Flight Centre has restarted its use of enterprise observability as a way of keeping its travel consultants selling and serving customers by more proactively managing the underlying tools and systems that support them in that work.
Chief information officer Chris Locke told iTnews that observability software, primarily ThousandEyes, was used pre-pandemic but mostly for reactive root-cause analysis of issues that consultants were already impacted by and complaining about.
The use of observability tools was “stopped” in the pandemic as part of broader cost-cutting efforts at Flight Centre.
With the resumption of travel, it has since been restarted but with a different purpose: to allow IT to move from a reactive to proactive capability when it comes to incident detection and response.
Proactive management of the IT environment is particularly about keeping travel consultants productive and available to serve customers.
Pre-pandemic, Locke said, if a single store’s systems were down or underperforming, “it wouldn't cause a lot of impact within the organisation” - owing to its size and number of employees.
But Flight Centre today is much smaller than it once was - both from a customer-facing and backoffice support perspective - and there is much more pressure to keep consultants productive all of the time.
“With the pressure that consultants are under at the moment, with the queues outside our stores and corporate travel returning massively, every consultant working is a massively important thing for the business, so consultant productivity is huge,” Locke said.
“It’s important every travel consultant [is] selling travel every day of the week.
“We've seen increases of about 50 to 80 percent in their productivity and the value of the sales they are committing over a month at the moment, so having every one of [them] busy and delivering travel is …absolutely critical.
“The ability for us to use a product like ThousandEyes to keep on the forefront of proactivity and keep those consultants working is massive.”
- This story is part of an iTnews Focal Point on enterprise observability. Click here to read and watch more.
Locke said the company is particularly using ThousandEyes for synthetic testing of how consultants experience software-as-a-service (SaaS) platforms that are now in use.
The tests had helped to identify and troubleshoot issues with a telephony platform used across “a couple of regions”; Locke said the tests had been kept running to monitor conditions that could be in breach of the service level agreement (SLA) with that provider.
“The minute we're seeing a breach of SLA, for want of a better term, we're able to hop straight on that problem and start triaging and working through the problem, prior to getting [complaint] phone calls [from users],” he said.
Flight Centre is continuing to use observability software to hold its vendors and providers “slightly more accountable than they were previously.”
“Where we can agree to a number of measures around a certain solution, we then track that in conjunction with a couple of our key vendors,” he said.
“In some instances, we're tracking those I won't say without their consent, but without their agreement that's how we're going to measure the service level agreement.”
Locke said the company had no interest in tracking SLA performance to claim service credits from providers.
“We're not using it to drive the return of service credits,” he said.
“We're using it to ensure that the platforms are up and available for our consultants.”
Locke said observability had also made it easier to pinpoint problems in a heavily SaaS-based environment, where there are many moving parts that make up the end-to-end connection between the user and the system.
He cites a recent issue with “a large SaaS solution that we use globally that needs to be hosted overseas from a GDPR perspective.”
“Based on it being in Europe, and a lot of our users in Australia, the network gets blamed straight away from a latency perspective,” Locke said.
“As it turned out, after doing some testing using ThousandeEyes, the problem was all within the codebase [of the SaaS application itself], so we were able to put the network [and] the end user devices out of scope very quickly [as the root cause of underperformance] and pinpoint where the problem was.
“I think we get a lot of value out of ThousandEyes from a mean time to resolution (MTTR) perspective. What used to take weeks is now down to hopefully only a couple of hours being able to pinpoint that problem, put in place the fix, and get users back using [these systems].”
Locke characterised Flight Centre’s observability status as being on a “maturity journey”.
The next phase of that is likely to be increased use of automation - such as automated responses to the detection of certain underperformance conditions, like spinning up additional storage capacity.
“I would say that's something we'll be tackling over the next six-to-12 months,” Locke said, adding it is budget dependent.
Future work could also explore how to harness and manage the large amount of telemetry data being generated by the observability environment.
Finding its place
In restarting observability and elevating its stature within IT, Locke said one of the biggest challenges has been in “finding the right home for it” as a discipline and function.
“We struggled early on with someone to own observability from a broader IT enterprise perspective,” he said.
“From a triage perspective, I think, we managed to find owners very quickly in our network, cloud and infrastructure teams. They’re very focused on how to get the most out of those sorts of tools, but they have a very reactive sort of view.”
At present, observability is housed in “service operations”, which effectively has responsibility for level two service desk and ITIL practices.
“It's found a really good home there at the moment,” Locke said.
“They're the team who really are focused on the true outcome for our internal customers.
“As a shared [IT] service provider [to the Flight Centre group], we refer to our internal travel consultants as our ‘customers’, and this team is really focused on that.”
In other companies, there has been a trend recently towards specific observability leads - as evidenced by a search of LinkedIn titles.
Locke liked the idea of having observability housed in a specialist domain area, with a domain lead, but said financial pressures on his industry persist.
“It's just, where do we squeeze that into a budget?” he said.
“Two years of half-billion dollar losses kind of hurt. The [IT] team shrunk from about 370 down to about 120 at one point.
“We're growing that back slowly. We won't get it back to where it was, but I guess it's finding the right niche for [observability] and the right amount of budget to make sure we're getting the value out of the product set without having to have a full person on it at the moment.”.
