A Blueprint for the Modern Insights Agency

Foreword

I have spent nearly three decades working in and for insights businesses. During this period, I’ve experienced—and often led—major shifts in how insights are gathered, analysed, and delivered. I conducted my first online survey in 1998, back when online research was little more than a curiosity. Since then, I’ve watched as technology has reshaped every facet of our industry, from offline to online, desktop to mobile, and finally to the full digitalisation of the economy.

Over this 25-year journey, I have had the privilege to lead insights organisations through nearly every function—from sales, customer success, and operations to product management, HR, and strategy. I have seen firsthand the business decisions that fundamentally shape the products and services we deliver. I’ve been responsible for innovation, transformation, and strategic realignment, including the challenging task of integrating billion-dollar acquisitions, deploying AI-driven operating models, and building global teams capable of delivering sustainable growth.

Yet despite these advances, the core operational model of many insights agencies remains firmly anchored in a legacy of consulting culture—manual, custom, and inefficient. It is clear to me, as someone who has both witnessed and actively shaped industry evolution, that incremental changes and superficial technology additions simply aren’t enough. A more profound shift in the entire operating model is not just advisable, but essential.

This e-book is a practical blueprint reflecting insights I’ve gained from both successes and lessons learned over my career. Its goal is to show what genuine transformation looks like and how it can be achieved in practice. I share it openly because, although the ideas here are clear, implementation requires thoughtful commitment and significant effort—exactly the qualities that separate ambitious leaders from the status quo.

I hope it serves you well.

Jonathan “JD” Deitch Paris, June 2025

Introduction

This e-book lays out a blueprint for a modern operating model for insights agencies in the AI age.

A new approach is sorely needed. Today, most insights agencies remain stuck in a cumbersome and costly legacy of consulting culture. A lead arrives through an email; senior researchers spend hours teasing out objectives and designing an instrument; a small army programs and fields the study, then grinds through data processing and PowerPoint creation. The researcher steps back in to interpret the results, and finally an invoice is issued that inevitably fails to capture the labour expended. This process would collapse under its own weight were it not for dedicated (and tired) people who manually oversee every critical step and handoff. Years of custom research and manual workflows have bound agencies into a structural straitjacket that incremental fixes or bolt-on technologies cannot loosen. True transformation means more than streamlining workflows—it demands fundamentally rebuilding the product and go-to-market strategies as well.

And yet, a do-it-yourself insights platform is not the answer. DIY platforms primarily solve suppliers’ problems, not clients’. Clients buy confidence, not data—they purchase the reassurance that their decisions rest on sound, defensible methods executed by someone other than themselves. That confidence disappears the moment the workload shifts back onto their desks, risking compromised decision-making and internal scrutiny.

What’s truly needed is an integrated stack that gives clients flexibility and the confidence that greater researchers are still there, delivering great research, while stripping out lengthy, low-value manual work that erodes margin. This means deploying technology to shoulder configuration and execution, reserving human intervention exclusively for framing, judgement, and interpretation. By “integrated,” I mean a single, complete lead-to-cash motion unified by one common taxonomy, where every step of the customer journey—from marketing-qualified lead, to commissioning, delivery, invoicing, and ultimately cash collection—is fully instrumented, credit-metered, transparent, and auditable.

Before presenting the blueprint itself, let’s first consider why achieving this transformation inherently requires a far-reaching approach across commercial, operational, technical, and human dimensions.

The Unpickable Knot

The reasons most research agencies remain trapped inside a labour-heavy, bespoke model are not a single problem, but rather an unpickable knot of commercial, operational, technical, and human constraints.

Commercial tension is the first knot. Clients still expect methodological certainty that historically came with six-figure budgets, yet the funding available for any individual study continues to shrink. Briefs are often incomplete or in flux because the client’s own stakeholders cannot agree on objectives or timing. When combined with accelerated launch dates, this widening gap between expectations and means repeatedly pushes work into the red.

Manual execution tightens the knot further. Each new project begins with a scoping call led by a researcher skilled enough to convert an ambiguous ask into a defensible study plan. Libraries of past questionnaires lie scattered across Word files and personal folders, which means setup often starts from a blank page. Late starts and mid-field changes—triggered by everything from sudden stakeholder additions to incorrect incidence assumptions—force analysts and project managers into urgent reprogramming, rapid sample adjustments, quota extensions, and more. Timelines compress, and individual acts of heroism become necessary to push the study over the line, adding untracked costs measured in monetary and psychological terms.

