We’ve all heard the term ‘data driven decision making’, but what does it actually mean? The definition refers to the simple premise of basing the business decisions we make on the empirical evidence presented to us through our company’s data. Data is the lifeblood of any organisation and the ability to extract understanding and insight from that data is the currency which makes us competitive in today’s market.
So, what does this mean for the analytical systems that we use? well by extension, the performance, reliability and stability of our analytical infrastructure is absolutely fundamental to our business operations. If we can’t get insight from our data, to the right people, at the right time then we cannot reliably make decisions. It also restricts our ability to deliver products and services to our customers – how long would a data services business last if for example the data products provided were not up to date or could not be trusted.
All of this points to the importance of ensuring that the analytical systems we use are kept in top working order. Consider the cost to an organisation if a critical system were to fail with no backup and no disaster recovery in place.
How can we avoid this? we’ve all heard the term prevention is the best cure, and that certainly applies when it comes to analytical systems. In some sense, by the time a system fails, it is already too late and you’re already up against the clock.
So how do we make sure our systems don’t fail? Well, unfortunately there are no 100% guarantees but what you can do is reduce the likelihood of failure by applying some simple principles. Really, the minimum we should do is service our analytical platforms to keep them up to date, but what we propose you do is go one step further and supercharge. In other words, rather than doing the bare minimum, let’s take the opportunity to dial up our capability so that processes run optimally.
Here are 8 things that you should think about implementing to supercharge your platform for success:
1. Understand your platform: Key to providing an analytics platform that is fit for purpose is understanding how it is used, by whom and for what purpose. Questions you should ask include when are there periods of peak workload, when are there periods where the platform must be stable and therefore be free of changes going into production and when can we expect the platform to be infrequently used to dial back resources – important when running workloads on the cloud.
2. Plan for change: Proactively set up a change schedule that allows hot fixes, maintenance releases, package updates and security fixes to be applied. Consistency generally works well for users and sets expectations around platform availability. It also means you lower the risk of exposure to vulnerabilities caused by security events.
3. Monitor performance: Don’t wait for users to complain that processes are running slowly. Proactive monitoring of platform performance extrapolating trends to anticipate safe thresholds being breached will help you alert your administrators to take action before problems occur. CPU, memory, I/O throughput and disk usage are all examples of metrics that should be monitored.
4. Secure your platform: Make sure you have the right RBAC and ABAC policies in place to authenticate and authorise access to analytics. Monitor access to the platform to ensure individuals only have access to the information they are permitted to access. Make sure you have streamlined onboarding and offboarding procedures in place.
5. Housekeeping: Allow diagnostics and log checking to highlight platform housekeeping tasks, to keep your analytics environments in good working order. Archive redundant content and make sure you have good version control mechanisms in place to ensure users are collaborating on the latest versions of processes mastered within a source control repository.
6. Optimise workloads: Processes may start off efficient but rarely stay that way. Requirements change, data volumes increase all of which can lead to a degradation in workload performance. Keep a regular review of batch process run times, perform regular reviews for bottlenecks and review concurrent activities to see if there is contention for resources that require prioritisation. If run times are increasing, consider what can be tuned and optimised – database, hardware or are there more efficient techniques to utilise in code.
7. Invest in your teams: Your analytics engine may be in tip-top condition, but how is it being used? are there skills and knowledge gaps preventing your user base from getting the best value from the capabilities available. Are there inefficient processes being maintained, bad practices being followed which directly lead to poor performance or unreliable results? Continued education and support for your users is every bit as important as keeping your analytics platform in good working order.
8. Modernise, don’t compromise: Keep abreast of the latest software versions, is it time to upgrade? It may be that simply continuing to update your platform isn’t going to get you the productivity boost that you need. The longer you avoid modernising the harder and the more expensive it becomes for an organisation to make the change.
At Katalyze Data, we believe all of these tips are fundamental to success – it’s about the people, the process and the technology. Want to know more? Please reach out to the Katalyze Data team for a free, no obligation consultation on how our platform health check services can accelerate your business.