The Compounding Firm: Engineering Network Effects Into an Expertise Business

Certifying people to deliver your methodology feels like building a platform. Most of the time, it isn't one. It's a franchise — a pipeline business wearing a platform costume.

Alex Moazed and Nicholas Johnson draw the line sharply in Modern Monopolies: a great framework plus trained deliverers is still a linear business. The shift to platform happens only when network effects switch on — when the ecosystem grows more valuable to every participant each time a new participant joins.

That one sentence is the entire game for an expertise business that wants to compound.

Geoffrey Parker, Marshall Van Alstyne, and Sangeet Paul Choudary make the same distinction in Platform Revolution. A pipeline creates value in a straight line: you produce the work, you sell the work, the client consumes the work. A platform creates value by orchestrating exchanges between producers and consumers. The platform itself doesn't perform the service — it makes the service findable, matchable, and trustworthy, and it earns from every exchange it facilitates.

Software companies get network effects by shipping code. A consultancy, agency, or training company has to get them by deliberate design, because the raw material is people and relationships rather than pixels. The design has three load-bearing parts: a core transaction that completes without the founder, a dataset that compounds, and a deliverable engineered to travel. Work through them in that order.

01 — The Transaction Test

Does Value Exchange Without You in the Room?

Moazed and Johnson insist that every platform stands on a Core Transaction: one repeatable exchange of value that the platform exists to make happen. Until you can name yours, every other network-effects conversation is premature. For a methodology business, the transaction breaks into four moves:

• Create. A certified practitioner puts themselves forward as available to deliver your methodology.
• Connect. A client runs your diagnostic assessment; its results surface their gaps and pair them with the practitioner whose specialism, region, and experience level fit.
• Consume. The practitioner delivers the transformation work to the client.
• Compensate. The client pays the practitioner, contributes feedback and case material, and their anonymized assessment data joins the benchmarking pool.

Read those four moves again and notice who is absent: you. That absence is the test. If client-practitioner matching still routes through your inbox, or no proposal goes out before you've reviewed it, the transaction is founder-dependent — and a founder-dependent transaction cannot compound, because you are its bottleneck.

Pass the test and something subtle changes. Clients now find practitioners through your ecosystem rather than through you. Practitioners pass work to one another. The diagnostic gets sharper with every completion. The business stops being a service you perform and starts being an exchange you govern.

02 — The Asset Nobody Can Reverse-Engineer

Why the Dataset Outvalues the Methodology

Your methodology can be imitated. Your assessment questions can be deconstructed by anyone who takes the assessment. What cannot be imitated is the accumulated record of every assessment ever completed across your network — structured, comparable, and growing daily. This is the data network effect, and in an expertise business it is both the strongest moat and the one founders most often neglect.

The flywheel turns like this:

• Turn 1: A practitioner assesses Client A, a financial services company.
• Turn 2: The anonymized results land in the benchmarking database.
• Turn 3: Client B, also in financial services, completes the assessment — and gets back not just a score but a position relative to industry peers.
• Turn 4: A benchmarked result carries far more weight than a raw number ever could.
• Turn 5: Client B takes the comparison to their board; the conversation sends Client C to get assessed.
• Turn 6: More completions enrich the benchmark, and a richer benchmark makes the next assessment worth more.

Choudary's phrase for this is "data is the new dollar." Parker, Van Alstyne, and Choudary describe the same dynamic as "demand economies of scale" — value that grows on the demand side as usage grows, rather than on the supply side as production grows. A rival can clone your framework over a weekend. Rebuilding thousands of comparable assessments across industries and regions would require rebuilding your entire ecosystem first.

Measure the flywheel in assessments per industry-geography segment. Statistically meaningful benchmarks need at least 10 assessments in a segment, so count the segments that have crossed that line. When 80% of your active segments carry meaningful benchmark data, the wheel is genuinely turning.

03 — Same-Side Gravity

A Bench That Trains Itself

Network effects come in two directions, and the easier one to start is same-side: participants on one side of the exchange making each other better. In practice, that means every practitioner in your network gains capability simply because the other practitioners exist.