Technical debt compounds the problem. Most insights workbenches are collections of rigid survey platforms, legacy tabulation tools, Excel-based sample plans, and desktop slide templates that rarely communicate with each other. Key information about questions, scales, samples, and costs lives isolated in different systems, preventing the easy application of automated coding schemes or normative databases. Without a unified taxonomy or orchestration layer, every automation initiative stalls precisely at these disconnected hand-off points.

Human factors complete the picture. Researchers enter the profession to interpret behaviour, yet spend much of their time rewriting questionnaires or troubleshooting data problems. Operations staff find themselves firefighting fieldwork issues or performing basic data manipulations rather than improving process quality. On the client side, insight professionals seek reassurance that methods are sound, while their internal business stakeholders just want to move quickly to the next step. Nobody enjoys the manual, inefficient process, yet piecemeal attempts at improvement fail: genuine transformation requires sustained, full-company commitment.

Transformation Principles

The following principles serve as the North Star for transformation, ensuring each design decision aims high toward the desired outcome and is not compromised by short-term constraints or lack of organisational will.

Configurable, not customised. Every recurring study type—concept test, U&A, brand tracker—becomes a recipe with declared inputs and outputs. Machines assemble these recipes, changing parameters without rewriting code.

Machines act first. Software handles essential scaffolding: initial briefs, instrument design, sample design, and baseline reporting. Humans appear only where professional judgement meaningfully raises confidence: probing assumptions, approving edge cases, or providing additional interpretation of results. Any unnecessary human intervention increases costs and introduces inconsistency.

Taxonomy at the core. A knowledge graph stores constructs, question wordings, sampling frames, risk tags, allowed parameter ranges, client preferences, tariffs and pricing, contracting, and more. It drives everything from sales scripts to study designs to operational handling to invoicing and auditing, ensuring alignment across all teams and eliminating ambiguity.

Single credit ledger. Credits attach value to every configurable element—sample units, analytics modules, expert time, post-project workshops—so scope, cost, and timeline surface in real time, not through back-office guesswork or assumptions.

Adaptive interfaces. Whether via email, voice, chat, web portal, or API, all exchanges with clients route through the same orchestration and configuration technology. The system enriches and validates each input, presenting only the controls relevant to that buyer persona. This reduces client friction and prevents duplicated effort internally.

Lead-to-cash continuity. Every artefact—data, configurations, instruments, presentations, credit debits—travels intact with the project until cash is posted, eliminating broken context, manual reconciliations, and duplicate entries.

A Single Integrated Motion

To unlock the economics of the new operating model, execution and technology must merge into one closed loop. Let’s start by explicitly identifying the elements of a typical agency operating model:

Lead → Order → Delivery → Invoice → Cash

Lead to order covers marketing and sales work, moving from awareness creation through to deal closure and project commission. The closing event then takes us to the delivery phase, where project specifications are finalized and a combination of humans and machines produce the commissioned work. With work delivered, the client is invoiced and ultimately pays.

This operating view highlights how individual functions must coordinate to deliver outcomes to clients. For instance, the terms “product” or “technology” do not explicitly appear here, yet their presence and coordination are fundamental. This operating model neatly links the company’s commercial motions with the underlying executional work.

Technology, Services, and Orchestration

A modern insights agency must leverage technology not merely to achieve efficiency but also to create a robust foundation for growth and innovation. The new technology stack will serve as the business’s central nervous system, built upon two key elements: an Orchestrator and a Knowledge Graph.

The Orchestrator

The Orchestrator is a slim cloud service holding the study specification and moving it seamlessly through three interconnected zones:

• Configurator (Lead to Order): Converts inbound signals into structured briefs, matches them to the nearest method template or recipe, prices the work in credits, freezes the specification upon client approval, and hands off to production.
• Execution Rail (Order to Delivery): Consumes the specification to auto-build instruments, order samples, process data, and draft deliverables via micro-services that draw rules from the shared Knowledge Graph.
• ERP Link (Delivery to Cash): Mirrors every credit debit, contract, change order, and acceptance event, and archives the complete project bundle for auditing and compliance.

Knowledge Graph

The Knowledge Graph holds the entirety of the agency’s experiential, methodological, and operational assets—from constructs, questions, responses, sample designs, and templates, to legal constraints and credit pricing. A common service layer exposes this knowledge to modules that evaluate changes (diff gate), manage risk, generate reports, and more. Security, versioning, and audit modules manage permissions and create audit trails, ensuring traceability of every decision.

