For a Headhunting Firm : A design for Productivity

 

 

WHITE PAPER The Pyramid of Productivity A Visionary System Architecture for the Future of Recruitment H.C. Parekh  |  3P Consultants  |  Founded 1990  |  Blueprint: January 2003

 

Executive Summary

 

In January 2003, H.C. Parekh — a former General Manager at Larsen & Toubro and founder of 3P Consultants — committed to paper a system architecture that was, in many respects, a decade ahead of its time.

The document, titled the 'Pyramid of Productivity,' laid out a five-layer technology and operations framework for a next-generation recruitment enterprise.

 

What makes this blueprint remarkable is not merely its ambition, but its structural coherence. Parekh had identified, by 2003, the core components of what would later be recognised as a recruitment technology platform: master data management, multi-entity databases, AI-assisted matching tools, a SaaS application layer, and a multi-sided user network. Each of these pillars is now a distinct product category attracting billions in venture capital.

 

This white paper examines the Pyramid layer by layer, contextualises it against today's HR technology landscape, and draws out its enduring lessons for entrepreneurs, HR leaders, and technology strategists.

 

1990 Founded Post-L&T startup

5 Architecture Pyramid layers

10,000+ Target Corporates User group

2003 System Blueprint Pyramid diagram

 

The Founder and the Vision

 

H.C. Parekh's path to 3P Consultants was shaped by a formative decade in industrial relations at one of India's most complex manufacturing enterprises. As General Manager at Larsen & Toubro's Powai factory (1979–1987), he had led a pioneering programme of participative management — building trust between workers and management through structured dialogue, transparency, and shared ownership of outcomes. That programme produced measurable productivity gains of 3% per year.

 

The founding of 3P Consultants in 1990 represented a pivot from managing people within an organisation to building the infrastructure through which organisations find, assess, and place people at scale. Parekh brought to this new venture the same systems orientation that had characterised his L&T work: the conviction that sustainable performance requires well-designed architecture, not improvisation.

 

The recruiter who wins is not the one with the best rolodex — it is the one with the best system.

 

The Pyramid of Productivity: Layer-by-Layer Analysis

 

The pyramid is structured with the most abstract, foundational elements at the top and the most concrete, user-facing elements at the base — an inversion of the conventional pyramid metaphor that reflects Parekh's engineering sensibility. Each layer depends on the integrity of the layers above it.

 

Layer	Components	Modern Equivalent
Masters (Top)	Industry · Function · Designation Level · Education · City/PIN/STD · Company-Group · Country-Currency · Edu. Institutes	Master Data Management (MDM) — normalised reference tables that ensure every record is consistently tagged
Databases	Member Resumes · Non-Member data · Client Profiles · Company Profiles · Job Ads & Histories	Multi-entity relational schema — candidates, clients, companies, and postings as first-class objects
Software Tools	Highlighter · Eliminator · Refiner · Compiler I/II · Composer I/II · Ad Viewer · Matchmaker · Desig. Splitter · Educator · Extractor · Classifier · Tabulator · Profiler	AI-assisted recruitment workflow engine — what LinkedIn's Recruiter and modern ATS platforms do today
Applications	OES · Manhattan · Lock-In · Communicator · Home Working / Video Interview · Web-Service	SaaS product suite — remote interviewing, engagement tooling, and web-based service delivery
User Groups (Base)	3P Consultants (Residents & Office) · 700 Newspapers · Content Providers · 16,000 Cybercafes · 1,000 Franchised Placements · Interview Experts · 3P Consultants (Home) · 10,000 Corporates · Recruiting Cos.	Platform network — multi-sided marketplace aggregating supply, demand, and distribution simultaneously

 

Deep Dive: The Five Layers

 

Layer 1 — Masters: The Foundation of Data Integrity

 

The topmost layer of the pyramid is not the most glamorous, but it is the most critical. The Masters layer defines the controlled vocabularies and reference data sets against which all other data is tagged: Industry, Function, Designation Level, Education Level, City/PIN/STD/State, Company-Group, Country, Currency, and Educational Institutions.

