Enabling a Scaled Enterprise SaaS Platform Through Self-Serve Data Access
LIVERAMP
A FRagmented Starting Point
LiveRamp offers a very powerful set of tools and products for clients to achieve their data-related business goals including activation, measurement, collaboration, and many more. Unfortunately, this product offering was fragmented, and its data ingestion story even more so. When I started looking at the space, I found the diagram to the right that showed the company to have six different data ingestion flows. This meant six different workflows that customers had to go through to make their data usable by a given feature set, with little to no interoperability. When customers used multiple products, they were forced not just to re-ingest their data, but possibly even reformat it.
Ingesting or connecting data is the third most common user task, with over 56% of users doing so on a regular basis. Exacerbating the issues, none of these flows were self-service and thus relied on support personnel to set up any new pipeline.
Early Discovery And What Users Were Telling us
User flow for creating a basic data object that contains customer data from a file.
With the early signals on product fragmentation in hand, I set out to form the initial situation and hypotheses by interviewing stakeholders and reviewing existing customer feedback in ProductBoard and previous user research. The user personas here are Supporters, internal users who assist customers with their data goals, and Uploaders, external users whose primary task is getting data into the platform.
Interviewing Supporters, I found several lesser-known ingestion flows, bringing my total from six to eleven. I learned more about the technology behind the main ingestions flows and that they were slow, inefficient, and costly to compute. I asked several support experts to walk me through the process of connecting a data source to one of our products and how they set up a data asset. The results were a little bit shocking: the bulk of our tech back end was propped up by Supporters carrying out manual, repetitive tasks. See the diagram, above. To me, this looked like a major case of “man-behind-the-curtain.”
I completed a heuristic evaluation on the few places we did surface ingestion information and found them to be lacking against accepted standards. I parsed as much existing feedback from Uploaders as I could find, and was able to break the user issues into four main categories:
Slow: “I was confused that when I wanted to upload a file and create a new audience, I need to submit a [support] ticket."
Unsophisticated: “[The UI] will say [the file] is delivered but it’s 4 or 5 days until it shows up. We get our stuff done and we don’t see the data in Safe Haven.“
Opaque: “It’s unclear where [the data] is in the process. Is LiveRamp aware of that data? If [LiveRamp] is aware, where is the data in that process? Is it being ingested? Did it fail? Was there an error? Was the format wrong? Was the encryption type wrong? We want a pizza tracker and an ETA on what’s happening.”
Rigid: “Files can get halted for a missing prefix or a misplaced capitalized letter and take days to resolve.”
Forming Hypotheses and strategy
Having assessed the situation, I started forming hypotheses to address the issues. Most of which span multiple problems:
Integrated Product Offering
Unifying the entry points for data ingestion and offering a modular configuration wizard will simplify the product offering for the user, increasing their satisfaction, and reduce the costs associated with maintaining duplicative functionality.
End-to-End Self Service
Enabling end-to-end self-service, including source connection, asset creation, and pipeline visibility will empower users and reduce time-to-value with LiveRamp’s product offering. Support costs will be dramatically reduced, which would open new revenue streams by enabling that team to provide white glove, premium service.
Intent Driven Workflows
LiveRamp’s tools have always been powerful, but allowing the user to tell us what they want to do with their data allows us to show them the right tools for the right job. This will reduce cognitive load, and in turn empower customers to successfully execute on their goals.
Efficient Dataset Management
At risk of stepping out of my technical depth, storing data once and applying transformations dynamically as product features require, will reduce compute costs and delight customers through dramatically speeding up ingestions SLAs.
Surviving and Thriving Within the LiveRamp Product Development Life cycle (PDLC)
Taking a detour in the story, a big part of what attracted to me to LiveRamp was the maturity of their Product Development Life Cycle, design processes, and stature of the design team. I wondered, “Why hadn’t this existed at my past companies?”
I started empowered to conduct discovery on the problems facing our customers, turn that into hypotheses, influence strategy, and validate everything.
For some time, I’ve believed in scaling the level of validation to the risk of the proposal. Riskier decisions call for more intense validation. This initiative had a variety of risk severities, and I had the opportunity to do more research than in my entire time at my last company:
LiveRamp’s PDLC at a high level when I was hired.
FEEDBACK
Feedback throughout the initiative was invaluable for challenging my assumptions and validating the initiative as a whole. Critical feedback helped me shape the structure and refine interactions. Observing alpha users refined that even more, as seeing where users try to click can spark insights that you just can’t get in a semi-interactive prototype. All of that said, the positive feedback and usability results reinforced the initiative as a whole:
Usability Wins - 92% ingestion task success rate in usability testing with zero direct failures.
File Tracking Visibility - “This is cool to see what step [my data] in and what it means”
Structural Clarity - “[Data Sources are] the origin point for getting the data into LiveRamp”
Empowering Self-Service - “Good! This will allow clients to QA their files” & “This is nice that it did it for me”
So What About Some Features?
Given the breadth of this initiative, there were and are a number of individual features delivered incrementally. I’ve included vertical slices of some of two of the most impactful individual pieces that have a UI.
Intent-based workflow
Only show relevant components based on the selections the user makes on what they’re trying to accomplish. We can show them an easy-mode template version or give users full power of customizing their schema based on their needs.
File Tracking Visibility - “The Pizza Tracker”
”Where’s my data?” is not a question that should require a support ticket.
Challenges, Results, and Takeaways
This initiative was successful, but it was not without its challenges. Over the course of two years:
There was significant turnover and I was paired with five different product partners. This affected momentum each time there was a switch. Thankfully, the last one stuck and she was instrumental in carrying this forward.
Company “North Star” strategy was constantly in flux. At LiveRamp, the most important thing for the company seemed to change quarterly. This could be frustrating, as you’d expect strategy to stay more constant and instead adjust tactically. This taught my product partner and I how to develop our narrative to be foundational to ANYTHING the company was trying to do.
The ever elusive “phase two.” because of the constant strategy shifts, it was sometimes difficult to secure resourcing for iteration. This can be overcome by showing the value of iteration in relation to cross functional initiatives.
The above challenges forced me to finally grow my ability to disagree and commit. Being right in hindsight isn’t satisfying in itself, but building a track record as a good decision maker helps to distinguish between a good designer and an effective one.
That track record is then supported by a few key data points:
The dataset management structure I proposed reduced ingestion times by 90% and reduced costs by 50-75%, not even counting the reduced support costs.
The introduction of the “Pizza Tracker” led to a 16 point increase in System Usability Scale (SUS) score.
Support tickets for repetitive manual tasks dropped by 47% year-over-year, showing how we empowered users.
Because of this type of project and its results, the design team at LiveRamp is looked at as the go-to for leading cross-team, cross-discipline discovery and collaboration. The framework developed here is proving to be flexible and scalable to new lines of business.
