How did Catalant scale to 100k experts?

Founders building talent marketplaces often assume that simply adding more experts will drive growth, but the real bottleneck is turning sporadic requests into a predictable, repeatable sourcing flow. At Catalant the challenge was not just attracting a large pool of specialists—it was making each client request reliably surface the right expertise at scale. This tension between quantity and reliable matching is frequently overlooked, leading many platforms to stall once they hit a few thousand experts. By rethinking the sourcing process as a data‑driven engine, the company uncovered a hidden lever that turned ad‑hoc gigs into a systematic pipeline. Now let's break this down.

Why does turning each request into a repeatable sourcing engine matter now

Founders often think that simply adding more consultants will unlock growth, but the real lever is predictability. When a client submits a brief, the platform must instantly surface the right expert without manual hunting. This repeatable sourcing engine transforms a sporadic gig into a reliable pipeline, allowing sales teams to promise delivery timelines and investors to model revenue with confidence. The shift from reactive matching to a systematic process also reduces acquisition cost because the same data set fuels multiple requests, creating network effects that amplify each new expert added. At scale, this predictability becomes the moat that separates a thriving marketplace from one that stalls after a few thousand specialists. Companies that embed this engine early can scale faster, retain clients longer, and attract top talent who see consistent work flow.

What common misconception leads platforms to stall after a few thousand experts

Many builders assume that a larger talent pool automatically improves match rates, yet the bottleneck often lies in signal quality. As the roster expands, the platform’s search algorithm can become noisy, presenting clients with irrelevant options and increasing decision fatigue. Without a disciplined data model that ranks experts by proven outcomes, the average time to fill a request rises, eroding client trust. This misconception also drives founders to invest heavily in marketing to recruit more experts instead of refining the matching logic. The result is a bloated marketplace that feels impressive on paper but delivers inconsistent outcomes in practice. Correcting this bias means focusing on depth of data per expert – past project performance, skill endorsements, and client satisfaction – rather than sheer headcount. When the engine learns from each placement, it continuously improves relevance, allowing the marketplace to grow without sacrificing quality.

How can founders design a data driven matching engine that scales reliably

The first step is to treat every request as a data point rather than an isolated transaction. Capture structured inputs such as project scope, budget range, timeline, and required competencies, then map these to a dynamic skill taxonomy. Next, enrich each expert profile with historical performance metrics, including success rates, repeat engagements, and client ratings. Apply a weighted scoring model that balances fit, availability, and proven outcomes, and continuously retrain the model with new placement data. Founders should also automate feedback loops: after each engagement, collect outcome data and feed it back into the algorithm to refine future scores. Finally, build a monitoring dashboard that tracks key indicators like match speed, fill rate, and client satisfaction, enabling rapid identification of drift. By iterating on this loop, the engine becomes self improving, turning every transaction into a learning opportunity and supporting exponential growth without manual intervention.

FAQ

How quickly can a marketplace grow its expert pool without sacrificing match quality

Growth speed depends on the strength of the matching engine rather than raw recruitment. When the data model reliably surfaces high quality experts, the platform can add new talent at a rate of several thousand per month while maintaining fill rates above ninety percent. The key is to onboard each new expert with a baseline performance profile – past project outcomes, skill verification, and client references – so the algorithm can rank them accurately from day one. This approach lets the marketplace expand rapidly without a dip in client experience.

What metrics indicate a matching engine is working effectively

Three core signals reveal engine health: average time to present a qualified expert, fill rate of posted requests, and post engagement satisfaction score. A decreasing time to present indicates the algorithm is narrowing the candidate set efficiently. A high fill rate shows that the pool and the scoring model are aligned with client needs. Finally, a rising satisfaction score confirms that the matches deliver real value. Monitoring these metrics together provides a clear picture of performance and highlights areas for improvement.

Can the data driven approach be applied to other service marketplaces

Yes, the principles are platform agnostic. Whether matching designers, developers, or logistics providers, the process of converting each request into structured data, enriching provider profiles with outcome history, and continuously retraining a scoring model yields the same predictability benefits. The main adaptation lies in tailoring the skill taxonomy and performance indicators to the specific service domain, but the feedback loop and metric framework remain identical.

Early founders often settle on a single, centralized system for managing talent before they scale.

In the stories they repeatedly encounter fragmented data sources, invisible work pipelines, and coordination breakdowns when teams rely on separate spreadsheets, email threads, and ad hoc tools. A centralized hub aggregates identities, assignments, progress updates and payment triggers into one view, restoring visibility and aligning actions across the network. At the system level this means a shared ledger of work objects that every participant reads and writes, rather than isolated silos that must be reconciled manually. Workhint serves as an example of such a hub, illustrating how a unified infrastructure can hold the network, the workflow logic and the execution data in one place. By grounding talent flow in a single structure, founders avoid the hidden cost of constantly stitching together disparate signals, allowing the marketplace to operate with consistent timing and clear oversight.

The story of Catalant shows that the real lever for reaching a hundred thousand experts is not sheer volume but the confidence that every client brief will instantly surface the right talent. By treating each request as a data point, enriching expert profiles with outcomes, and feeding results back into a weighted scoring model, the platform turned a chaotic stream of gigs into a self‑reinforcing engine. That engine created predictability, lowered acquisition costs and built a moat that grew faster than the headcount itself. The lasting insight is simple: growth follows the reliability of the match, not the size of the roster. Consistent, data‑driven matching is the quiet engine that powers scale. Predictability is the hidden currency of marketplace growth.