 

The annotation 'Actual Designation ?' beside the Designation field is telling. Parekh was already grappling with what remains one of HR data's most stubborn problems: job title proliferation. When every company invents its own titles — 'Associate Vice President,' 'Senior Consultant II,' 'Principal Engineer (Band 6)' — cross-company comparisons become impossible. The Masters layer was designed to impose order on this chaos through standardised designation hierarchies.

 

In today's parlance, this is Master Data Management (MDM). The fact that Parekh designed it in 2003 — before most Indian companies had formalised HR information systems — speaks to the depth of his data architecture thinking.

Layer 2 — Databases: The Multi-Entity Schema

 

The Databases layer separates five distinct data entities: Member Resumes, Non-Member data, Client Profiles, Company Profiles, and Job Advertisements with their Histories. This separation is not merely organisational tidiness — it reflects a sophisticated understanding of data relationships.

 

A candidate is not the same as a member. A client is not the same as a company. A job advertisement is not the same as a vacancy. By maintaining these as distinct entities with their own schemas and relationships, Parekh designed a system capable of nuanced queries that flat-file or single-table approaches could not support. This is, in essence, a normalised relational database design — the kind that enterprise software companies charge significant licence fees to provide.

Layer 3 — Software Tools: The AI Layer Before AI

 

This is the layer that most astonishes a modern reader. Parekh identified thirteen distinct software tools, each performing a specific function in the recruitment workflow:

 

•       Highlighter & Eliminator — automatic shortlisting based on defined criteria

•       Refiner — iterative narrowing of candidate pools

•       Compiler I/II — aggregating profiles from multiple sources

•       Composer I/II — generating structured outputs (letters, reports, shortlists)

•       Ad Viewer — browsing and parsing job advertisements

•       Matchmaker — the core engine, pairing candidates to roles by profile fit

•       Desig. Splitter — decomposing non-standard titles into standard components

•       Educator — training and upskilling module

•       Extractor — parsing unstructured resumes into structured data

•       Classifier — categorising candidates by function, industry, level

•       Tabulator — generating comparative reports

•       Profiler — building rich candidate profiles from multiple data points

 

The Extractor and Matchmaker tools, in particular, describe what is today a multi-billion dollar software category. Resume parsing (Extractor) and AI-driven candidate-job matching (Matchmaker) are the core features of platforms like Eightfold AI, HireVue, and LinkedIn Recruiter. Parekh conceived them when the standard technology for resume processing was a fax machine.

Layer 4 — Applications: The Product Suite

 

The Applications layer translates the underlying data and tools into user-facing products. Six applications are identified:

 

•       OES (Online Employment Service) — the core job portal

•       Manhattan — likely the enterprise client management application

•       Lock-In / Blackhole — engagement and retention tooling, designed to keep both candidates and clients within the 3P ecosystem

•       Communicator / Octopus — multi-channel communication platform (email, SMS, alerts)

•       Home Working / Video Interview — remote assessment and placement

•       Web-Service — the API and integration layer connecting 3P to third-party systems

 

The inclusion of a Home Working and Video Interview module in 2003 — seventeen years before remote interviewing became a business necessity — is perhaps the single most striking element of the entire pyramid. Parekh had recognised that geography was an artificial constraint on talent placement, and designed a product to dissolve it.

Layer 5 — User Groups: Platform Economics Before the Term Existed

 

The base of the pyramid is the most expansive layer, and the one that reveals Parekh's most distinctive insight: that 3P Consultants was not a recruitment agency but a platform. The user groups identified include:

 

•       3P Consultants (Residents and Office-based) — the supply-side professionals

•       700 Newspapers — distribution and sourcing partners

•       Content Providers — data enrichment partners

•       16,000 Cybercafes — distributed access points for candidates without home internet

•       1,000 Franchised Placements — distributed delivery network

•       Interview Experts (Home-based) — distributed assessment capacity

•       3P Consultants (Home-based) — remote workforce

•       10,000 Corporates and Recruiting Companies — the demand side

 

The 16,000-cybercafe distribution network is particularly brilliant in its context. In 2003, broadband penetration in India was minimal. Parekh designed around this constraint by treating cybercafes as branch offices — distributed access points through which candidates across India could engage with the 3P system. This is the same problem that M-Pesa solved in Kenya a decade later by using mobile airtime agents as banking branches.

