Getting value from software observability

Most top executives understand that delivering great software-based experiences to customers is table stakes for business success, but they don't appreciate just how much business value is under the hood.

It's about capturing constant, real-time, highly actionable insights on the performance of a digital business and being able to answer to the business on key questions, including:

• Has a slowdown in the responsiveness of a web-page caused prospective customers to wander off to a competitor's site?
• Does changing a "buy now" button from red to blue cause sales to increase in Sydney, or decline in Auckland?
• Does a surge of consumer requests for a new kind of service indicate a major new business opportunity?

Technologists refer to this ability to look under the covers of software in production as "observability.

It's not a new concept; the application performance monitoring industry is two decades old.

But until recently, this data was used almost exclusively in an IT context, often to figure out why an application was running slowly or why a website had crashed.

Not anymore.

With software increasingly at the core of how companies interact with customers, employees and partners, major organisations such as ANZ Bank, Australia Post, the ABC, Culture Amp, Service NSW, and Xero now routinely mine software performance data to achieve ambitious business results.

Executives at companies like these often say they feel as though they've been given the gift of sight, instead of operating in the dark.

Previously, companies tried to get this insight from big data analytics.

But this might take weeks or months, and then longer to figure out who should fix the problem.

By building observability into the software delivery infrastructure, the data is captured as a matter of course and is readily available to relevant parties in a context that makes sense to them.

So what's the best approach to gain observability and unlock maximum value from the cloud?

Here are some hallmarks of the most successful efforts:

No tech silos. To be able to respond quickly to software performance data, software development teams need to work in close coordination with the IT operations teams responsible for delivering their applications to customers.

Move to more CX-centric KPIs. KPIs around uptime targets are no longer good enough. Leading IT shops are moving to metrics with direct impact on the customer experience and on the health of the business, such as customer-engagement or sales per visit.

Get a panoramic view. In today's multi-channel world, it's crucial to maintain the customer experience across all of your digital touchpoints. That means having observability into everything that impacts that overall experience - from network performance to digital experiences - and having that information visible in one place for everyone.

OKRs more than OK. Rather than plan around the best hunches of top executives, the Objectives and Key Results (OKRs) approach relies on more frequent analysis of what's possible by the teams that best understand how to turn opportunities into measurable results.

At the end of the day, delivering software is becoming a more crucial aspect of business, and more complex.

Performance is critical.

And if you're not doing all you can to get observability: those constant, real-time, highly actionable insights, then your competitors will be.

Ken Gavranovic is New Relic's product senior vice president, where he helps shape the strategic direction of the New Relic platform and supports customers as they migrate their organisations to modern technologies.