Ingestion Mapping

Customers were originally expected to work with our Solution Engineers (SEs) to ingest their data. That wasn't scalable, and we'd been making efforts to provide self-service tools to lessen the load on SEs.

Customers need a way to quickly and easily ingest and map their data on the Jumio portal in order to onboard and begin monitoring and investigations, without the hassle of reaching out to a Jumio SE.

Scale made this urgent. Each new customer consumed an estimated 10 to 15 hours of SE time across kickoff calls, sample file reviews, and rework, and ingestion was the longest stretch of onboarding: often two to three weeks before a customer saw their own data in the portal.

1. 01

   Self-serve from day one

   Let a new customer generate an API key and start sending data without waiting on a Jumio SE.

2. 02

   Two ingestion paths

   Support real-time ingestion through the API and batch files through SFTP, since customers' technical capabilities vary widely.

3. 03

   Data that reads like theirs

   Let customers map fields, naming conventions, and formats so the portal speaks their internal vocabulary.

With Product we framed success around two measures:

The first step to creating a seamless experience is simply understanding the process. This involved meetings with product and engineering to understand what the customer needs and how each step connects, plus a knowledge base of existing documentation on successful data ingestion.

We added settings pages for API, Ingestion, and Mapping. For ingestion, I approached the design by understanding how it's done technically; the customer was expected to submit a JSON file in an expected format.

After ingesting data, users needed to customize the mapping of fields to fit their internal vocabulary and formats.

The API page started by simply giving customers a way to create an API key, along with the other endpoints they could utilize, plus links to the API documentation.

I initially partitioned information under accordions, hoping to reduce visual clutter and direct the user toward the steps needed on each page. After discussion, it proved an unnecessary extra click to access information they'd need anyway, so I iterated to display all information and options upfront.

Accordions were unnecessary here, causing more disruption than the visual clutter they were meant to prevent.

The business call was to skip formal validation testing and learn from production instead; the surface area was contained and internal alignment was strong. I pushed to make testing in production mean something: we tracked where customers stalled in setup and reviewed support tickets with the SE team in the months after launch.

70%

of new accounts completed ingestion and mapping without an SE session in the first two quarters

3x

faster time to first data in the portal for self-serve accounts, days instead of weeks

There were also plans to expand this feature with custom field mapping and proxy assignment.