Three mechanisms do the work:

• Collective pattern recognition. Put 50 practitioners into the field across different sectors and regions and the patterns they see together dwarf what any one of them could spot alone. The healthcare practitioner runs into a data-maturity pattern that a financial-services colleague already documented six months earlier — and the shared pattern library deepens with every engagement.
• Peer problem-solving. A monthly community call where practitioners trade what's working becomes a living knowledge base. The one wrestling with a resistant executive team in manufacturing hears from the one who cracked the identical problem in retail. That pooled experience makes everyone in the network sharper than an equivalent solo operator.
• Complementary specialization. Once practitioners specialize — by discipline, sector, or territory — competition turns into collaboration. The architecture expert hands data gaps to the data expert; the talent expert routes implementation work to the automation expert. Each specialist's position gets stronger as the network's coverage gets more complete.

"The honest gauge of same-side effects is cross-practitioner referral activity: referrals made, referrals received, referrals converted. Quarter-over-quarter growth means the network is alive. A flat line means you have isolated consultants renting a common logo."

04 — The Empty-Room Problem

Sequencing Supply Before Demand

Cross-side effects are the defensible kind: more practitioners produce better coverage by geography and specialty, better coverage produces better matches, better matches draw more clients, more clients pull in more practitioners. Past critical mass, the loop feeds itself.

Getting it started is the hard part, and the sequencing rules come from Andrew Chen's The Cold Start Problem. Build the supply side — practitioners — before the demand side, because a client who walks into an empty network leaves and rarely returns. Chen calls this the "moment opposite of magic": a manufacturing buyer who opens your practitioner directory, counts three names, and quietly closes the tab.

His remedy is "Flintstoning" — powering the machine with human legs until it can run on its own. Suppose you have 25 practitioners and only 3 of them know manufacturing. Don't let a manufacturing client browse the directory; route them by hand to the best fit and present it as concierge matching. Never let the client see the empty shelves.

Two corollaries follow. First, perfect a single Core Transaction before layering on a second one — a half-working exchange compounds nothing. Second, resist the urge to build platform software ahead of platform behavior. Technology should formalize and scale exchanges that are already occurring naturally between practitioners and clients; it cannot conjure them.

05 — Make the Deliverable Do the Marketing

Designing Results That Leave the Building

Choudary puts it bluntly: "Virality is a business design problem, not a marketing or engineering effort." Translated for an expertise business — your assessment result, not your ad budget, has to carry your growth. The artifact must be built to exit the platform and circulate.

Engineer the output for three rooms it will travel into:

• The boardroom. One visual page: maturity score, industry benchmark, three priority moves. It should slot into a board deck untouched.
• The feed. A finding that writes its own post. "We came in at 2.1 on data maturity. Our industry sits at 3.4. Here's how we're closing the gap."
• The stage. Anonymized findings practitioners can present. "European financial services firms average 2.8 on automation maturity — and the top quartile operates very differently." That's a keynote with built-in demand.

The loop those rooms create: a CEO encounters a benchmark, wants their own number, takes the assessment, and their data thickens the benchmark — which a peer CEO then encounters. Every revolution adds participants and data at once.

Building the loop takes five deliberate moves. One: shape the spreadable unit. A 50-page report doesn't circulate; a single-page maturity snapshot with a clear score and a handful of benchmark comparisons does — and it must provoke the question "where would we land?"

Two: build sharing into the artifact itself. Every result ships with its own URL, a social-ready image, and a pre-written summary the client can paste. Clients won't draft their own narrative; hand them the words.

Three: publish a public data product. An openly released benchmark report — the state of maturity in your industry, issued quarterly or annually — sits at the top of the viral funnel, drawing press, speaking slots, and search traffic.

Four: turn practitioners into amplifiers. A practitioner presenting case data at a conference, posting a benchmark insight, or walking a prospect through a trend chart is distribution you didn't pay for. Equip them with shareable assets, not only client deliverables.

Five: close every loop. The public report ends in "Get Your Own Score." The post links to the assessment. The talk ends with "Find a Certified Practitioner in Your Region." A viral artifact with no path back to the ecosystem is reach without compounding.

