If digital presence now determines commercial success, how should markets measure it?
Curated by Business Science Daily — peer-reviewed sources, human-verified.
Learn more
About Our Curation Process
Business Science Daily curates academic research in business and management. Each featured study is selected from reputable, peer-reviewed journals, institutional repositories, or working papers (e.g., Elsevier, Sage, NBER, SSRN).
Articles are carefully summarized to ensure clarity and accuracy, with direct citations or links to original sources. Our process emphasizes transparency, academic integrity, and accessibility for a broader audience.
Learn more in our Editorial Standards & AI Policy.
The answer, according to new research published in The Accounting Review, is digital traffic. Christopher Armstrong, Yaniv Konchitchki, and Biwen Zhang, may be the most comprehensive examination of how web traffic data maps onto financial performance and stock valuation.
The Researchers’ findings challenge the efficiency of markets, the completeness of forecasts, and the adequacy of traditional financial metrics.
The Data Behind the Digital Economy
They obtained proprietary data from Similarweb, a platform that aggregates digital traffic statistics across every public-facing domain affiliated with more than 1,000 U.S. corporations.
Accordingly, the sample includes the hundred largest firms in each of the eleven GICS sectors, collectively accounting for roughly 90% of aggregate domestic market capitalization. This is not a sample of internet darlings or tech giants. It captures the mainstream of American commerce.
Two measures form the basis for the analysis: TotalVisits and PageViews. TotalVisits captures breadth—the number of sessions users initiate on a website. PageViews captures depth—how thoroughly users explore content. A consumer who visits Amazon once but browses 20 pages contributes 1 TotalVisit but 20 PageViews. Together, these metrics capture both reach and engagement.
The findings are striking. Quarterly digital traffic exhibits a strong relationship with contemporaneous revenue, gross profit, and operating profit. More importantly, traffic data from only the first month of a quarter already predicts financial performance for that entire quarter, and predictive power increases as more months accumulate. The information is real-time, granular, and remarkably timely.
The story does not end there…
If digital traffic contains such powerful signals about future performance, one might expect financial analysts to incorporate it into their forecasts. The evidence suggests otherwise.
The researchers measured revenue surprises relative to two benchmarks: consensus analyst forecasts and a seasonal random walk model based on historical data. In both cases, changes in digital traffic predicted surprises. A one-decile increase in TotalVisits corresponded to a meaningful upward movement in revenue surprise. The same held for PageViews, though with slightly less statistical power.
Analysts, it appears, do not fully internalize the signals available in real-time traffic data. This is not necessarily a failure of diligence. Digital traffic data is costly to obtain, requiring subscriptions to specialized vendors. Even when acquired, it is difficult to process. Identifying and extracting traffic data from every URL affiliated with a firm, then calibrating its implications for financial performance, presents nontrivial challenges.
The cost of information matters. And that cost creates opportunities.
The Investment Case for Digital Traffic
Here is where the research becomes genuinely provocative. The authors constructed a simple investment strategy: each month, sort firms into quartiles based on changes in digital traffic. Long the highest traffic-change quartile, short the lowest. Hold for one month, rebalance.
The zero-cost hedge portfolio generated abnormal returns between 41 and 94 basis points per month, after adjusting for the Fama-French five factors plus momentum, liquidity, investment, and profitability. Annualized, that represents returns that would make any quantitative fund manager take notice.
But the pattern becomes even more revealing when you examine where the returns come from.
The strategy’s profitability concentrates among firms with below-median institutional ownership. This is exactly what one would expect if information acquisition costs explain the mispricing. Institutional investors have the resources to purchase alternative datasets and process them efficiently. They are the primary market for vendors like Similarweb. When they act on this information, they arbitrage away the mispricing. Among firms where institutional presence is weaker, the opportunities persist.
Markets are not perfectly efficient when information is costly to obtain and process. This is not a radical claim—it follows directly from Grossman and Stiglitz’s foundational work on the impossibility of informationally efficient markets. But the empirical demonstration, with real trading strategies and real returns, is compelling.
The Mechanism That Matters
Perhaps the most important contribution of the research is not the finding that digital traffic matters, but the explanation of why it matters.
The authors conjecture that the value implications of digital traffic lie in its ability to capture consumer behavior that can lead to transactions. To test this, they manually examined more than 3,600 websites affiliated with their sample firms, classifying each as consumer-oriented or not. Consumer-oriented websites were defined as those enabling direct transactions (Amazon.com), providing digital products (Netflix.com), or hosting detailed product information (Tesla.com). Firms were classified as having consumer-oriented digital traffic if at least half of their total visits came from such sites.
The results are unambiguous. The link between digital traffic and financial performance exists almost exclusively among firms with consumer-oriented websites. The interaction term between traffic and the consumer-oriented indicator drives out the main effect entirely. For non-consumer-oriented firms, digital traffic carries little to no predictive power.
This is not merely a statistical curiosity. It identifies the economic mechanism: consumer-initiated visits, not just any digital presence, are what translate into revenue. A software company’s corporate website, visited by developers seeking documentation, does not generate the same commercial signal as Amazon’s product pages visited by shoppers. The distinction matters for investors seeking to extract value from alternative data.
What This Means for Capital Markets
The implications extend beyond investment strategies. They speak to how we understand firm value in an increasingly digital economy.
Traditional financial reporting was designed for an industrial era. Tangible assets, inventory turnover, and physical capital dominated value creation. Intangible assets—brands, networks, user engagement—were secondary. That world has inverted. The most valuable firms today are built on attention and engagement. Yet accounting standards have been slow to adapt. Research and development expenses are expensed rather than capitalized. Brand value appears nowhere on balance sheets. User growth is reported, if at all, in press releases rather than financial statements.
Digital traffic data offers a partial solution. It provides a real-time, market-wide measure of consumer engagement that correlates with fundamental performance. It is not a panacea—traffic does not equal revenue, and engagement does not guarantee conversion—but it is a meaningful signal that existing metrics fail to capture.
The research also raises questions about the role of financial analysts. If analysts are not incorporating available data into their forecasts, what does that imply about the efficiency of information intermediaries? One interpretation is that the costs of acquiring and processing alternative data are simply too high for traditional analysts to justify. Another is that the institutional structures of sell-side research—short-term incentives, coverage breadth, and reliance on management guidance—discourage investment in novel information sources.
Neither explanation is comforting.
Using proprietary Similarweb data for 1,067 U.S. firms (representing ~90% of market cap), the study shows that digital traffic—monthly visits and page views—is a strong coincident and leading indicator of revenue, gross profit, and operating profit. Importantly, this information is not fully priced by markets, and analysts do not fully incorporate it.
Quarterly digital traffic (TotalVisits and PageViews) is strongly tied to same-quarter revenue, gross profit, and operating profit. Moreover, extrapolated data from only the first month of a quarter already predicts quarterly financials, and predictive power increases as more monthly data accumulate.
FullQuarter3mon (actual quarterly traffic) yields the largest coefficients: e.g., 0.060 for Revenue, 0.129 for GP, 0.083 for OP.
Even Extrapolate1mon (Month1 × 3) predicts revenue (coef. 0.037***) and GP (0.087***), demonstrating real-time usefulness.
Changes in digital traffic significantly predict future revenue surprises relative to both analysts’ consensus forecasts (SUR) and a seasonal random walk model (SURGE). This indicates that analysts do not fully internalize real-time digital traffic signals.
- SUR: revenue surprise relative to most recent analyst consensus, deflated by actual revenue.
- SURGE: standardized unexpected revenue growth relative to seasonal random walk with drift.
- Economic magnitude: one-decile increase in ΔTotalVisits corresponds to ~0.036 decile increase in SUR and ~0.061 decile increase in SURGE.
A zero-cost hedge portfolio that longs firms with the most positive monthly traffic change and shorts the most negative generates 41–94 basis points per month in risk-adjusted returns. Returns are concentrated among firms with low institutional ownership, consistent with costly information processing.
Hedge returns are insignificant (e.g., -0.13% to 0.34%, t-stats < 1.4), suggesting sophisticated investors arbitrage away mispricing when acquisition costs are justified.
ΔTotalVisits and ΔPageViews have incremental explanatory power for quarterly returns even after controlling for net income and changes in net income (Table 8).
The valuation implications of digital traffic are driven entirely by firms with consumer-oriented websites (where at least 50% of traffic comes from consumer-facing sites). For these firms, the traffic-financial link is up to 24× stronger, and investment strategies yield significant abnormal returns (e.g., 1.07–1.31% per month).
- Consumer-oriented: websites enabling transactions (e.g., Amazon.com), digital products (Netflix.com), or detailed product info (Tesla.com).
- Non-consumer: B2B, corporate information, or other stakeholder traffic.
- Robustness: results hold with change specifications, industry-year fixed effects, and Vuong tests (Tables 11–14).
The paper bridges accounting, finance, and alternative data literatures by demonstrating that digital traffic is a valuable signal for forecasting financial performance, predicting analyst forecast errors, and generating abnormal stock returns. It also identifies the economic mechanism—consumer-initiated traffic—and highlights the role of information acquisition costs.
- Novel data: comprehensive coverage across all sectors, not just brick-and-mortar or digital natives.
- Economic mechanism: consumer-oriented websites are the driver, not mere digital presence.
- Costly information: abnormal returns concentrated among low institutional ownership firms.
- Policy implications: as e-commerce grows, digital traffic will become even more central to valuation.
The Limits of the Signal
No research is without caveats, and this study is no exception.
The sample period ends in March 2020, just as the COVID-19 pandemic began reshaping digital behavior. The surge in e-commerce that followed may have amplified the relationship between traffic and performance, or it may have introduced structural shifts that alter the mapping. The authors acknowledge that their data precede most of the pandemic period, and they note that a longer sample would have reduced generalizability because digital intelligence technology was neither developed nor accurate in earlier years.
There is also the question of causality. Digital traffic correlates with financial performance, but the direction of causation is not always clear. Does traffic drive revenue, or does revenue-driven marketing activity drive traffic? The within-quarter analyses, which use early-month traffic to predict full-quarter performance, suggest a predictive relationship, but the underlying mechanism remains a matter of interpretation.
Perhaps most importantly, the research does not address the ethical dimensions of alternative data. Digital traffic information is derived from consumer behavior, often collected without explicit consent. Investors using such data benefit from information asymmetry—they know more about a company’s prospects than the company’s own customers do. This raises questions about privacy, fairness, and market integrity that the research does not resolve.
Why It Matters
For investors, the implications are direct. Digital traffic data offers an edge that is not yet fully priced into markets. The abnormal returns generated by simple trading strategies suggest that sophisticated investors can profit from incorporating alternative data into their models. The concentration of returns among low institutional ownership firms also suggests a window of opportunity—one that may close as competition intensifies.
For analysts, the findings are a challenge. If digital traffic predicts revenue surprises, failing to incorporate it represents a missed opportunity for forecast accuracy. The cost of acquiring such data may be a barrier, but so is the cost of being consistently wrong.
For researchers, the study opens new avenues. The identification of consumer-oriented websites as the mechanism suggests that sector-specific approaches may be more productive than broad-brush analyses. Future work could examine whether other forms of digital engagement—social media activity, search trends, app downloads—offer similar predictive power. The relationship between alternative data and institutional ownership also merits deeper investigation.
For regulators and standard-setters, the research adds to the growing pressure to reconsider how intangible value is reported. If capital markets are relying on data outside financial statements to assess firm value, the adequacy of current reporting standards is called into question.
The Broader Debate
This research participates in a larger conversation about the nature of value in a digital economy. For decades, accounting scholars have wrestled with the growing gap between book value and market value. The rise of intangible assets has rendered traditional financial metrics increasingly irrelevant to valuation. Digital traffic is one of many alternative data sources attempting to fill the gap.
But there is a tension here. If the most valuable information about firms comes not from their financial statements but from third-party data vendors, what does that imply about the role of financial reporting? Are we moving toward a world where fundamental analysis requires subscriptions to proprietary datasets? If so, who gets left behind?
Individual investors, smaller institutions, and retail traders lack the resources to acquire and process alternative data. The information asymmetry between sophisticated and unsophisticated investors may widen, with consequences for market fairness and capital allocation. The research does not address these normative questions, but they hover in the background.