Case Study

Indeed: Optimizing and scaling Internal Platforms

Abstract

This case study showcases my accomplishments and contributions in the role of Internal Platforms Manager at Indeed, a global job search platform. It highlights the strategies, initiatives, and outcomes achieved under my leadership to enhance Internal Platforms, improve collaboration, and drive organizational efficiency. The study sheds light on the challenges faced, the actions taken, and the overall impact of my work in optimizing Internal Platforms at Indeed.

Introduction

Indeed, headquartered in Austin, Texas, is a renowned online job search platform that connects job seekers and employers worldwide. As the Internal Platforms Manager, I assumed a pivotal role in identifying, developing, and optimizing Internal Platforms to streamline internal processes, foster collaboration, and enhance employee productivity.

Challenges

As an Internal Platforms team, we struggled with a few things that didn’t seem to affect other UX teams as much, namely:

  • Our work was only seen internally, which affected
    • how the impact of our work was perceived, and that, in turn affected
    • promotion cases.
      Note: promotions needed to be put forth by managers and were evaluated by a panel of Directors and VPs
  • Some people would see it as an opportunity to transfer to other teams with externally-facing work to have a better chance at getting promoted.
  • Our budget within Internal Platforms to hire additional team members was tied to the needs of Internal Platforms as a whole.

Team Growth

Initial Team Make-up

Our small but mighty team at the beginning of my time with INTUX (Internal Platforms UX) was composed of a few roles: research, design, and content strategy. Most of the designers reported to me and some to my manager. Research had their own manager. There was a content strategist with a dotted line to our team; in essence they were embedded. Our team also had access to a development team in order to accelerate the build of some of the projects we would work on.

As you might imagine, this made streamlining our working processes a little cumbersome. There were lost efficiencies in not having everyone working as one team.

Inheriting Teams

As the managers of content strategy and development moved to other teams and got promoted, the team members that reported to them began to report to me. Our already established team, having had a good working relationship with the soon-to-be INTUXers, welcomed them with open arms.

Expanding Our Influence

As our team grew in capabilities, we were able to create efficiencies in how we communicated and how we worked. This allowed my manager and I put together a capabilities document that allowed us to pitch our services to a variety of internal teams in order to grow our influence and strengthen existing working relationships. As more teams learned of our capabilities, the desire to work with INTUX grew, and so did the demand for our services. We were able to address the needs of plenty of our partners in Austin and Seattle. COVID put a stop to our trip to Tokyo.

Our soon-to-be partners were working on interesting projects that had the potential to make lasting impact at Indeed, aligned with out OKRs, and would provide our newly assembled team the opportunity to help build and polish some standard operating procedures. These SOPs we would eventually document and formalize into an INTUX Handbook, which would serve as our go-to manual as well as an onboarding accelerator as we built out the team further.

Working Together

A growing team brought with it some additional challenges. We would need to more closely address how we worked. To that end we worked on:

  • a team handbook (as mentioned above), which outlined who was on our team, our meetings, how our asynchronous stand-up worked, which tools we used, what teams we interacted with most and tons more.
  • an internal way to resolve conflict. Our team participated in an internal training based on the book Crucial Conversations. Having established a common language to resolve conflict meant less time spent resolving disputes between reports and more time helping them figure out an approach to do it themselves.
  • hiring additional management. As we hired more individual contributors, we also needed to hire more managers to be able to handle a growing team. This helped INTUX each more closely partner and address the needs of each of the platforms we served, and it also helped our clients feel like they had an embedded UX partner for their projects.

On the positive side, a growing team also meant we had more clout. More clout meant we could more easily show progress, impact, and help get our team get the promotions they so deserved.

Advancing Careers

Our team held weekly one-on-one’s between manager and report. These were short, half hour meetings. This meant that there were small touch points every week where needs, wants, grievances, and goals could be addressed. This also meant that as a manager I could coach my team to success and promotions and help them make small adjustments while measuring progress week after week.