Building the Knowledge Graph this way enables the business to leverage its accumulated expertise into new products and services, such as:

• Derivative products created from existing data, avoiding unnecessary new data collection.
• Norms developed from studies in advertising effectiveness or brand lift.
• Synthetic panel applications.
• And more…

A Simplified Workflow within the Stack

The Configurator wakes the moment a lead lands—email, web form, voicemail transcript, or API push—enriches it with firmographics and interaction history, maps it to a recipe, and surfaces real-time costs, timelines, and risks. Once signed, it streams the locked specification to the Execution Rail, monitors a diff gate for mid-stream changes, and triggers expert review only when a material risk is identified. A live credit ledger remains in lock-step with ERP, and a compliance vault captures every artefact—specifications, data, decks, invoices—under a unified project ID.

Below is a more detailed process flow.

Operating Flow

Go-to-Market (Lead to Order)

1. Signal captured

An email, voicemail, demo request, or social interaction enters CRM. Enrichment services append firmographics, buying stage, and prior interactions. The Configurator, integrated with CRM, assigns a provisional recipe and credit range before any human intervention.

2. Smart invitation:

The Configurator offers two paths depending on buyer characteristics and preferences:

• Schedule a scoping call with a sales host.
• Continue in a self-guided, structured flow.

For smaller clients, this can be fully self-service. Enterprises typically interact through a customer-success host who absorbs context and shields buyers from configuration minutiae.

3. Scoping conversation

During the call, the sales host speaks freely as the Configurator listens, transcribes, and tags objectives, audience, geography, deadlines, legacy practices, legal constraints, and more. Missing fields appear as silent prompts, guiding the host’s questions.

4. Live specification & risk log

As soon as mandatory attributes are complete, the Configurator instantly assembles a study specification and risk log. Both appear in real-time, showing estimated credits, delivery date, method recipe, and flags like “sample size under-powered.” If no high-risk flags arise, the client approves immediately; credits debit automatically.

Credit-based pricing mirrors procurement needs: smaller firms buy credit bundles online, while enterprises sign recurring contracts that top up quarterly. Credits can be represented in local currency.

5. Expert gate (for risk behind a certain threshold)

If flagged as “review required”—due to conflicting scales, sensitive PII, or unrealistic timelines—the client sees “Under review · SLA 24 hrs.” A researcher amends or approves the brief, and the updated spec returns to the client within 24 hours. Approval triggers automatic credit debits. (Expert review access can be gated to high-value clients.)

6. Contract & finance handshake

Approved specs move automatically into ERP. Contract terms populate a Statement of Work from templates tied to credit volume and geography, and DocuSign executes without manual editing. ERP immediately books revenue upon client signature.

7. Handoff to execution rail

Upon order completion, the Configurator triggers production—instrument creation, sample ordering, data QC, and reporting—as described next. Clients receive milestone notifications until delivery.

This Go-to-Market loop eliminates lengthy proposals, maintains buyer clarity and simplicity, and ensures all decisions regarding scope, cost, and risk happen transparently before the project begins.

Execution (Order to Delivery)

1. Launch signal

Once signed, the Configurator transmits the locked spec, risk log, budget, and delivery SLA to production.

2. Instrument generator

A micro-service auto-assembles the questionnaire from the taxonomy—canonical wording, routing logic, scales, and required client-specific items. Compliance checks run automatically, and a PDF proof is archived.

3. Sample procurement & live QC

Sample specs transmit via panel APIs. Completes return through real-time fraud and speed checks. If incidence drifts, the service proposes remedies and incremental-credit quotes before timelines suffer.

4. Built-in change control

Mid-stream requests enter a diff engine that evaluates cost, timeline, and data quality impacts. Significant changes debit credits or prompt expert review.

5. Data landing on governed schema

Every question in the library has stable IDs and processing rules, enabling instant data processing. The ETL service automatically maps responses, cleans data, generates weights, codes open-ends, and aligns results to tracker histories and norms.

6. Automated analytics & narrative drafts

Descriptive statistics, weighting, significance tests, and norms populate via analytics micro-services. Reports produce interactive dashboards and draft decks that auto-update.

7. Optional Confidence Call

If purchased, a researcher refines the deck and hosts a live client call. The presented version is archived.

8. Delivery

Completed deliverables—dashboards, decks, background files—publish via Configurator. Clients receive timestamped notifications, automatically recording acceptance.

