Finding a needle in a thousand haystacks - How to fix customer experience

20 years ago I worked for Accenture at the London Stock Exchange. A few months before I joined, the Exchange suffered a catastrophic outage that delayed trading for 8 hours. The impact was hundreds of millions of dollars lost. The fall-out from that day was still reverberating months later, as we desperately tried to repair the lost goodwill and reputation harm.?

What’s changed since then? Well…everything and nothing.

In the year 2000 there were no hypervisors, no micro-services and no containers. Applications were monolithic and servers had names like pets. Complexity, scale and the rate of change of technology systems is now exponentially greater and amplifying every year.?

Over the past 12 months alone, virtually every major exchange on the globe suffered some form of outage or performance degradation, meaning that delivering stable, high performing and engaging digital platforms is more complex now than at any other time in history.

When things go wrong, we are no longer looking for a needle in a haystack - we are looking through a thousand haystacks to find that one needle.

But introducing stability into platforms is just table stakes. Modern digital businesses are also grappling with challenges such as…

• Understanding customer journeys
• Faster and higher quality software delivery
• Optimising for core web vitals
• Delivering simpler more streamlined operating platforms
• Developing an agile adaptable workforce.
• Meeting SLAs via faster issue identification and remediation.
• Scaling to biggest days
• Modernisation of applications
• Shortening and enhancing testing cycles
• Cost optimisation/tools rationalisation
• and migration to cloud.

These challenges are almost universal and require a well considered observability strategy and a single source of truth for telemetry data if they are to be executed successfully.

But many large companies have between 20 and 40 monitoring tool sets, which can cause a loss of visibility and delays due to context switching; ultimately tying up teams and curbing their ability to innovate.

Things typically evolve this way, because traditional vendor pricing models have prevented organisations from instrumenting everything, resulting in piecemeal monitoring and tool sprawl.

So, how do we solve these issues? And how do winning companies treat digital delivery, and avoid performance degradation and outages??

Creating a unified team

Successful Digital Businesses focus their teams on the two millimetres of glass that sits between them and their customer. I’m referring to the glass of the mobile phone, tablet, smart TV or laptop.

It’s the most valuable real estate in the world, because it provides a window through which customers interact with any business.

A business and its engineering teams should be relentlessly focussed on delivering a strong customer experience via these screens - creating engagement, conversion and brand affinity. They must also be focused on innovation and growth: integrating customer, business and operational data, so that engineering teams can focus on high impact work, experiment and drive competitive advantages.

For IT Operations and site reliability engineering teams, the focus must be on ensuring uptime, performance and reliability. This can be done by introducing stability into platforms, and proactively identifying and resolving issues before they impact customers. Once burning platform issues are resolved, teams are then free to focus on creating operational efficiency, while building a modern digital business that is able to realise the benefit of their cloud investments faster.

A new approach

Up until now, almost every organisation has been forced to compromise on their observability strategy. They have had to leave parts of their environment un-instrumented or to choose best of breed or best of cost solutions. It’s time for a new approach.

The lineal relationship between inflating observability costs and increasing scale is not sustainable.

Organisations shouldn't have to count hosts or purchase shelf-ware in order to negotiate better unit prices. The future is a consumption model where organisations pay only for what they consume.?

Instead of being charged more for platforms as they expand in size and complexity, pricing models should align to engineering practices.

No more host counting. Instead pay only for the users of your observability platform and the data that’s ingested.

If your team wants to foster a high performance culture where a single engineer can manage 20 services rather than 2, or 100 hosts rather than 50, the consumption model should reward you for that.

The pay off when organisations get their observability strategy right is inevitably improved, uptime performance and reliability, cost savings, greater developer productivity and more engaging and profitable user experiences. It’s a worthwhile pursuit and at last it can be done without breaking the bank.

Chris Marshment is the Director, Sales Strategy and Global Accounts (APAC) at New Relic. The views and opinions expressed in this article are his own and do not necessarily reflect those of New Relic.