Unblocking a struggling team member & enabling $27M cost reduction.

At Target I founded, built, and managed the team responsible for building a suite of 8 internal Data Science software products that were used by 500,000+ team members & vendors.

This is the story of how I guided a strategic pivot and unblocked a struggling team member (enabled career growth), resulting in $27M productivity cost reduction & 3X decrease in data discovery use cases.

SITUATION

Make Greenfield the data visualization & reporting hub for all of Target's data.

At Target, I founded, built, and managed the team responsible for building internal Data Science software products that were used by hundreds of thousands of Team Members and Vendors alike. There were 8 main products my team and I supported:

Our main reporting and data visualization tool, Greenfield, had the largest current and potential user base. It was where employees and vendors alike came to consume data contained in Target's millions of data tables. They used this information to help them make decisions.

Data Scientists, Analysts, Product Owners, Merchandisers, and many other types of Team Members use a variety of data visualization and reporting tools at Target. Many of these are 3rd party tools the company spends tens of millions of dollars each year to retain access to. The Greenfield team was charged with improving the experience and effectiveness of our platform so that these other tools would not be needed. We wanted control of all systems we used to do business.

Data-informed Strategic Re-direction

One day our Product Director, Rob Koste, came to me and wanted to prioritize a project to make it easier for users to create dashboard cards. This was a huge pain point, but I had a few concerns with focusing here first.

Rob Koste

Director of Product

Data Sciences

"We need to make it easier for people to create dashboard cards"

1. Only 3% of users were 'builders'

When I had the team look more into usage data, we discovered that only 3% of users currently were building cards. Some of this was undoubtedly due to the unintuitive, frustrating experience it was to create a card, but there were hundreds of thousands of users who just wanted to consume the data. We said we wanted to become THE data reporting and visualization platform, but we weren't resolving the real pain of the highest % of our user base.

In fact, the users who had the worst experience were also our biggest active & potential user base.

Even though it wasn’t ideal, our card builders could search our data catalog tool and run queries to find the data they needed to visualize or report on, and even move data from our various data storage locations into Greenfield if it was one of the few datasets that Greenfield didn’t support naturally. The experience was far worst for the non-technical users.

The vast majority of those making billion-dollar decisions were among 97%, and their experience was “very poor”. We said we wanted to become the data visualization platform, but we were ignoring the less technical users who needed us the most and who would be the most critical factor in helping us reach our goal.

2. We were wasting $52M each year in preventable wasted productivity

Because Beginner and Basic / Mid-level users couldn't easily use Greenfield, they were relying on Analysts & Product Managers to help them find data sets, reports, & dashboards they should have been able to find on their own. This correlated to a wasted $52 Million each year that was totally preventable if we only made it easier for people to find the data they needed more easily.

In fact, our Data Analysts and Product Owners were spending more than half of their time just helping users find a data set, or learn how to use Greenfield. There were multiple day-long training classes, a Slack channel dedicated to helping people find data, and a team of trainers that had been set up to help people do things that were too confusing in the Greenfield UI.

3. 85% of analyst teams failed to answer a simple busines question in 1 week or less

In fact, when we ran a hackathon-like internal "contest" to see how many of 35 teams of "mid-level" analysts could answer a data-related business question posed to them, only 15% were successful in the 1 week time limit. If teams of analysts couldn't do it, average business people with even less technical acumen didn't have a prayer. This wasn't only costing the company money in wasted productivity for "not my job" tasks, it was wasting even more for the analysts regular work as well.

TASK

Increase adoption by improving Data Discovery

Rob really wanted to "do this right", so after some deliberation and discussion with our Engineering Director partner, we agreed to pivot to focusing on 3 main 'data discovery' objectives:

1

Increase adoption by those less familiar with data science

2

Reduce time spent on helping others find data sets

3

Reduce "time to insights" for answering business questions

A key career growth opportunity for one of my direct reports

About this same time, the Lead Designer I was planning to put on this problem left the team and I had a new team member join the team.

