Yi-Ying Lin

Work

About

Yi-Ying Lin

Work

About

Yi-Ying Lin

Work

About

Lookup Columns: Simplifying access to scientific data

Lookup Columns: Simplifying access to scientific data

Lookup Columns is a no-code query builder on Benchling’s life sciences R&D platform, built to let scientists trace complex data relationships without admin support. Launched in 2021, it now powers 14 million lookups a year and serves as the foundation for data tracing across Benchling. I led the design with a PM and four engineers, taking it from concept to launch.

The problem

The data was there, just out of reach

At Benchling, scientists manage data through schemas, but getting the right data in context was slow, technical, and frustrating. Answering a simple question like “Is this protein’s source cell line still viable?” often meant digging through layers of samples, containers, and results.

The only tracing tool, Lookup Fields, didn’t cover inventory, couldn’t work across table types, and depended on formula syntax that required schema expertise. Without a better option, scientists had to jump between records, constantly losing context and wasting time. In interviews, they were clear: they needed a faster, more intuitive way to get the right data into their tables.

The only tracing tool, Lookup Fields, didn’t cover inventory, couldn’t work across table types, and depended on formula syntax that required schema expertise. Without a better option, scientists had to jump between records, constantly losing context and wasting time. In interviews, they were clear: they needed a faster, more intuitive way to get the right data into their tables.

Lookup Fields

Screenshot from the Benchling Help Center showing Lookup Fields, a formula-based tool that was hard to read and limited in scope.

Entity lineage map: Purified protein

Example of the complex, multi-layered data relationships scientists navigate. This map centers on purified protein, with each linked entity branching into its own network.

The solution

Lookup Columns: scientists’ data companion

The breakthrough came from how scientists themselves described “looking up” data. I translated their language into steps and designed a no-code query builder that lets them pull multi-layered, linked data directly into their tables. With plain language, a step-by-step builder, and helpful prompts, it keeps their focus on science instead of technical details.

The breakthrough came from how scientists themselves described “looking up” data. I translated their language into steps and designed a no-code query builder that lets them pull multi-layered, linked data directly into their tables. With plain language, a step-by-step builder, and helpful prompts, it keeps their focus on science instead of technical details.

To make it extensible, I partnered with PMs, engineers, and the Lab Automation designer to create LookupConfig, the flexible component behind Lookup Columns. By abstracting the configuration logic, we set the foundation for bringing the same no-code query experience to other parts of the platform.

To make it extensible, I partnered with PMs, engineers, and the Lab Automation designer to create LookupConfig, the flexible component behind Lookup Columns. By abstracting the configuration logic, we set the foundation for bringing the same no-code query experience to other parts of the platform.

Configure lookup

Lookup steps

Trace data lineage step by step, with options updating by linked item type.
Trace data lineage step by step, with options updating by linked item type.

Lookup filters

Filter lookup results, such as barcodes from containers with more than 20 mL.
Filter lookup results, such as barcodes from containers with more than 20 mL.

Lookup Tables

Lookup Tables bring reference data into entries and account for 21% of structured tables.
Lookup Tables bring reference data into entries and account for 21% of structured tables.

Extensible framework

Designed for reuse, LookupConfig now powers features across Benchling, such as Run Schemas (shown here).
Designed for reuse, LookupConfig now powers features across Benchling, such as Run Schemas (shown here).

The impact

Adopted widely, integrated deeply

Since launch, Lookup Columns have seen rapid adoption and strong feedback. In the past year alone, scientists created 14 million lookups and 1.5 million Lookup Tables across 7 million structured tables. The approachable, step-by-step design made it easy for scientists to configure queries themselves, driving quick adoption. Beyond usage, LookupConfig has become a core building block for data tracing across Benchling, now powering Notebook tables, Computed Fields, Run Schemas, and Task Schemas.

Since launch, Lookup Columns have seen rapid adoption and strong feedback. In the past year alone, scientists created 14 million lookups and 1.5 million Lookup Tables across 7 million structured tables. The approachable, step-by-step design made it easy for scientists to configure queries themselves, driving quick adoption. Beyond usage, LookupConfig has become a core building block for data tracing across Benchling, now powering Notebook tables, Computed Fields, Run Schemas, and Task Schemas.

This is the next level of Benchling implementation. Very intuitive and focused. I believe our scientists will be able to configure Lookup Columns themselves.

— Enterprise customer

I really like the idea of using simple lookup steps, like looking up results of an entity, to get to the data I need. I didn’t always know what to expect at first, but once I tried it, the design made a lot of sense.

— Benchling Customer Success

With Lookup Columns, I can see the freeze dates of my VPC containers right in my table, so I know right away if they’re still good for my experiments without having to dig through inventory records.

— Enterprise customer

Yi-Ying Lin, 2025

LINKEDIN

EMAIL

Yi-Ying Lin, 2025

LINKEDIN

EMAIL

Yi-Ying Lin, 2025

LINKEDIN

EMAIL