Key Takeaways

  • Matt Bank revealed that GEM flipped its software budget within twelve months, moving from 90% third-party tools to 90% in-house software development.
  • GEM generates approximately 30,000 lines of code per month, a tenfold increase over the prior year, building bespoke tools to track manager personnel moves.
  • GEM uses Glean, a model-agnostic internal search engine, to drive internal data query and retrieval costs across historical firm files to zero.
  • John Lawrence at Rice Management Company is assembling a centralized data lakehouse to synthesize unstructured manager letters, conference takeaways, podcasts, and track-record metrics.
  • Institutional allocators are moving away from vendor SaaS suites because off-the-shelf software cannot organize the private, conversational data where institutional memory lives.

The 90% Shift From Commercial SaaS to In-House Code

For decades, institutional investment offices bought software off the shelf. They bought CRMs, portfolio accounting packages, and manager databases, then hired consultants to stitch them together. That vendor-first model is breaking down.

Matt Bank explained how fast the economics reversed at GEM. “However, if you rewind the calendar a year, about 90% of our spend and our toolset was externally developed. We were buying off-the-shelf things, trying to integrate them. Today, that has flipped. 90% of the things we're developing are internal.”

The driver behind this inversion is the collapse in software creation costs. Investment teams no longer need large software engineering squads to build tailored tracking tools. “Our team generates about 30,000 lines of code per month,” Bank stated. “I'm told that's a lot, I don't know, for a firm of our size, but it's up 10x from where that was last year.”

Instead of waiting for commercial software vendors to build workflows for tracking GP departures or mapping fund relationships, GEM builds custom internal systems. Bank also relies on Glean to search across firm files: “We have an internal search tool called Glean, which is model agnostic. It has basically driven the cost of search across the firm's various data sets to zero.”

When search costs hit zero and code generation accelerates tenfold, the value proposition of generic SaaS seats evaporates.

The Lakehouse Bottleneck: Unifying Unstructured Allocator Data

Building code is fast, but structuring the underlying institutional data remains a hard operational obstacle. Allocator information does not live in neat SQL tables. It sits in quarterly LP letters, PDF memos, partner meeting notes, and audio transcripts.

John Lawrence described the multi-year challenge Rice Management Company faces while attempting to unify these sources into a single analytical foundation. “The real challenge is aggregating our data into one centralized data lakehouse, which we're working on. It's easier said than done. That has taken more time than I would have expected.”

The volume of incoming qualitative intelligence gives multi-asset allocators an informational advantage, provided they can structure it. “Allocators have tremendous amount of access to managers, to letters, to peer conversations, to conferences, to podcasts,” Lawrence pointed out. “Aggregating all that data together can be a challenge, and then making better decisions, that's ultimately where we want to get.”

Without a unified lakehouse, manager commentary remains locked in individual inboxes and siloed drives. The allocators winning the next cycle will not be the ones buying the most third-party data feeds. They will be the ones who successfully index their own private history.

Why It Matters

Institutional allocators sit on decades of proprietary qualitative data that third-party vendors cannot access. By substituting generic software subscriptions for custom internal applications, top endowments and foundations are turning internal operational archives into private intelligence engines. This dynamic threatens standard vertical SaaS pricing power while widening the analytical gap between tech-enabled investment offices and slow-moving peers.