"Watch the viral coefficient: the share of new assessment completions arriving via referral or sharing rather than paid acquisition. Above 30%, the assessment spreads on its own. Below 10%, what looks like a product is really a marketing dependency."

06 — Six Names, One Pattern

Who Crossed From Franchise to Platform — and Who Didn't

None of this is theoretical. Study the methodology businesses that made the crossing — and the one that stayed behind — and the pattern repeats.

• EOS. Gino Wickman codified a way to run a company, then certified implementers to deliver it. Hundreds of EOS Implementers now serve 200,000+ companies, and the implementer network, accumulated assessment data, and community of EOS-run businesses generate real network effects. Platform.
• Gallup (CliftonStrengths). Tens of thousands of certified coaches plus the largest strengths dataset on earth — 30+ million completed assessments — plus tooling built around the instrument. Arguably the deepest data moat any methodology business has dug; no challenger replicates 30 million data points.
• SAFe. Certification, a surrounding tool ecosystem, and accumulated implementation data. Each new SAFe practitioner expands organizational adoption, which expands demand for practitioners. The loop is live.
• FranklinCovey. Stephen Covey's 7 Habits became the seed for a certified facilitator network, a subscription content platform (the All Access Pass), and organizational-effectiveness data spanning thousands of clients. Facilitator breadth plus subscription economics produce genuine cross-side pull.
• StoryBrand. Donald Miller trained guides on his messaging framework — but the decisive move was the software the certified guides run with their clients. The tooling is what produced the data flywheel and the switching costs.
• Sandler Training. Franchisees have delivered David Sandler's sales methodology since 1967 — one of the longest-running methodology-as-franchise models anywhere. Yet the platform step, aggregating sales performance data across all franchisees' clients, never happened at scale. Sandler remains the cautionary half of the pattern: franchise, not platform.

Every crossing shares the same three ingredients: a standardized diagnostic, an accumulating dataset, and a practitioner network big enough for matching and referrals to mean something. Training deliverers is the entry fee. Wiring them into a network whose collective activity compounds is the business.

07 — When the Network Starts Pulling

Reading the Tipping Point Honestly

Chen describes a tipping point past which a network sustains itself — new participants arrive because the network is valuable, not because your marketing reached them. For a service methodology network, the thresholds look like this:

• Supply: 50+ active practitioners with enough specialization spread that 80% of client inquiries get matched within 48 hours.
• Demand: 500+ completed assessments — sufficient for statistically meaningful benchmarks across your top 5-8 industry segments.
• Data: A dataset deep enough to publish a credible annual state-of-the-industry report that earns coverage and inbound interest on its own.
• Referrals: 30%+ of new clients arriving via practitioner referrals or shared assessment results.

Below those lines, you push. Above them, the network pulls. Founders who have felt the change say it's impossible to mistake.

Three confirmations that you've crossed:

• Inbound overtakes outbound. Practitioners and clients find you faster than you find them, and marketing spend falls under 15% of revenue.
• Content appears without an editorial calendar. Practitioners publish case studies, benchmark insights, and commentary unprompted. The network produces its own thought leadership.
• Growth problems give way to quality problems. The daily question stops being "where do we find practitioners and clients?" and becomes "how do we hold quality while demand outruns capacity?" There is no better problem.

Expect the crossing to take 18-36 months, in three recognizable phases. The first 12 months run in franchise mode: you train practitioners, they deliver linearly. Months 12-24 are network mode: practitioner-to-practitioner referrals begin, shared learning speeds up, community value becomes tangible. Months 24-36 are platform mode: cross-side effects activate, the data becomes a product, and the system starts producing value you personally never created.

"Certification count is a vanity metric. Count completed engagements instead — full diagnostic-to-transformation cycles, not badges issued or assessments started. One practitioner who delivers 20 engagements a year contributes more to the network than five who deliver none."

And resist shortcuts. Don't ship technology before the behavior it formalizes exists. Don't open a second transaction before the first one is reliable. And never let a client wander into an empty room.

Compounding was never reserved for software. It belongs to any founder willing to stop being the service and start governing the ecosystem that delivers it.