Metrics and management: Best practices from engineering heads on how to lead effectively

As technology’s role in our everyday lives becomes increasingly important, so too does the work of the engineers that build and maintain the systems and platforms that bring these solutions to us. And for engineering leads, the ability to effectively manage not just the technical aspects but also the people under them is crucial.

In a recent closed-door online panel discussion hosted by Tech in Asia and co-moderated by Amazon Web Services (AWS), a few of these engineering leads came together to discuss the best practices of managing complex tech stacks and the teams that run them. The panel included AWS principal solution architect for startups Supreet Sethi, Tokopedia vice president of engineering Manoj Awasthi, , and CXA Group head of infrastructure Riyad Twair.

Measure effectively

To make sure that things run smoothly in an organization, keeping an eye on metrics is important, Twair said. But that doesn’t mean everything needs to be measured. “It’s a matter of figuring out which metrics work best for which situations,” he shared. “I split them up into metric categories rather than [look at] specific metrics.”

In order to keep improving their products and solutions, organizations are also constantly evolving their tech stacks. Throughout these changes, the user experience needs to remain seamless, Twair emphasized. Some ways to do this according to the CXA Group exec would be to monitor for specific “heartbeats” and have alerts for when something’s going wrong.

To that end, it’s paramount for a platform to keep track of its uptime – or its reliability and availability. But for an organization like Tokopedia, which runs over 300 microservices that work in tandem to provide the user experience to customers and merchants, keeping track of all the technical metrics is difficult and time-consuming.

Manoj Awasthi, vice president of engineering at Tokopedia / Photo credit: AWS

“One good alternative we found that was effective was using a business metric as a proxy,” Tokopedia’s Awasthi explained. “And when that business metric is below a certain threshold, that’s counted as downtime.” Some examples of metrics to track include the expected total gross merchandising value or number of orders on the platform at a given time of day.

Twair agreed on this point, taking his past experience working at ecommerce giant Lazada as an example. “Sales would drop for a five-minute timeframe, and we wouldn’t see any issues at the infrastructure level, but obviously something was wrong,” he said. “That would trigger an escalation, and we’d start an investigation just based on business metrics.”

That said, when it comes to determining thresholds and acceptable boundaries of performance deviation, Twair said that it’s simply a matter of preparing beforehand with lots of trial and error – multiple load tests using simulated traffic help to determine acceptable limits.

Hiring and development

A tech stack is only as good as the team working on it, and for both Awasthi and Twair, hiring for culture fit is a best practice to stand by.

“Skills are important, and there needs to be a baseline for that, but throughout my years in this industry, I’ve seen a lot of people with very good technical skills that can’t quite work in a team,” Twair shared.

In order to combat biases in the hiring process, both Tokopedia and CXA Group conduct internal meetings that consist of everyone in the hiring chain. In these meetings, members discuss the questions that were asked during interviews, which guarantees that the assessment of candidates are based on the right grounds.

Companies must also look at ways to develop the skills of its engineers. For Tokopedia, this manifests in a system that offers two paths of progression: individual contributor or managerial. The company also has a mentorship program for the top 10% of its engineers, which gives them the chance to pick up leadership skills.

Taking a page out of Alibaba’s playbook, the Indonesian ecommerce player asks candidates who are up for promotion to present their own case for taking the step up. “We find that this has made people more curious about what their next-level role would be like and helps them to better understand and be able to articulate it,” said Awasthi.

As for CXA Group, Twair shared that employees set their own key performance indicators, which helps them develop along their personal paths.

“It really helps fit things into place – a lot of people who came in as individual engineering contributors are now team leads and they’re doing a great job,” he said. “I’ve even had people go from engineering roles to project management roles.”

Photo credit: Dmitrii Shironosov / 123RF

Dealing with technical detours

Regardless of how fastidious a business is with how it sets up and maintains its platform, sometimes coding shortcuts need to be taken in order to push a feature through quickly. This prioritization of delivery over clean code is known as “technical debt” and, according to Twair, it’s “a necessary evil.”

“It’s very important to have a systematic approach when it comes to technical debt,” he said. “For example, if an engineer has to do something which will require more work later, those tickets should be created on the spot, and the timeline for corrective work should be set before you even approach those tasks.”

In order to prevent technical debt from building up, refactoring – which refers to improving code to build a better structure that allows for enhanced maintainability and sustainability – should be an ongoing activity, said Awasthi. In Tokopedia, 20% to 30% of each team’s capacity is dedicated to taking care of technical debt, he added.

Chronic issues such as recurring bugs also hint that there’s bad design or architecture in place, he added. “You’ll have to actually look at it, as patching will not work,” he said.

“[Engineering teams] need to work with the product team and make sure [these fixes are] prioritized. The impact on the business must be articulated – you definitely want to ship that next killer feature, but not on a broken platform,” concluded Awasthi.

Amazon Web Services is the world’s most comprehensive and broadly adopted cloud platform, offering over 175 fully featured services from data centers globally.

Learn more about how AWS can help your startup scale quickly, reliably, and at low cost here.

This content was produced by Tech in Asia Studios, which connects brands with Asia’s tech community. Learn more about partnering with Tech in Asia Studios.