Toll economics: why some businesses keep compounding
The customer makes another payment, runs another workflow, or renews another essential service. A well-positioned business gets paid again. Over time, that repetition can becomes a powerful growth flywheel.
Long-term stock-market winners need more than a good quarter. They need an economic engine that can keep producing value after the excitement of the first purchase has passed.
One useful place to look is toll economics: businesses that earn repeat revenue because customers keep using a network, renewing access, or relying on a product already embedded in their operations. The “toll” may be a transaction fee, a software subscription, a licensing payment, a registry renewal, or replacement products sold into an installed base.
The appeal is not simply that revenue repeats. It is that the company may avoid having to win the entire relationship again with every sale. If customers receive continuing value, changing providers is costly, and the market keeps expanding, an established relationship can support years of additional revenue.
What turns a repeat payment into a compounding engine?
A subscription button alone does not create a moat. A compelling toll usually combines customer dependence with room to grow: the product does something important, remains economical relative to its value, and benefits as its customers transact, expand, or do more work.
That can improve the economics of growth. An installed network or software platform may serve additional activity without an equivalent increase in cost. Cash generated today can fund product improvements, distribution, acquisitions, or repurchases. When those investments earn attractive returns, the next year's earning base grows.
A mechanism to investigate, not an automatic outcome. Each link can weaken through competition, churn, regulation, costs, or poor capital allocation.
The important distinction is between repeat purchasing and durable economics. A business can have loyal customers and still face shrinking prices, rising service costs, heavy capital requirements, or low returns on reinvestment. The toll becomes interesting when repeat demand translates into improving economics for each share.
Ten historical examples of the engine at work
These companies illustrate several forms of toll economics. Across fiscal 2014–2024, their reported revenue compounded at roughly 4%–12% a year. That is business growth, not a stock-return calculation. Several businesses also expanded through acquisitions.
| Business | The repeat-revenue mechanism | FY2014 → FY2024 revenue | Annual growth | Sources |
|---|---|---|---|---|
| VisaV | Service, processing and cross-border fees let Visa earn from payment activity across its network. | $12.70bn → $35.93bn | 11.0% | 2014 · 2024 |
| MastercardMA | Network assessments and transaction-processing fees scale with payment volumes and transactions. | $9.47bn → $28.17bn | 11.5% | 2014 · 2024 |
| MSCIMSCI | Recurring index/data subscriptions and asset-based index licensing fees monetize investment workflows. | $1.00bn → $2.86bn | 11.1% | 2014 · 2024 |
| S&P GlobalSPGI | Ratings, benchmark licensing and financial-data subscriptions earn fees around capital-market activity. | $5.05bn → $14.21bn | 10.9% | 2014 · 2024 |
| Moody’sMCO | Credit ratings and recurring analytics services monetize issuance and risk-management needs. | $3.33bn → $7.09bn | 7.8% | 2014 · 2024 |
| CME GroupCME | Clearing and transaction fees monetize trading activity; market-data subscriptions add recurring revenue. | $3.11bn → $6.13bn | 7.0% | 2014 · 2024 |
| Intercontinental ExchangeICE | Exchange, clearing, data and workflow services collect fees across financial-market infrastructure. | $3.09bn → $9.28bn | 11.6% | 2014 · 2024 |
| VerisignVRSN | Domain-registry fees recur as registrars register and renew names in the .com and .net registries. | $1.01bn → $1.56bn | 4.4% | 2014 · 2024 |
| ADPADP | Recurring payroll and human-capital-management work embeds ADP in customers’ regular operations. | $10.23bn → $19.20bn | 6.5% | 2014 · 2024 |
| Roper TechnologiesROP | Specialized software and technology businesses earn repeat payments inside customer workflows; Roper reinvests cash through acquisitions. | $3.55bn → $7.04bn | 7.1% | 2014 · 2024 |
USD billions, rounded. Annual growth = (FY2024 revenue ÷ FY2014 revenue)1/10 − 1. ICE uses revenue less transaction-based expenses. Visa fiscal years end September 30; ADP June 30; the other examples December 31. This deliberately fixed historical decade is not a presentation of latest financial results.
Acquisitions, changing business scope, and other comparison notes
- Visa: Visa Europe acquisition and other acquisitions contribute; growth is not wholly organic.
- Mastercard: Rebates and incentives reduce revenue; acquisitions and new services also contribute.
- MSCI: Asset-linked fees vary with markets and fund flows; acquisitions also contribute.
- S&P Global: 2014 predecessor is McGraw Hill Financial; IHS Markit merger and divestitures materially change scope. Not like-for-like organic growth.
- Moody’s: Ratings revenue is issuance-sensitive; acquisitions and analytics expansion contribute.
- CME Group: Revenue varies with volumes and mix; acquisitions contribute.