Close and Bill (Delivery to Invoice to Cash)

1. Deliverables accepted or challenged

Clients receive notifications upon publication and may raise issues (defects or scope changes) within a specified window, routed via a built-in triage system.

2. Ledger finalisation

Acceptance closes the provisional ledger, reconciling variances automatically. Unused credits revert; overruns debit instantly. Signed project bundles flow into ERP.

3. Invoice issued

Prepaid credit-pack clients require no invoice; ledger entries suffice. For pay-as-you-go, ERP auto-generates invoices with supporting project files attached, requiring no manual reconciliation.

4. Payment applied

Payments match invoices automatically via project IDs and credits. Exceptions or late payments trigger finance alerts.

5. Feedback and upsell loop

Post-delivery, a brief “impact pulse” survey collects satisfaction metrics and ROI signals, triggering follow-up or feeding into product backlog refinement.

6. Compliance archive

All artefacts—instrument proofs, sample specs, reports, invoices—archive immutably, inheriting permissions from the credit ledger, providing traceable records without manual intervention.

The entire lead-to-cash journey completes without spreadsheets, orphaned invoices, or manual write-offs. All financial, methodological, and compliance details reside in one continuous, traceable system of record.

Strategic Workforce Planning

The new stack cannot take root unless the organisation’s people map cleanly to the new workflows. The operating model requires four clearly defined talent lanes:

1. Growth and Market Activation (Revenue, Sales, Marketing)

Marketing focuses exclusively on demand-generation, where collateral such as white-papers, case studies, and ROI calculators help buyers move themselves down the funnel, ultimately channelling them into the Configurator. Marketing’s primary success metric is qualified opportunities (MQLs).

The Sales team operates in three distinct roles:

• Hunters source and negotiate credit frameworks with enterprise clients. Their variable compensation ties to annual credit throughput rather than margin on individual projects. A small group of senior hunters handles complex, high-value enterprise relationships.
• Farmers retain and ideally expand spend within existing accounts. Their variable compensation ties to a blended revenue retention and expansion target.
• CSMs conduct project scoping conversations, guiding buyers through the intake flow and helping to close orders. Think of them as the agency’s “sales engineers”—commercially minded professionals who understand enough about research to assist clients at the bottom of the funnel and facilitate final approvals.

2. Insight and Advisory (Research & Insights)

Researchers no longer build questionnaires or clean data. Instead, they perform one or more of the following functions:

• Curating the knowledge graph, enhancing taxonomies or refining recipe cards.
• Providing expert advice directly to clients who purchase advisory services.
• Acting as thought leaders within the insights industry or in their clients’ industries.

Junior hires typically begin as taxonomy stewards and may progress into narrative roles, thought leadership, or method governance for an entire study family. Workload and effectiveness are measured based on contributions to enhancing the knowledge graph, providing advisory support, and developing thought leadership—not on project management or traditional operational outputs.

3. Platform Reliability and Data Quality (Operations)

Operations transforms into a reliability engineering team. They monitor service health, investigate API failures, triage data anomalies, tackle escalated fieldwork issues, and communicate fixes back to engineering and product teams. Primary KPIs include uptime, incident resolution time, and defect recurrence rates. Capacity planning is based on error budgets and cycle time, not project count.

4. Product, Engineering, and Data Science (Product & Engineering)

Product management owns the knowledge graph, credit tariffs, and both internal and external user experiences of the platform. Engineering manages micro-services, version control, security, and auditing. Data science fine-tunes machine-assistance modules—such as semantic similarity search, power analysis, and narrative drafting—and validates them against expert-generated output. Their shared goal is reducing human intervention per study while maintaining or improving decision quality.

With the workforce clearly re-anchored around these four strategic lanes—growth, curation, reliability, and platform evolution—the organisation can scale credit volume (revenue) without maintaining battalions of manual project workers (cost). The executive team’s role is to ensure alignment across these lanes, manage cultural change proactively, and navigate the organisation smoothly through transformation.

Goal Setting and Reporting

While a good blueprint provides structure, it must also give leaders the visibility and flexibility they need to steer the business. Goals will differ by organisation, by leadership style, by market conditions, and by strategic intent. The real strength of the operating model described here is its built-in capability to provide clear, actionable data—regardless of which goals you choose.

When each stage of your business—lead to order, order to delivery, delivery to cash—is instrumented end-to-end, reporting becomes automatic rather than a manual, retrospective exercise. Leadership can choose and refine goals as market conditions change, and because the underlying telemetry stays constant, you will always see precisely how your business performs against those goals.