 

Ahead of Its Time: The 3P Blueprint vs. Today's HR Tech

 

The table below maps each element of the 3P Pyramid to its contemporary equivalent in the global HR technology landscape — products and platforms that were built, in most cases, years or decades after Parekh's 2003 blueprint:

 

3P Vision (2003)	Industry Reality (Today)
"Matchmaker" tool for candidate-job fit	LinkedIn AI Recruiter, HireVue, Eightfold.ai
Video Interview / Home Working module	Zoom, Teams, HireVue — mainstream post-COVID
16,000 cybercafe distribution network	Mobile-first job portals (Naukri app, Indeed)
Franchise placement network (1,000 nodes)	Staffing aggregators, gig platforms
Extractor tool for resume parsing	AI resume parsing (Sovren, Textkernel, Rchilli)
Web-Service application layer	Cloud-native SaaS HR platforms
Multi-sided user groups (corporates + consultants + newspapers)	Indeed, Naukri — two-sided marketplaces

 

Lessons for Entrepreneurs and HR Leaders

 

Systems Before Scale

The most common failure mode for recruitment businesses is to grow headcount before building infrastructure. Parekh's instinct was the inverse: design the system first, then populate it with users. The pyramid's five-layer architecture could support a business of any size because its foundations were not human — they were structural.

Data as Competitive Moat

The Masters and Databases layers of the pyramid represent a form of competitive advantage that is difficult to replicate: clean, structured, proprietary data. Every placement, every assessment, every client interaction was designed to feed back into the database — making the system smarter with each transaction. This is the flywheel that powers every successful platform business.

Distribution is Strategy

The 16,000-cybercafe and 1,000-franchise network was not a marketing plan — it was a distribution strategy. By recognising that candidate reach was constrained by infrastructure, and designing a distribution layer to overcome that constraint, Parekh was thinking about the problem in a way that most recruitment agencies never did.

Remote Work is a Feature, Not a Compromise

The Home Working and Video Interview application was designed into the core product suite, not bolted on as an afterthought. This architectural decision reflected a belief that remote work was a capability to be built, not a problem to be managed. Most of the world took until 2020 to reach the same conclusion.

 

Platform thinking asks not 'how do we serve more clients?' but 'how do we build the system through which clients and candidates find each other?'

 

Conclusion

 

The Pyramid of Productivity is a document from 2003 that reads like a product roadmap from 2015.

H.C. Parekh had, in a single hand-drawn diagram, sketched the architecture of what the global HR technology industry would spend the following two decades building — at a cost of tens of billions of dollars in venture capital and engineering talent.

 

That the vision preceded the technology is not a tragedy. It is a testament to the quality of the thinking. The constraint was not imagination — it was infrastructure: broadband penetration, cloud computing, mobile devices, and AI tooling that simply did not yet exist at the scale the pyramid required.

 

What endures from the pyramid is not any individual feature, but the architectural discipline it embodies: the insistence on building from foundations upward; the commitment to data integrity before data volume; the recognition that a recruitment business is not a collection of individual transactions but a system for creating matches at scale.

 

For the entrepreneurs, HR leaders, and technology strategists of today, the pyramid offers a simple challenge: draw your own. If you cannot describe your business as a coherent layered architecture — from master data through to user groups — you may be building a practice, not a platform.

 

About H.C. Parekh & 3P Consultants   Hemen C. Parekh served as General Manager (Industrial Relations) at Larsen & Toubro's Mumbai factory from 1979 to 1987, where he pioneered a participative management programme that produced sustained productivity gains. He founded 3P Consultants in 1990 as a technology-forward recruitment enterprise. The Pyramid of Productivity diagram was drawn on 6 January 2003 as a system architecture blueprint. Parekh launched the blog 'Letters to L&T Employee' on his 80th birthday (27 June 2013) and subsequently created a digital avatar at www.hemenparekh.ai.