- Intercontinental Exchange: Both figures are revenue less transaction-based expenses. Acquisitions, including mortgage technology, materially expand scope.
- Verisign: Registry contracts and pricing terms matter; renewals and domain demand are not assured.
- ADP: 2014 uses continuing operations presented in the 2015 release after the CDK spin-off. PEO revenue and client-funds interest mean revenue is not purely software fees.
- Roper Technologies: Substantial divestitures and acquisitions mean the endpoint revenue comparison is especially imperfect. Its company-reported adjusted FCF history below better explains its reinvestment model.
The variety matters. Visa and Mastercard collect fees around payment activity. Exchanges earn from trading and clearing. Index providers license benchmarks and data. A registry earns when a domain is registered or renewed. Payroll and specialized software providers earn within recurring business processes. The common feature is an enduring relationship with activity the customer continues to need.
For Roper, cash generation adds another perspective: company-reported adjusted free cash flow rose from $800 million in 2014 to $2.282 billion in 2024—approximately 11% annual growth. That non-GAAP measure and the revenue history both reflect a changing portfolio of businesses, rather than purely organic growth. Roper 2025 proxy, cash-flow reconciliation.
Twenty recent winners—and the economics behind the next question
Historical compounders show why a repeat-revenue engine deserves attention. Recent winners show why that engine belongs within a broader model. The following 20 stocks more than doubled over two years, across technology, power, industrials, materials, and refining.
We selected these examples by their strong past returns and the availability of a measured toll-economics score, then inspected their current readings. This is a retrospective study of winners, not a list the model selected at the start of the period. The first five identities are reserved to preserve the value of member research.
| Row | Company | Business group | 2-year price change | Current toll score /100 |
|---|---|---|---|---|
| 01 | Member reserved 1 | Technology / industrials | +2,594.8% | 62.54 measured inputs |
| 02 | Member reserved 2 | Technology / industrials | +1,569.9% | 80.43 measured inputs |
| 03 | Member reserved 3 | Technology / industrials | +983.3% | 84.74 measured inputs |
| 04 | Member reserved 4 | Technology / industrials | +705.8% | 100.03 measured inputs |
| 05 | Member reserved 5 | Technology / industrials | +669.3% | 71.12 measured inputs |
| 06 | RKLBRocket Lab Corp | Space systems | +661.3%Price history | 64.33 measured inputs |
| 07 | CRDOCredo Technology Group Holding Ltd | Connectivity semiconductors | +618.5%Price history | 100.02 measured inputs |
| 08 | PLTRPalantir Technologies Inc. | Enterprise software | +403.5%Price history | 84.74 measured inputs |
| 09 | PARRPAR PACIFIC HOLDINGS, INC. | Oil refining | +363.8%Price history | 71.12 measured inputs |
| 10 | DKDelek US Holdings, Inc. | Oil refining | +330.6%Price history | 100.02 measured inputs |
| 11 | ARISAris Mining Corp | Gold mining | +307.6%Price history | 71.12 measured inputs |
| 12 | GEVGE Vernova Inc. | Power equipment | +293.5%Price history | 80.43 measured inputs |
| 13 | COHRCOHERENT CORP. | Photonics | +268.5%Price history | 80.43 measured inputs |
| 14 | APHAMPHENOL CORP /DE/ | Electronic connectors | +184.1%Price history | 100.02 measured inputs |
| 15 | MPMP Materials Corp. / DE | Rare-earth materials | +176.9%Price history | 71.12 measured inputs |
| 16 | MPCMarathon Petroleum Corp | Oil refining / midstream | +165.6%Price history | 71.12 measured inputs |
| 17 | POWLPOWELL INDUSTRIES INC | Electrical equipment | +155.7%Price history | 84.74 measured inputs |
| 18 | VRTVertiv Holdings Co | Data-center infrastructure | +147.2%Price history | 66.13 measured inputs |
| 19 | AVGOBroadcom Inc. | Semiconductors / software | +111.7%Price history | 80.43 measured inputs |
| 20 | PSXPhillips 66 | Oil refining / midstream | +110.3%Price history | 71.12 measured inputs |
Price period: October 2, 2024–October 2, 2026. Cumulative change in Yahoo adjusted closing prices, including the provider's dividend and split adjustments. Scores: October 8, 2026 research snapshot. Rows are ordered by past return, not current model rank. Business groups are editorial labels.
A strong price move prompts the next research question: what can sustain the business from here? Toll economics helps examine repeat demand and customer dependence. Forward demand, capital returns, cash generation, valuation, and market behavior help distinguish a durable opportunity from a temporary surge.
A refinery or miner can appreciate sharply through favorable prices, margins, or capacity conditions without resembling a classic subscription or transaction-fee business. Conversely, recurring revenue does not guarantee a strong return over every holding period. The model's value is in examining those different economic drivers together.
How to interpret this winners table
The initial read-only scan found 5,913 active companies with both stored price endpoints. The 20 examples were chosen to span several business groups, each with more than 100% adjusted-price appreciation and an available toll score. Each selected price pair was fetched directly from Yahoo. This is an illustrative sample, not the market's top 20 or a complete, survivorship-free historical universe. No minimum toll score was imposed.
Current scores describe the latest available reported inputs, including older reporting periods. They cannot establish what the model knew two years earlier. The table does not attribute the gains to toll economics or show that Matterhorn identified these stocks before they rose.
These readings have two to four measured inputs, frequently gross-margin trend, R&D efficiency, or customer concentration. None of these selected snapshots has a passing direct recurring-revenue or retention observation in its toll lens. Some have neutral recurring-revenue estimates. A high reading summarizes the available inputs; it does not independently prove a subscription-like business model.
The toll lens applies existing checklist weights: pass earns full credit, neutral half, and fail zero, normalized across its evaluated inputs. This explanatory lens uses different arithmetic from the overall score and must not be read as a percentage contribution to it.
One lens does not describe the whole company.
Conceptual diagram—not a company score, probability, or weighting chart. The spokes connect five research lenses with the Proprietary Algo; they do not represent six equally weighted factors.
Toll Economics
Why customers keep paying. Recurring revenue, retention, switching friction, pricing power, and products designed into customer workflows.
Forward Book
Where future business may come from. Backlog, remaining performance obligations, orders, bookings, and recurring-business growth. These indicate demand; they are not guaranteed future revenue.
Fundamentals
Whether the business is strengthening. Revenue and profit growth, margins, returns on capital, leverage, and dilution. These test whether growth creates durable per-share value.
Cash Generation
How much growth turns into cash. Free-cash-flow growth, margins and yield, reinvestment, and repurchases relative to stock-based compensation. These examine what remains after the costs of operating and investing.
Market Confirmation
How the market is pricing the case. Relative strength, price and risk behavior, institutional accumulation, and valuation measures. These provide context on recognition and expectations.
Proprietary Algo
How the analysis comes together. The ML-enhanced model blends 70% financial-component analysis with 30% learned-pattern ranking to help prioritize research.
The five lenses organize overlapping business and market inputs. Proprietary Algo adds the model perspective; it is not a sixth checklist lens. The 70/30 blend is explained below.
Proprietary Algo: combining financial strength with learned patterns
Toll economics helps explain how a business can keep earning. Matterhorn brings that perspective together with deeper financial analysis and patterns learned from historical company performance.
In the ML-enhanced model, the overall score uses a 70/30 blend:
| Proprietary Algo | Weight | What it contributes |
|---|---|---|
| Financial-component analysis | 70% | Growth, forward demand, quality, market confirmation, valuation, data support, and a separate durability estimate. |
| Learned-pattern ranking | 30% | Historical combinations of company characteristics associated with stronger subsequent returns relative to sector and size peers. |
The financial analysis examines the business across multiple dimensions. The learned-pattern component examines how combinations of those characteristics have related to subsequent performance. Together, they help prioritize companies for further research.
Toll economics remains a useful perspective within that broader analysis. Repeat demand matters most when the company can translate it into growth, attractive capital returns, and cash generation—and when the investment price makes sense.
The 70/30 split describes Matterhorn’s ML-enhanced Beta score. Its financial-component score is distinct from the standard five-component score. The spider diagram shows research perspectives and the algorithm together; its spokes are not equal weights or additional independent contributions. The current toll readings in the table are lens scores, not ML-enhanced scores or historical predictions. The blend does not attribute past stock returns to individual inputs.
One lens within a broader investment model
Toll economics is one of five investor lenses in Matterhorn, drawing on an industry-specific set of signals that we analyze alongside growth, forward demand, cash generation, valuation, and market behavior.
The lenses reorganize existing checklist results; they are not five additional, independent votes. Their inputs can overlap, and the appropriate tests differ by industry. There is no single universal signal count that accurately describes every company.
The objective is deeper analysis of the economics that can sustain an investment: how reliably customers return, whether revenue becomes cash, how efficiently that cash is reinvested, and what expectations the share price already contains.
For a potential long-term runner, a compelling combination is repeat demand, attractive incremental returns, expanding opportunity, and a sensible entry valuation. Toll economics helps identify where repeat demand may come from. The rest of the research determines whether that demand can reward shareholders.
Look beyond the headline growth rate.
Explore how Matterhorn brings business economics, financial trends, and market behavior into one investment-research workflow.
Explore Matterhorn →Historical examples and screened companies illustrate a research process. They are not personalized investment recommendations. Historical returns do not establish future results.