It’s been more than 3 years since I joined LinkedIn in the summer of 2021. In that time, I’ve moved from the Business org to Engineering, gone a level more technical from Business Analytics to Data Engineering, served as the SF Office lead for LinkedIn Asian Alliance, and worked with dozens of teams and colleagues across the world. As I look back on my time at LinkedIn, here are a few things I’ve learned:
Write things down
Technical skills are great, knowledge transfer is paramount. Writing things down, documenting decisions and why they were made, and having an accessible wiki/doc/platform to explain things in more detail save hundreds of hours otherwise wasted trying to piece together the past.
Documentation looks like different things depending on the org and project at hand. For big initiatives, you might spin up an entire wiki with multiple pages, links to other docs, timeline trackers, and more. For a single dataset, you can write a RFC (Request for Comments — initial doc outlining a proposed build), create an ERD (Entity Relationship Diagram), and create a table explaining each variable. Ongoing adjustments to production code have history stored through the publishing mechanism (often GitHub).
As companies adopt more intelligent search for intranet data and internal documentation, these records become even more easily accessible and help quickly transfer knowledge to new owners over time, minimizing miscommunication.
Urgent vs. Important
How do you prioritize what to work on today? My former Director recommended a simple “Urgent vs. Important” matrix to get it done. The logic is simple:
- Urgent and Important: Do. These are priority tasks, take action now!
- Urgent and Not Important: Delegate. Pass it on to someone with more time or interest.
- Not Urgent and Important: Delay. Schedule it for later, after top priorities are done.
- Not Urgent and Not Important: Delete. Don’t waste time on busy work.
For example, if I have a one silver layer table to build, an urgent ask from a partner team to fix a broken dataset, a data ingestion flow that needs an adjustment for one variable, and a vendor connect for a LinkedIn Asian Alliance vendor fair, I can review each and then categorize them as:
- Urgent and Important: Silver layer table. I need this table to finish a downstream table that 100+ users are expecting by the end of the month to track $1B of bookings data from one product suite. This should be top priority.
- Urgent and Not Important: Vendor connect. This is relatively “less important” since it only impacts a local office, and I can delegate to someone on my cabinet to manage the connect instead.
- Not Urgent and Important: Data ingestion flow. It’s used by dozens of stakeholders, but the variable in question isn’t going to be used until a downstream team has the capacity to build a dashboard related to it in 6 months. I’ll keep it on my list for now and if I have downtime (or if we start getting closer to the deadline), I’ll work on it then.
- Not Urgent and Not Important: Urgent partner ask. Turns out that the “urgent ask” was just a misunderstanding on the use case of the dataset. Remaining work will not be prioritized because it won’t actually solve what the partner wants.
Act as One LinkedIn
LinkedIn has a set of internal values, and “One LinkedIn” remains an enduring phrase on the list. It’s a reminder that we are all colleagues of the same company and we’re all working towards the mission of “creating economic opportunity for every member of the global workforce”.
I’m not saying that everything perfectly aligns all the time, but I’ve found that refocusing on what’s best for company, customers, and colleagues is a great way to steer tense conversations back to productive dialogue.
This value, expanded more broadly, is very important in work and conversations. Figure out the common goal and work from there to create useful action.
Andrew Jabara 