For example, operational teams might prioritise cycle times or human minutes per project, using these metrics to pinpoint bottlenecks or opportunities for automation. Sales might focus on net revenue retention, expansion metrics, or discount utilisation. Finance teams may look closely at contribution margins, costs per delivered credit, or accounts receivable aging. Executives and boards will likely favour aggregated views, focusing on net growth, margin performance, strategic client health, and progress against key automation initiatives.

Regardless of the metrics you emphasise, the blueprint ensures the numbers are consistent, auditable, and timely. The integrated reporting layer means that operational, commercial, financial, and executive reporting all draw from the same real-time dataset—no spreadsheets, no reconciliations, no ambiguity.

In short, the operating model provides a stable foundation on which to set clear, flexible, and measurable goals. Leaders gain confidence in their ability to steer the business, respond swiftly to change, and remain accountable for outcomes—whatever those outcomes may be.

Planning and Funding the Transformation

No two organisations begin this journey from exactly the same place. Each business will have its unique strengths and weaknesses, legacy systems to phase out, and foundational capabilities to strengthen before scaling. A practical migration plan will reflect these realities, laying out clearly where major efforts are required and where only minor adjustments may suffice.

One of the core challenges in any ambitious transformation—especially within publicly traded businesses—is the necessity of operating the legacy model and the future state simultaneously. While building new rails and processes, the current business must continue running without disruption. Doing this effectively means executive commitment must be total. Leaders should anticipate the complexity, resource the initiative fully, and be ready to reinforce priorities continuously.

From a funding perspective, businesses must be strategic in balancing immediate P&L pressures against the investment required. Some parts of the initiative, such as substantial platform builds or significant infrastructure work, may qualify for balance-sheet funding (capex), while others—operational changes, training, or incremental improvements—will impact operating expenses. The involvement of the CFO from day one is therefore critical. Beyond mere budgeting, finance leadership must help decide how to best fund, stage, and track progress throughout the transformation.

Additionally, careful sequencing and prioritisation are vital. Organisations rarely have the luxury to implement everything at once. Early wins that build internal credibility, tangible efficiency gains, and risk-reduction strategies can help ensure the initiative keeps momentum and maintains support.

Finally, thoughtful risk management and clear guardrails are non-negotiable. Operating parallel models introduces complexity and can heighten short-term operational risk. Leadership must define clear risk thresholds, specify rollback conditions, and monitor these proactively to ensure that short-term execution does not slip.

With clear planning, sufficient resourcing, close involvement from finance leadership, disciplined sequencing, and structured risk management, the transformation from the current model to the future blueprint becomes achievable—even while the business remains fully operational.

Conclusion

The transformation outlined in this blueprint is ambitious—but it’s also both necessary and achievable.

The insights industry today stands at a crossroads. The traditional consulting-driven model, heavy with manual intervention, inconsistent quality, and structural inefficiencies, cannot survive intact in the era of rapid technological evolution and shifting client expectations. Yet the solution isn’t as simple as bolting on technology or pushing the workload onto the client through DIY tools. Instead, genuine transformation requires rebuilding the core of the business—rethinking product design, redefining the go-to-market strategy, realigning people, and implementing technology that orchestrates it all seamlessly.

The integrated stack described here provides the structural backbone to do just that: configuring and automating recurring processes, embedding methodological rigour into machine-assisted flows, and ensuring visibility at every point—from the first marketing touchpoint through to final cash reconciliation. With human expertise focused solely on high-value framing, judgement, and interpretation, and technology managing repeatable execution, agencies gain scalability, efficiency, and the flexibility to innovate.

But as powerful as technology can be, the journey remains inherently human. Strategic workforce planning, cultural alignment, clear goal setting, vigilant reporting, and, above all, total executive commitment are non-negotiable components of this transformation. Indeed, your talent—fully aligned, inspired, and acting in concert—will ultimately determine whether the blueprint translates into reality.

The choice ahead is clear: agencies can either continue struggling within the constraints of legacy operating models—the unpickable knots, incremental tweaks, and operational heroics— or replace them with clarity, confidence, and sustainable scale.

The blueprint presented here provides a structured framework—principles, roles, an integrated operating model, and aligned technology—that agencies can adapt thoughtfully to their own context. Each agency’s transformation will look slightly different, reflecting unique strengths, constraints, and goals. But the underlying direction remains consistent: toward a leaner, integrated operating model explicitly built to manage complexity, maintain margins, and create lasting value in an increasingly uncertain market.