Social matters. We now live in a world where social media streams span an entire spectrum of content from consumer-level entertainment and frippery, onward to bona fide news channels and enterprise-level corporate messaging… and even outward to government-level broadcasts, all made in text, image and video moving picture format. Because we have so much social media in our lives, it has now evolved to offer different grades of membership, user status and service.
Firmly established at the business, workplace and employment end of the social media spectrum is LinkedIn. Acquired by Microsoft in 2016, the site runs without any blatant use of branding from Redmond and operates from its own headquarters in Sunnyvale, California.
Like any social stream, there is a proportion of white noise, time wasters and ne’er-do-well types populating the content. The site itself operates various filters at the backend to cleanse a degree of this and of course, offers LinkedIn Premium for those that wish to pay for more control on the platform itself. How then does this element of the LinkedIn total offering get built, how does it run and what do we need to know about its internal mechanics if we’re thinking about engaging with it more directly? One person who knows the answer to these questions is Prashanthi Padmanabhan in her role as senior director and head of engineering at LinkedIn Premium.
Why, what… when
Explaining how LinkedIn mechanics are architected on the inside, Padmanabhan says her team’s work is primarily driven by three main elements of LinkedIn’s engineering principles for platform development: start with the why; align on the what; optimize for when.
“The why factor in product development is all about avoiding adding technology for the sake of technology. As engineers, we know there are multiple ways to approach a given problem or task. Understanding the why is extremely important as that can lead to specific technology choices. This can be anything from analyzing the fastest way to get a new product into the users’ hands to looking for the most resource-efficient way to maximize value and enhance productivity,” said Padmanabhan.
But even with good intentions, building software products that users actually want to interact with in their day-to-day experiences can be tough. Being realistic, Padmanabhan admits (like any software engineer) that the coding function can get removed from the customer experience, falling into the trap of thinking that product managers, UX designers and marketing teams are the only ones who need to engage with users. But, she says, for engineering teams, hearing from users about their needs, pain points and feedback on the products is far more powerful than any product requirements specs.
Detailing her own special experiences at LinkedIn, she details opportunities to be one of the engineers invited to member feedback sessions to better understand the business and financial implications of technical decisions. After these sessions, the team’s engineers have worked to proactively prioritize issues that need fixing or go the extra mile to solve problems just because they heard or saw how those issues impacted the member in a real way.
Beyond bugs, jumping Jira!
According to Padmanabhan, “That relevant knowledge is so much more powerful than reading a bug description in a Jira ticket! Once we build a minimum viable product, we start with a limited test with a small segment target audience. This test is where we assess the product performance, review user feedback and sentiment and continue iterating the product. Involving and educating engineers in this process of user research, persona definition, member feedback review etc. is a great way to keep the products we build simple, intuitive and trustworthy.”
Pleasingly (for those of us who spend a good amount of time on the platform), the LinkedIn team say that they emphasize opportunities to foster collaboration among engineers. One way Padmanabhan tries to level-up the engineering team is by conducting sessions known as ‘Demystifying the Business: 101’, where the company brings together all elements of the business – including finance, user research, product marketing and data science teams. These sessions help build business awareness, foster cross-functional collaboration and promote data savviness within engineering teams.
“When I took over the LinkedIn Premium team, I partnered with the business operations and finance functions to do one of these sessions. The goal was to help the team connect the dots between their work and the business outcomes, which engineers care a lot about. Also, being business-aware and data/metrics-savvy empowers engineers to prioritize their product roadmap to optimize for the right customer value and business outcomes,” said Padmanabhan. “It’s all about creating an environment where everyone understands our customers, stays aware of business dynamics, thinks critically, experiments rapidly and keeps learning.”
Technical chops + business savvy UX flair
To excel as an engineering leader, she insists that relying solely on technical prowess is not enough. While strong technical chops are really important, they must be complemented by solid business acumen and a sense of product and user experience (UX) flair.
“Imagine a Venn diagram representing the multifaceted responsibilities of an engineering leader. In one circle lies the technical strategy, architecture and roadmap – essentially the ‘how to build’ aspect, where they traditionally excel. However, the other circle encompasses equally vital skills: business strategy, product and user experience roadmap, AI/data strategy and go-to-market approaches. These interdisciplinary skills are the bridge that connects technical excellence with business impact. By developing this holistic skill set, engineers become more than technical experts; they transform into cross-functional leaders who can meaningfully drive business growth,” explained Padmanabhan, in a definitively positive voice i.e. she clearly takes this central approach seriously.
Having covered a lot of the why and what, the team’s when factor is drawn from a focus on the fact that all product features first launch as experiments that ramp to a small segment of members. They then progressively expand towards the full member base based on experiment results and learning. This experimentation infrastructure and tools help run multivariate tests and monitor performance against key product/business metrics. Not just first-order impact, but also downstream impact to extended parts of the product ecosystem.
“When rolling out the updates to LinkedIn Premium, we wanted to move quickly, however, we also wanted to get this right. They might seem contradictory initially but reconciling them is building a culture of rapid experimentation grounded in data-driven decisions,” said Padmanabhan. “In practice, this meant we made product decisions based on our best judgment, measured the feature’s performance – and then were completely willing to adjust if the data showed us a better path forward.”
For example, when the team started experimenting with the LinkedIn AI-assisted message drafting tool, it had target metrics for feature adoption, the acceptance rate of suggested drafts etc. They iterated on the product experience and the underlying prompt engineering until they achieved those target metrics before broadly rolling out the product to the member base. Meeting these metric goals gave them confidence that the initial product works as envisioned in solving member needs. This is the essence of Agile product development i.e. put a minimum viable product in front of a focus group of users to assess product market fit and performance, then incorporate feedback so software application developers can rapidly iterate and improve and pivot when and where needed.
Day one is day zero
“I often tell my engineers that the day we put the product in front of our members is day zero [with day two to inevitably follow] of the product development journey! This is also why the time to market is a critical vector to optimize for – don’t let perfection get in the way of unlocking early learnings,” concluded Padmanabhan. “As product builders, we shouldn’t shy away from involving our users and embracing their feedback and sentiments early and continuously into our product iteration process. Ultimately, user feedback is a precious gift.”
With Twitter becoming X and TikTok’s approval in some territories (repeatedly) coming into question, many professional social media advocates, users and evangelists are watching the social media space very carefully. While social media-focused enterprise technology influencer powerhouse players such as Evan Kirstel exist across pretty much every platform in much the same way they have done for the last decade, the next big thing is generally always five years around the next corner, so it’s worth considering how LinkedIn is evolving in the period to come.
