join us at DataDecisionMakers.

You might even consider contributing an article yourself!