A new study by experts at Wharton and elsewhere shows how AI
can stop innovation declines at early-stage firms after an IPO.
There’s a new frontier for the ever-expanding capabilities of AI: reviving innovation at firms that have slackened on that front after an initial public offering (IPO).
With increased deployment of AI analytics funded by the capital raised in an IPO, firms can reduce some of the “innovative penalties” they face, according to a paper titled, “Innovation Strategy after IPO: How AI Analytics Spurs Innovation after IPO.” Wharton professors of operations, information and decisions Lynn Wu and Lorin M. Hitt wrote the paper with Bowen Lou, professor of data sciences and operations at the University of Southern California.
The paper noted that the decline in innovation after an IPO is counterintuitive. After all, following an IPO, a firm should have reduced financial pressures because it has newly raised capital. The new financial strength should increase incentives to engage in risky innovation. Firms could also reap reputational gains after an IPO, which could make it easier to acquire talented and innovative employees, strengthening bargaining power with suppliers and signaling quality to consumers.
The reality, however, is that the “quality of the innovative output” tends to drop after IPOs, especially at early-stage firms, the paper noted. That “innovation decline” occurs because of three factors: the newly public firm prioritizes short-term financial goals (e.g., quarterly earnings) at the cost of longer-run innovation payoffs; it must meet higher disclosure requirements, which discourage innovation; and it has lower managerial and employee incentives to pursue innovation following the dilution in ownership with an IPO.
Wu explained how the incentive to innovate could be reduced when an IPO suddenly increases the number of owners. “You don’t want to take risks in innovation anymore because if you’re successful, the gains are shared with a lot more people. And you don’t always get appropriately recognized for those gains. Furthermore, to meet quarterly earnings and financial disclosure requirements, it is difficult to think beyond a few quarters. These can all discourage long-term innovation,” she said. “So managers tend to be less innovative in their endeavors.”
Those factors may encourage newly public firms to shift their focus to lower risk, incremental innovations that are more likely to be built upon their existing stock of knowledge, and discourage them from pursuing riskier novel or breakthrough innovation, the paper pointed out. Innovation declines are seen also in firms that get acquired.
AI Analytics a Superior Option to Protect Innovation
For sure, options other than an IPO are available to firms that want access to more resources but also protect their innovation: for instance, they could use leveraged buyouts or mergers and acquisitions, the paper stated. But the success of those strategies is not certain, and they are often expensive, risky, and entail significant organizational change, it added.
In other settings, firms that recognize those threats to innovation have taken conscious steps to avoid them, the paper noted, citing prior research. Dell, for instance, reverted to private ownership through a leveraged buyout in 2013, to prioritize long-term innovation. Swedish furniture maker Ikea, on the other hand, chose to stay private to maintain its focus on innovation.
AI analytics is a more attractive option for newly public firms to maintain the innovation momentum, according to the paper. (The authors defined AI analytics as analytics related to data analysis and machine learning.) Recent advances in analytics technology, enabled by new machine learning capabilities and increased digitization efforts, could potentially mitigate some of the innovation declines that newly public firms face, without requiring complex financial restructuring, the paper noted.
A particular casualty at newly public firms is innovation that involves “re-combinations,” or new combinations of existing technologies. “But AI analytics helps to generate new combinations, Wu noted. That is because AI analytics brings capabilities to facilitate “the process of searching, aggregating, and mining diverse knowledge, which enables new combinations of [existing] technologies,” the paper explained.
According to the study, firms acquiring AI analytics capabilities after an IPO experienced less of a decline in innovation quality compared to similar firms that had not acquired those capabilities. Thus, AI analytics helps such firms overcome the negative effects of short-term focus on earnings and higher disclosure requirements. Specifically, AI analytics, as measured by the skillsets of employees, can help generate innovations that are new combinations of existing innovations. However, AI analytics is not a panacea for all innovation woes in a firm. AI analytics had only a limited effect in addressing the concerns around reduced managerial incentives for innovation, and in generating radically new innovation.
The study analyzed patent data at 1,471 firms from 1988 to 2013, of which 1,080 had an IPO and the remainder stayed private. It extended the data to 2019 to examine the effects from recent advances in machine learning. Wu said she expected the trends shown in the study to continue with the advent of GenAI. “The large language models are really about turbocharging combinations,” she noted.
Highlights from the Study
The biggest innovation gains for newly public firms that invested in AI analytics were in manufacturing (61%) and information technology (19%); the least responsive were those in real estate (0.14%) and utilities (0.20%).
AI analytics is helpful in driving innovation of combinations of existing technologies, but that could not be established for radically new innovations.
The requirement of information disclosures can be daunting for newly public firms in industries with long product cycles such as energy and biopharma, and impede their innovation efforts. AI analytics can be particularly helpful in those settings by “decreasing the negative impact of information disclosures on innovation.”
Within AI technologies, machine learning skills are critical in driving innovation, especially in new combinations of existing technologies.
The study tracked job postings, resumes and online job reviews to study how post-IPO firms are ramping up their AI capabilities.
The study is among the first to examine how AI analytics and related emerging technologies can be used to mitigate the innovation decline after IPO and the underlying mechanisms.
All considered, Wu pointed out that the takeaway from the study is not that AI analytics is a cure-all for innovation problems. “AI is general-purpose technology, but it is not something that you should use as a hammer on everything,” she said. AI can only partially address the innovation decline after an IPO, and help solve problems with short-termism or increased disclosure requirements. It still does not solve any of the innovation declines caused by reduced incentives, she noted.