Promotion panels met every six months, about two to three weeks after evaluations. Most promotion cases needed to be written well in advance in order for the promotion panel to have a chance to review, and, in full transparency, for INTUX management to help show the real weight and impact of the work before the panel was able to meet to decide on who was ready for a promotion.

Before leaving Indeed, everyone who reported to me that was working towards a promotion, got a promotion.

Sample Projects

Peregrine

Peregrine is a JIRA velocity analysis utility. It provides insights and actionable recommendations for teams to improve their engineering and product velocity.

The MVP helped answer three main questions:

  • Delivery lead time (DLT): How many days from starting work on a ticket to closing it?
  • Deploy throughput: How many production deploys per team per quarter?
  • Experiment throughput: How many Proctor (internal tool used for A/B) tests complete per team per quarter?

Key Results:

  • Standardized velocity measurement dashboards are used by 80% of all teams.
  • Delivery lead time average < 7 days in prior 12 weeks (Nov ‘18 baseline: ~13.5)
  • Deploys/team >= 85 in prior 12 weeks (Nov ‘18 baseline: ~57)
  • 4x as many experiments completed using features in Product Platforms and Front-End Core products in 2019 vs. 2018.

Alexandria

Indeedians use Alexandria, the company’s data dictionary, to find documentation about data sources. When the Alexandria product team learned of usability issues through discovery research, they partnered with INTUX to improve the user experience of the product, beginning with search functionality improvements.

UX Design, Research, and Content Strategy worked together to simplify a search experience that wasn’t meeting users’ mental models. The team replaced confusing dropdown filters with a new filter modal that allows users to narrow down results without having to learn Elasticsearch syntax. And they introduced helper text and placeholder text to provide users with clear guidance.

The new search functionality also boasts a more robust autocomplete feature inspired by Google Drive that organizes results by data source type and includes associated metadata.

Key Results:

  • Consistent increase in user satisfaction score on Alexandria. In a Data Platforms CSAT survey, Alexandria was the highest-scoring tool (3.52/5) from all 17 tools listed in terms of user satisfaction, up from 3.48 in previous quarters. In the following quarter’s survey, Alexandria tool satisfaction continued to increase to 3.7/5.
  • Quotes from CSAT:
    • “I think many steps have been taken in the right direction already - training opportunities, the Alexandria revamp, standard index filters in IQL1 - I really appreciate all of this work.” - Job Seeker user
    • “Alexandria is extremely helpful, and I appreciate the communication when IQL indexes are being deprecated.” - Client Success user
  • Weekly full-time employee usage of Alexandria for the past two quarters increased by 2.8% from the previous two quarters.

TestStats

The INTUX team partnered with a team based out of the Tokyo office to improve TestStats, the internal experimentation tool that helps Indeedians analyze Proctor tests (internal tool used for A/B testing).

One initiative focused on exposing the allocation history of an experiment, allowing users to make informed, efficient decisions on tests. UX Research performed concept testing validation that indicated users found value in seeing both list and graph views of allocation history, as well as time frame options. The INTUX team used that feedback to create a side pane experience for an experiment’s allocation history.

Key Results:

  • 114 PMs have used this new feature during the quarter after release
  • Highest adoption rate seen among Job Search and Search Quality teams
  • Consistent QoQ increase in TestStats usage by SMB (Small and Medium Business team) and Job Seeker PMs
  • "To me it's currently the easiest way to quickly go through different allocations and [see] how the percentage changes across time. Otherwise, you need to go to Proctor and it's not very easy to view those...As far as I know, [it’s] the easiest way to look at the allocation history." - Product Manager

Empowering Impactful Digital Experiences

I bring nearly two decades of experience building and optimizing teams and digital products. I’ve led design teams for big brands, creative agencies and academic institutions. Most recently, I led the Internal Platforms team at Indeed. Prior to that, I helped put together and lead a separate team at Bloomberg, where we built and improved some highly influential business tools. These last few years, I've focused primarily on working with public benefit organizations, non-profits, and other mission-driven organizations looking to make a positive impact in the world. Oh, and making sure my dog, Kona, is living her best life.

Ready to get started?

Call or email to book a free 15 minute consultation.