Imagine you’re in a department store. You’ve roamed the aisles and filled your basket, and after queuing to reach the till, your transaction can’t be completed. The cashier is having a bad day so tells you go back to the end of the queue, and try again later. Let’s be kind and say they allowed you to keep the contents of your basket, but after queuing again, the same thing happens. Before you know it, you’re dizzy and ready to leave the goods behind.
Of course, this would be unlikely in real life. The duty manager would look over the balcony and quickly spot a brewing commotion. But machines can have bad days too. How does this translate in the virtual world of retail? No humans. No duty managers. No balconies. The cashier is just code running as a series of micro services; running on docker containers; running on a virtual machine; running on a hypervisor situated in all corners of the earth; overlooked by an anonymous somebody.
Let’s take a look at how one of the UK’s largest department store, John Lewis deals with it, and how it uses Splunk to watch over that cyber balcony.
Paul Adams has been in IT Ops for a long time. He’s part of the consultancy company Octamis but spends the vast majority of his time helping John Lewis in its operations. I’ve known him for over five years at John Lewis, and can vouch that he’s the poster child of a new generation of data-thirsty operations gurus.
He recently presented at Gartner IT Infrastructure & Operations Management in Berlin on behalf of John Lewis, on why machine data is so critical to how the organisation approaches its IT operations. Why? Because only with this data can great customer service be guaranteed.
Splunk is central to analysing John Lewis’ data. Across the many different applications and components that make up John Lewis, Splunk is what ensures the machines avoid having bad days, and guarantees that every single customer has a seamless experience online. Paul calls Splunk his “macroscope” over all the other monitoring tools they have.
"Sometimes the system errors with a smile on its face”, which is Paul’s way of saying there isn’t always an error generated. You have to analyse data constantly, just like a duty manager walking the aisles in a physical store. At the Gartner event, he gave the example of how John Lewis created a checkout flow process using Splunk ITSI. It holds KPIs of every single customer going through the different steps of a checkout process. It’s a notoriously complex process, as a lot can go wrong. Some of it can be in the control of the retailer, and some dependent on third-parties. Error or no error, issues cannot elude Paul - he can see everything customers are doing.
Before you question if constantly asking questions is feasible, with Splunk ITSI it is! It allows easy mapping of the different components that make up John Lewis’ service, and quickly creates KPIs based on raw data. Then, ITSI does the hard work by applying adaptive thresholds using machine learning algorithms, baselining what looks good and bad at specific times of the day, based on past trends. Splunk ITSI helps find performance anomalies within those KPIs, and provides dashboards similar to Powerpoint presentations thanks to Glasstables. This takes more workload off of Paul, allowing him to concentrate on improving the service and customer experience further, which includes looking beyond the web and into fulfilment and business optimisation.
This data-centric approach to customer service is why John Lewis is cherished within the retail market, and are able to maintain such a great brand in 2017 as they did in 1867.
If you want to discover out how to do this for your own organisation, register for our webinar on June 29th, “Earn a Seat at the Business Table with Splunk ITSI”. You can also view the full presentation from Gartner IOM 2017 below: