New research from UC Berkeley Haas and Stanford GSB, published in The Accounting Review, reveals digital footprints can forecast revenue and identify mispriced stocks months before earnings announcements.
With e-commerce projected to account for 40% of global retail sales by 2027, a new study has uncovered something Wall Street has been missing: the number of people visiting a company’s website can predict its financial performance—and its stock price—more accurately than traditional metrics alone.
Professors Yaniv Konchitchki and Biwen Zhang of UC Berkeley Haas and Christopher S. Armstrong of Stanford University Graduate School of Business analyzed digital traffic data from over 1,000 of America’s largest companies, representing roughly 90% of the U.S. stock market’s capitalization. They tracked total visits and page views across all company websites across several years.
“We found that stock prices do not fully incorporate digital traffic information—Wall Street is leaving money on the table,” Konchitcki said. “Capturing this mispricing is not straightforward, but we employ novel data and techniques to extract these gains from publicly traded securities.”
“We found that stock prices do not fully incorporate digital traffic information—Wall Street is leaving money on the table.”
—Assoc. Professor Yaniv Konchitchki, UC Berkeley Haas
Among other insights, the study reveals four key findings:
- Website traffic works as both a real-time and forward-looking gauge of financial health—including sales and profits. Using just one month of traffic data within a quarter, the researchers could forecast quarterly results with reasonable accuracy.
- The data captures information that Wall Street analysts and time-series-driven forecasts miss entirely. Companies with surging web visits consistently beat revenue estimates, while those with declining traffic missed expectations.
- Stock prices haven’t caught up. Investors who bought stocks of companies with the highest website traffic growth and shorted those with falling traffic earned substantial above-market returns, after adjusting for standard risk factors. These above-market returns suggest information is hiding in plain sight.
- The effect is driven by companies whose websites that actually sell things or deliver digital products like Amazon, Netflix, or Tesla—not sites that merely provide corporate information. The commercial activity is what matters.
The high cost of alternative data
The mispricing was particularly pronounced among companies owned mostly by individual investors rather than institutional giants like mutual funds and hedge funds, suggesting that sophisticated investors who can afford expensive alternative data sources are more likely than retail investors to exploit this gap. And even the institutional investors aren’t exploiting it fully, the findings suggest.
“Institutional investors have the resources to purchase such datasets,” said Zhang, an assistant professor of accounting at UC Berkeley Haas. “Because gathering and analyzing alternative data is expensive, markets can’t be perfectly efficient—creating opportunities for those willing and able to invest in the information.”
“Because gathering and analyzing alternative data is expensive, markets can’t be perfectly efficient—creating opportunities for those willing and able to invest in the information.”
—Asst. Professor Biwen Zhang, UC Berkeley Haas
The predictive power of website traffic strengthens as more data accumulates during a quarter. By the second month, predictions improved further—yet stock prices still failed to fully reflect this information even at quarter’s end.
Methodology and implications
The researchers obtained proprietary data from Similarweb, a leading web analytics platform that tracks billions of daily digital interactions worldwide. The dataset covered the 100 largest firms in each of the 11 major industry sectors.
The findings are particularly relevant as digitalization accelerates, the researchers note. Companies like Yum Brands have announced goals to achieve 100% digitally-driven transactions, while Nike CEO John Donahoe called digital “the new normal.”
As consumers continue to shift toward e-commerce, the study suggests that digital presence will have an even more pronounced impact on corporate operations and valuation.
For now, however, the researchers found that a significant information advantage exists for those capable of conducting a sophisticated analysis of website traffic data—before the broader market catches on.
For investors, academics, and practitioners, Konchitchki concludes: “Ultimately, when it comes to capital markets, the ‘secret sauce’ of an advanced strategic investor hinges on gaining an edge through proprietary data, costly training, unique analytical methods, and/or brilliant independent thinking that diverges from the ‘average.’ Indeed, our research paper leveraged unique data and analyses that led to the notable implications that we found.”
Read the full paper:
Digital Traffic, Financial Performance, and Stock Valuation
By Christopher Armstrong, Yaniv Konchitchki, and Biwen Zhang
The Accounting Review, November 2025