She was a Senior Design but was single-mindedly seeking a promotion to Lead. She thought she was ready and was a little frustrated it hadn’t happened yet. All of her previous managers felt she wasn’t ready, but she disagreed and wanted a chance to prove herself. Plus, she had a background as a data analyst prior to career switching into UX so I decided to give her a chance to prove herself.

I tasked Liz with 3 key goals:

Multi-method
Research Project

Develop & Pitch Experience Vision

Improve UX
Partnership

MULTI-METHOD RESEARCH PROJECT

We needed to conduct a multi-method research project to understand more about where people were getting hung up on their data discovery journey. We also needed to establish regular user research as a consistent part of the process on this team.

DEVELOP & PITCH EXPERIENCE VISION

Once we understood more about the data discovery journey, I tasked Liz with taking the results and developing and pitching an experience vision of what data discovery at Target could look like in the future, and then partner with the POs of Greenfield and Data Portal (our data asset cataloging tool) on a proposal for the first couple of phases to get there.

IMPROVE UX PARTNERSHIP

The Greenfield team was historically very engineering-driven. They weren’t used to working with UX so I also asked Liz to strengthen the partnerships and establish a working agreement with her partners to improve things going forward. (This all would be good experience to show she was ready for a promotion).

ACTION

Autonomy-driven leadership

I wouldn't normally be intricately involved in decision making with a Lead, so I gave Liz a lot of autonomy. My interactions with her included:

Setting objectives and goals for the project in partnership with Product Director and Engineering Director partners.
Weekly 1:1
Regular team Design Critique
Periodic check-ins with cross-functional leadership group

Outside of these, I gave her autonomy on how she approached solving the objectives. As a cross-functional leadership group, we wanted to see what she and the team could come up with.

Research went very well

The multi-method research project went very well. Liz is an extremely capable researcher (probably one of the best designer-researchers at Target). She uncovered some very crucial insights. A couple examples include:

Almost no one could find dashboards / cards / reports that already existed without someone sending them a bookmark.
There were big data inaccuracy & trust concerns. Data in Greenfield cards and the same datasets surfaced elsewhere didn't match up. As a result, a lot of waste was being created in comparing / contrasting data from different sources, and with general avoidance of Greenfield by some users.
In Data Portal, our data cataloging and data set management tool, there were conflicting ways to find the same data assets and none of them were intuitive. Users only could use DP if they already knew the exact name of a data table (for example). The search was only being used to pull up a data set, not to discover one. In 1 example Liz discovered "cart_size" returned 152 unique results. Distinguishing them was extremely difficult and time consuming.

In fact, the data inaccuracy and trust concerns were such a big problem that solving them became a Data Science org wide "Big Rock" OKR. Liz's research was instrumental in uncovering and pitching that.

Partnership did not go well

The partnership was not going well. Engineers simultaneously avoided looping Liz into the big decisions and looped her into too much. They looped her into small UI decisions and she wanted to go deep on everything which frustrated everyone involved. An associate product manager was trying to control UX backlog items, all UX final decisions, and micromanage her work.

Liz was also a very blunt person and didn't approach some of these difficult situations with the tact they needed.

We were also getting behind schedule and while great work was happening we didn't have anything to show for the weeks of work we had been doing. Time was starting to run out to show progress.

I got more involved in the details

I started being more involved in the work, team meetings, etc in an effort to help. We set up a UX Sprint Planning meeting and I participated frequently in Design Reviews with partners. This resulted in some wins (we gained some autonomy over our own "backlog"), but unfortunately, without realizing it, I had began to smother Liz a bit and undermine her decision-making in front of her team. I was actually hurting, not helping.

Hey there, this is the default text for a new paragraph. Feel free to edit this paragraph by clicking on the yellow edit icon. After you are done just click on the yellow checkmark button on the top right. Have Fun!

Radical Candor saves the day

Radical Candor is something in which I really believe. I give and seek feedback regularly and strive to get to know and invest in each team member. I had previously done the work to set this kind of an environment up with Liz which made ALL the difference in the world.

One day, in a 1:1 Liz expressed that I was too in the details. She felt that I had overcorrected in an effort to help, but that it wasn't helping. I was giving feedback to her in front of her team which unintentionally devalued her as the UX subject matter expert on the team. I was trying to ensure the work would be successful, but I was in the weeds too much.

Supporting without micromanaging

To continue supporting her growth without micromanaging I did the following:

I PULLED OUT OF HER BALANCED TEAM MEETINGS
I let her handle her balanced team discussions alone. I gave her ideas, but didn't try to solve problems for her.
I DID MORE PAIR DESIGN SESSIONS WITH HER WHERE I COULD GIVE HER FEEDBACK ON THE WORK 1:1
(This allowed her to look like the expert when meeting with her product and engineering lead partners).
I WAS EXPLICIT WHETHER I WAS PROVIDING 'FEEDBACK' OR 'DIRECTION'
When your manager gives you feedback, sometimes it can be confusing whether they're telling you to "consider" (feedback) or to "go do" (direction). I started to be explicit on which I was providing in our sessions together. I always try to over index on "consider", but that hadn't been made clear. So I was more clear going forward, and encouraged her to ask if it wasn't clear.

I established leadership-level Roles & Responsibilities alignment

Since the working agreements at a team level were not working well, but I still wanted Liz to have her autonomy, I approached Rob (Product Director) and Jason (Engineering Director) to align at a leadership level on Roles & Responsibilities and collaboration model. I facilitated the discussions and created the visualization.

We then shared our expectations down with the team and left it up to them to figure out the specific of how they would achieve this collaboration process-wise on their team. We didn’t want to be prescriptive of the ”how” but did feel it necessary to put guard rails around the autonomy & how we expected partnership to work in our space to facilitate more effective collaboration.

We focused on what the unique responsibility of each role area was individually, overlapping with each other role area, and collectively.

I put in place a new resourcing model

Up until then, Liz had been trying to work in an embedded model. The problem was that there were 2 Greenfield teams and only 1 Liz. She was being pulled in too many directions and wasn’t able to focus the time needed to the most important work.

Also, Rob was product managing one of the teams, while a more junior product manager was doing some of the work. There wasn’t good scope over which team did what.

I pitched & worked with Rob & Jason to restructure the teams into a Dual-Track Resourcing Model (with the support of my UX Director). Rob, Jason, and I worked together as a cross-functional unit on setting the overall vision and prioritizing how to get there.

A problem statement would then be delivered to the Discovery Team and they would flush it out, come up with a vision and project plan on how to best solve the problem.

For example, the Leadership Team identified that Data Discovery & Trust was an initiative we wanted to go after first. The Discovery Team then did research and brainstorming and determined that they would focus on standardizing and locking down Dataset Management & Certification first, then Greenfield Navigation, then Card Search in Greenfield. When the Dataset Management & Certification standardization designs were done each product manager worked with their own Development team to refine & implement their part of the designs.

I set up Leadership - Discovery team check-ins

It is important to note that the joint Leadership team was in constant contact with the Discovery team, providing feedback and helping to guide the direction as needed, but the work was being done by the Discovery team. This was a cross-functional leadership team leading the work of cross-functional individual contributor teams.

RESULTS

Major cultural and business result wins

The noteworthy accomplishments coming from this project include:

Business

We saved the company $27M in reduced wasted productivity costs (annually)
Greenfield onboarding time decreased by 175%.
530% reduction in "Time to insights" (from question to data-informed decision).

Culture

Liz's work (partnering with the Discovery team) resulted in a Data Science org wide strategic investment to focus more on Data Discovery & Trust.
Research was established as a regular part of 3 teams' process (not just 1) with cross-functional representation in 100% of sessions.
Greenfield teams hit 300% of customer exposure hours goal
Goal setting & strategic planning became unified instead of separate & shared outcome-focused initiatives were established Data Science wide.
Product & Engineering begged for more UX resources and aided me in the business case to secure them

Perhaps the accomplishment I'm the most proud of, however, is how much Liz grew as a designer and partner. I also was grateful for the experience and learned a lot about coaching without micromanaging along the way.